NYU Startup School: Creating Your Social Media Marketing Plan


I've done these sessions on different topics at Google Analytics.

This is a first for social, although I do teach a semester-long class as part of SPS as well as at General Assembly.

So, I'm going to introduce a whole bunch of different frameworks and we can power along as we go, because the assumption is that you're all coming from different backgrounds, you have different experiences.

So, the goal is that you really can take the frameworks and apply them into what you want to do and how you want to apply them.

Alright, so, that's pretty simple agenda we'll reframe social media as a discipline looking at the landscape and then look through the rules and guidelines.

What makes up social media marketing today? And then the frameworks as part of the strategy.

So, this is my trust me slide.

Sarah already talked about most of this stuff.

I lead the digital marketing team at Kickstarter.

We do email marketing, we do content marketing, we do paid advertising on social.

So, super relevant.

I also teach here at NYU and previously I was at Stack Overflow, Maker Bot.

So, the theme here is that I've worked with really large scale devoted communities at organizations, at these like high-powered technical organizations.

So, a lot of my examples would be using community to drive ROI forward.

And there's my dog.

Always a fun thing of put in here.

This is Lucy.

She also has a social media account which I run.

And just a proof point is she did get into a Swiffer commercial before, not sponsored by Swiffer here.

But there's power for a dog social media account to actually make me her owner revenue.

Alright, so I'm gonna split this into four sections.

We have research, planning, execution, then iters, iterating and optimizing.

You can apply this to really like product design.

You can apply it to most things that are modern skillsets, so, I feel like it's a good schema to start.

Alright, so for research first and foremost here is a quick video, it's Socialnomics.

This dude does this video every year, there's a 2018 one coming out.

It's just a really great overview of the current statistics and information of how social media and digital advertising works.

So, we'll start here and then move into the primary content.

Ignore the music.

[Music] Alright, so there's some themes there.

But overall I do want to highlight that, I mean, digital is growing, its, the amount of consumers and users on digital.

But in addition to that the spend is predicted to increase over time, no surprise to anyone.

But what I do want to point out is the spend on social media both organic and paid is increasing.

So, we'll talk about both because more and more they work hand in hand.

That's one of the things that we are working on at Kickstarter.

It's if we start a new account how does paid actually drive it? If it's not just ROI what do we use and how do we justify spending on social media? And if you're looking for additional data Force really does awesome reports on the stuff, so does IAB and Barrel.

They conduct annual reports about how SMBs and other organizations use social media.

So, first framework of many is prospect quality and conversion probability, two terms that I think will frame all of these channels pretty well.

Why do we use so many frameworks? One, social media changes constantly.

And two, each individual discipline is kind of like a puzzle, piece in a puzzle that you put together yourself.

So, this framework helps because it combines brand awareness with that kind of ROI mentality and conversions.

So, prospect quality is the idea that someone needs to understand the values of your organization to be able to be an active participant.

So, for Kickstarter, for example, when I talk to people about Kickstarter and they have heard about us they think of us primarily as crowdfunding.

I would say that is good prospect awareness or prospect quality, but on the lower end.

Because we are actually a public benefit corporation.

How many of you knew that? So, Kickstarter is a public benefit corporation and which means that we have to, we are held accountable for sacrificing revenue to increase the quality of the creative and arts community.

One example is we give 5% of all revenue every year to arts organizations, arts nonprofits.

So, if you understand that fully, there's high prospect quality.

And it's like that deep core of the organization.

Do they know that? I mean, that comes after many touch points in overtime.

So, first and foremost this prosthetic quality.

Some channels are better at aligning with it than others, but it is an overall journey.

Secondly you have conversion probability.

So, the likelihood the person will take your intended action.

Conversion really is an umbrella term here.

It can be a transaction, some monetary conversion could be a sign up or a download.

There are so many different kinds of conversions that you can align.

And in that thought you can also say well, perhaps if sign up is a lower conversion probability where eventually we want them to be a monetary participate.

So, all these things aligned in spectrums first is display, just seeing advertisements, or I guess maybe not seeing but triggering advertisements on properties such as websites and apps.

The conversion probability historically has been super low here.

It's just because with when people see these kinds of placements, it's interrupting their current experience in the long-form content.

So, if I'm looking for ten ways to bake a cake I've already found the content I want, I don't necessarily need to click an ad to go somewhere else.

You're very much removing them from the experience.

And then it's along the prospect quality matrix on that y-axis because it could be super poorly targeted.

So, Carnival Cruise is one of my favorite examples because they got some bad press after showing display ads for their beautiful cruises right after their cruise ship had stopped working for like 48 hours.

Do you remember this? And the bathroom stopped working for all people on the cruise ship for 48 hours.

And they were showing ads alongside these articles So, this is an example of contextual placement gone wrong.

And display, the technology is getting better definitely.

But it's not necessarily, it doesn't guarantee a higher prospect quality there.

Then we have retargeting, it's kind of a catch-all.

So, this is when someone visits your property and then you use cookie tracking or some other kind of technical tracking to then show them your message again.

It has inherent high prospect quality just because they've had to have been on your property.

Primary examples of properties are websites and apps.

These are things that you own as an organization, you have exclusive rights to the data that's collected about the user.

So, for example, NewYorkTimes.


Although we have an article, like Kickstarter might be mentioned in an article there, we don't.

Well, it's up to New York Times to give us that data for the performance.

We don't by default have access to that, the page views, the time on page, and stuff like that.

The weird middle ground, just to solidify this information, is social media accounts, because let's say Kickstarter, we have a Facebook account and when we post organically Facebook Insights tells us how many likes we have, the demographics of the users.

I've screenshots and we'll walk through the reporting later.

But you can bet Facebook has a whole bunch more information about these users, about the post performance that they're not giving us.

So, we get some information there, and it's technically our property because we choose what to post and how to interact with users.

But it's in that really weird middle ground.

So, your properties allow you to retarget because once the user comes to your website you can then target their cookie across other properties and show them ads.

So, they've been there so long as your website is great there's higher prospect quality.

Next, email.

Email in the scheme of things has both high conversion probability and prospect quality.

Because of can spam laws.

The user must consent to receive messages from you otherwise it's illegal, super-duper illegal.

So, the prospect quality is high and, given the one-on-one capabilities of sending this person an email because they sign up for this product and we want to show them similar products, the conversion probability or the potential for that is higher.

Two more.

So, search.

Search is known as the high intent channel.

I like thinking of it as finding the needle in the haystack for a user.

You want to find the best cat gif.

So, you search for it in Google as the head librarian of the largest library in the world needs to send you the best content related to your search query.

So, it's just up there because it puts what the user wants to see directly in front of them in the shortest amount of time possible.

And then we have social media.

If it was up to me, overall it kind of is, I put social media just across the entire screen.

When you have organic social media and paid social media together it's like really compounded reach and compounded impact.

So, there's higher prospect quality for organic especially, because to see a post you need to follow the brand right or a friend needs to post it.

We're looking at the standard newsfeed.

Or it's trending.

We have just good faith in Twitter Facebook and other channels.

They're actually showing as the content that we value.

So, that's why it has inherently high prospect quality.

And then because of the ad targeting capabilities we then allow it to the top right along with search and email.

So, there's a lot of ways to use it.

And that's why we have some of these frameworks.

As a quick warmup I would like you to answer these five questions over the next minute.

Take out your fancy dancy devices, please, and open any social media app.

It doesn't need to be Western, Eastern, whatever fits your fancy.

Find a post about the Oscars.

Does the content match your expectations? And then, is the post organic or paid and how do you know? And finally, What is the objective of the post? These are very broad questions intentionally, so that you can interpret them in multiple ways.

Don't worry the buzzer will make an annoying sound, so you will know when time is over.

Alright, who, so, raise your hand please who wants to tell me how they found it and what posts they found.


Oh, Yes, sorry, the recording.

Which is why I'm doing the Britney Spears right now.

-I just went on Instagram and searched Oscars, the -top Oscars or whatever.

And then the post that came up, actually I clicked on three, and each of them were a hairstylist -posting.

So, and they were in the top views or whatever, so I don't know, it's probably not paid necessarily, but I mean, -it's engagement I guess or notoriety.


I mean, for that platform specifically it wouldn't be paid.

But the question is why do you see those posts? And there's a combination of things.

If you, let's flip this, if you are the social media channel, how do you determine what users should see? So, on average on Facebook users have the potential to see 6,000 posts every single day.

And we're just looking at the newsfeed.

I will never see 6,000 posts, because I don't have the patience or attention span to do it.

But we're looking at the average Facebook user has approximately 2,000 friends and those friends post 3 to 10 times a day.

Then they also follow a dozen or two businesses.

Those businesses post way too much and then like yada, yada, yada.

But so, there is an oversaturation here and as the social media channel you need to figure out how to show the right posts to the right people.

So, for these it could be, perhaps, these are the posts related to the Oscars that had the most amount of likes or engagement and therefore that's the signal.

In addition to that we also look at something called affinity and affinity categories within social media.

So, that aligns you with excellent quantifiable score to individual topics such as dogs, pets, pet food, I don't know, you know, I'm like a dog guy, and like NFL, whatever it is based on your previous interactions and on your financial data.

It is known at this point that Facebook and Google especially they use data from large organizations such as Nielsen, and Axiom, and Epsilon to combine your actual digital presence with your financial information.

So, we know that if you tend to buy pet products then you should be put in that category for advertisers to target you.

If this is new to you the only questions I really need to ask on your phone, which has a unique mobile ID, do you use an email app, a banking app, and a social media app? All we really need to do in the US, privacy laws are different all across the world, All we need to need to do in the US is tied together with a unique device ID.

There's other ways as well.

If you do want to nerd about this let me know.

There's something coming up in the EU called the GDPR where a lot of this cookie tracking and so forth, I mean, well, I guess not be allowed.

I agree with the premise as a whole though.

It's like we have our privacy and data, and we should like to use it however we'd like.

So, depending on where you live there's different levels of this.

But we can use the moral of the storie's that these channels have so much information about us to make sure that we see what we do see.

That is the sharing app that promised there's more interaction later.

All right, so, let's look at the landscape.

This is the most practical way.

I've actually been able to layout the landscape.

And it's based on user intent.

So, if you're you are a PM or product person, this should look fairly familiar.

There is discovery and search.

The intent and user basic behavior here might be better explained if we replace the word discovery with native.

So, native placements.

This is where users aren't necessarily looking to find your brand, your service, and so forth.

But it appears, and they're very delighted, they're super happy about it, and search is that needle in the haystack.

They're like, I want to find cat gifs, so please show that to me.

So, this is how I've aligned it and they can mix and match.

It's actually really hard to just put them into one individual line.

So, Snapchat is probably the most discovery oriented app.

They don't like to be called a social media channel, but there's a inherent social media aspect to it.

So, this is like you aren't searching necessarily for information about New York and like a New York event, but maybe they'll surface something for you and you follow people whose surface content that you weren't necessarily looking for.

The opposite side of the spectrum, Pinterest is, Pinterest has really great search capabilities.

They're arguably the top search competitor to Google other than, I guess, Amazon at this point.

If you want to learn more about it I would just look at this product they're launching called Pinterest lens.

It allows for auto tagging of photos.

So, you can take a photo of, let's say, me, an old tag, the blazer with the organization, a link to the product and so forth.

That's the dream of it.

Google is countering it with the same kind of functionality within Google images.

So, Pinterest Lens.

You can search for things and actually get.

You can like use it with high intense search and then so many things in between.

So, this is how I like laying it out.

The additional layers on top of here include the user type.

So, Pinterest, for example, skews female, Snapchat skews younger, and then we just saw that Twitter is skewing grandparents #AARP.

[Laughter] And all of.

These are things that I don't necessarily need to go over because every organization publishes these stats.

Sprout Social is one good one.

They have blog posts where they use their own data to show the demographics and the use per social media channel if you don't want to pay there's like million dollars to the Nielsen report.

So, discovery through search.

This is how I think we should align our efforts.

Are we going do we want the user to discover us or do we want to put us like right in front of them? There's true merits to both of them and we'll uncover these don't unpack them as we go.

So, additional layers all these channels have paid and organic where the organic, again, you either follow the brand or your friend shares it.

And then based on those two things that services through the affinity categories.

Like you have a strong affinity towards dogs, therefore, when it's someone shares their post about dogs it's going to show up in front of you.

On the other side I'm sure a lot of you saw paid posts as well during that one minute search.

That was targeted towards me, you know that's a great placement, I immediately bought the.

Dog domain.

And this is where we use information about the user to then show exactly what we think they want to see in that placement.

Caveat on both sides though is still native.

So, I want to explore the word contextual.

Contextual placements has existed far, like, before social media advertising.

We can even use, let's say, like a magazine as an example.

So, you would talk to an ad sales rep of Popular Mechanics and you would talk to them about, alright, well, I have this ad about this product and it's in this category.

So, what should be the content before and after it? Like, they have editorial content and we want to make sure that if someone reads about crowdfunding our ad is either immediately before or right after it because it warms them up.

So, translate that to your product feeds now.

On Facebook you have a feed and it's like political, political post, ad about Kickstarter, and then babies.

Something there doesn't quite fit as well.

And that's where you really get the advantage with all this like creepy per se targeting to show that right message to the right person the right time.

So, additional framework, we are all scaffold on top of it.

So, this is the sales funnel.

Moving on.

I think that overall it's not dead, it's just less practical in social media marketing because it is ever-changing.

And we have so many data points to contact users in the lowest prospect quality stage possible to the highest prospect quality phase because we still want to interact with them.

It's a great way to build brand loyalty.

And this is the problem.

This is not adaptable to the standard sales funnel.

So, what we have is awareness, so many different ways to make people aware and reach them about our brand.

But then we need to bring them back to our properties or bring them back to our core messaging.


So, I read another report.

Does anyone know how many digital touch points on average it takes to convert a user? So, first of all it's like all of the core depending where you read.

So, to convert someone into a user, which isn't necessarily monetary, is six to eight touch points and then on for a monetary participant it's like, it can be as high as twenty eight.

I definitely I have some links at the end of this to cite these stats as well.

So, we're like, alright, well, we need to contact people how many times? It definitely affects your strategy.

And this is one of the main proof points for social media because you have different channels where people absorb.

And they ingest different kinds of content.

So, you actually have all this power, all this targeting at you.

So we're trying to scaffold on this knowledge right now.

And this is where retargeting can come back into play.

Based on what the user has seen on your properties you would, it's in your best interest to show them similar content.

So, kind of like a loaded question, but what is, what will have a higher conversion probability retargeting someone from your home or a product page? You're like hey, I'm gonna show you an ad to get you to be a monetary participant.

They go to your homepage.


Yeah, it's a loaded question because ideally it's the product page because they've been able to get further down in that prospect quality.

And I think that's one of the.

That's one of the reasons why it takes so many touch points because the home page is great.

It's the entry way into your brand, but it doesn't necessarily instill the values that cause them or exper them on to actually purchase.

So, this part of the funnel, its consideration.

Moreso nurturing the user.

There's a couple ways to go about it.

One, you supplement across multiple channels, social media is key here.

The other is we.

Let's get their email address.

Let's use lead generation and then follow-up nurturing through email marketing.

I'm a huge fan of email marketing, but the counter-argument we're like, because we are all here for social media, counter-argument is that email marketing is one placement, is one inbox per email address, whereas social media has multiple.

We'll go over that as well.

So, there's definitely a strength there.

Even though for email marketing once you generate their email address you aren't paying for repeated visits you're competing against not just your competitors, you're repeating, you're competing against.

You're competing against all their friends, all their family members, all their co-workers for the same individual placement.

So, it's problematic because marketers have ruined e-mail for everyone.

And then there is like evaluation and that fits into this model specifically.

It's the Harvard Business Review customer decision model.

Here it goes into consideration versus evaluation.

Evaluation is like when you're searching for the right shoe for you and you have 40 tabs open.

When you get to that point, you know what one, two, or three product considerations are best for you.

So, this is like the highest prospect quality before conversion.

They're like alright, I'm looking for shoes, and I really really care that they're durable, so I'm gonna look at credibility and reviews.

And I also care that, I don't know, they're like my kind of style, or customizable, whatever it is.

Whereas consideration is more so like a topic consideration.

Where they're like, is this industry, is this kind of thing right for me? So, does anyone know what this is? It's the Met, absolutely.

I like using the Met here as an example because if we look at consideration versus evaluation within this funnel.

And we're, I swear we'll go straight back to social media soon.

There's two sets of competitors.

And the former chief digital officer of the Met, I saw him speak recently ,so we use this exact example.

For evaluation the competitors are like the MoMA, you know, it's like other museums where they're like, should I go to this one or that one? What about for the consideration? What do you think the competitors are? His answer is Netflix and Hulu.

They need to get people to be like, I should want to go to a museum today, this would be great for my life.

And then afterwards they evaluate the suite of museums and products within that.

So, this is how I've broken it down overall.

So, now I'm combining multiple frameworks into one.

You get the awareness.

It's like, have they or have they not heard about this thing at all? Secondly, do they understand the importance of the topic? So, this is the like, all right, am I gonna go watch Black Mirror or am I kind of go to a museum? After that then it's product consideration.

And I think the fallacy here and when people are constructing their marketing campaigns is that they think that they can jump from awareness to product consideration.

Which is why I think this step needs to exist especially in social media marketing.

And then you have evaluation and then things in between for advocating and bonding.

Also for those of you taking pictures.

You will get the deck so long as you fill out the survey at the end of this session, so you'll get all like 70 slides.

All right.

So, my assumption is that you are all here representing businesses or your personal brands.

What I'd like you to do is another quick one minute knowledge check.

And I want you to ID your topic.

So, the consideration.


Who are your consideration competitors? What is your topic? And then, what is your intended conversion action? What is that key for success? You now have to write these down to remember them because we will reference to these three things as we go along for the rest of the lesson.

All right.

-Hello, I am Usica Felula, I lead strategy and marketing at Payant.

We're a FinTech solution to medical bill payments.

Our -product values are clarity surrounding billing, convenience for payments, and options for payment methodology.


-And our conversion is both download and using the platform to handle payments Perfect.

And then we'll spice up the social media channels to figure out what's best for each of those as we go along.

So, combining all this together again, love social because it really is the combination of being able to use your, like flex your brand muscles in addition to targeting to meet these desires of users.

The full slide is actually this is just storytelling.

We have combined the Mad Men era with the programmatic days.

We're like Mad Men.

We're like all we care about is the prefrontal cortex and how people naturally think about things.

And then programmatic we're like why not just let robots optimize all this stuff for us.

And we're finally at the age that social media has enabled and allowed us to do.

We can be like, what if we make, what if we are Red Bull and we make a documentary and then we spice it up we show it all the different social channels.

It's sliced differently per social channel because they're different people.

So, we are allowed and enabled to do this kind of stuff now.

Which brings me to how do we actually figure out the splicing of the user base? We have two models here.

This is the first one.

My background is actually product marketing.

So, this is when as a product marketer or PMM I've worked really closely with UX and UX research.

So, this is a pretty standard UX research user breakdown.

We look at the overall user base, so the current user base, and this can be very very very low touch points like they have made an account, but and have never done anything.

Then we break it down to consideration.

And this expands beyond the just the concentric circles because there are people that are not in your user base who are actually interested, or could be interested in your product or service.

A main KPI here is actually just visits to your website or property in addition to email signups.

So, if we transition slightly to growth marketing as a discipline, like, you need to grow as quickly if not faster than the industry that you currently exist in, so, your topics.

And you, and since it's such a soft metric here, you really are looking at can you get people interested enough to give you data about themselves? I'm explicitly giving you an email address.

After that we have evaluation and then four, which is the loyalty loop, so, people who are active participants within your brand or your, within your brand product or service.

And then I always feel like unless you, I mean, even if you're like Amazon or Google, there's an additional addressable audience on top of that.

The only reason I like sort this out is because the user behavior and the likelihood they participate and all of your posts is so much higher.

And if you are going to use advertising those people in purple tend to come with a premium.

And so, I won't give you all the stats, but it's approximately 10x to target.

The purple for Kickstarter is user base.

Then it is for two which is overall consideration.

And you need to figure out whether or not that's worth it for you.

So, breakdown.

Total user base, overall consideration, the evaluation, and this quantity of people those who are converted and becoming advocates for your brand, product, or service, and the immediately addressable audience, oftentimes the premium, but a much higher conversion rate.

So, one additional way to look at this between two and four is the spectrum on the conversion probability.

All right, so, two.

Then how do you segment users? Every single social media platform and other ad platforms as well can work within this wedding-cake model.

So, the wedding-cake model here as the base which is demographic and geographic information.

Historically we've been like alright, well, if we target people in St.

Louis whose skew male are in our 18 to 22.

We can assume they don't have much like disposable income.

They're probably college students, so forth and so forth.

But it's not very defensible and if we were already talking about over saturation in the market you can bet that people new to advertising, that's how they default for targeting.

So, we need to get further along for an additional reason.

It's because we have, again, all this technology, so let's craft particular messaging to the right people at the right user stage of the funnel.

And that's where behavior and psychographics come in.

So, behavior comes in to like retargeting, let's say.

They have done something on your site, may not be the homepage, but perhaps you combined it.

They have been to the FAQ page and they've been to the product page.

There is high intent there.

Or perhaps they read a user review.

These are all things that you can use as targeting.

With Facebook in particular what you need to do is just install the or inject the tracking pixel.

All of these platforms have their own pixels that allow them to access data from your site.

So, a lot of these sites also have what we call a global installation or universal code.

All I need to do is go to your ads manager, generate the pixel, and then copy and paste it either in the header or the footer of your homepage.

And then it globally deploy your, or just kind of like blast across your website.

If you would like, I can send you a tutorial after this session.

So, when some.

And that's how we're able to actually take those users and track things and then surface ads in front of them on those platforms later.

We're kind of just combining the data from both properties in that sense.

So, behavior can come in the form of just what they're actually doing.

At Kickstarter, and I'll show this exact example, we have these very well branded video advertisements.

And what I do is I create retargeting audiences based on people who have watched 3 or more seconds of it because we put the value proposition 3 seconds in the video.

So, and to me this is kind of like loose math.

I'm like, well, if they're on it for three seconds, they didn't just accidentally watch.

It didn't like auto trigger, and then they paid attention to something else for a second.

And then we have psychographics.

So, what are they actually interested in? And this is where we can use that financial information.

The most granular example I have which is also not something I can ever see myself using.

With Twitter advertising allows you to target people by CPG category and CPG product.

So, I can, you, we can all target anyone who has bought Heinz ketchup specifically.

And I'm like, alright, well, if that's your use case, great.

But it's just how far down you can go.

So, for us instead what if someone has bought art supplies and we want to ask them to be a crafts creator.

Or someone has bought tickets to, tickets to a museum or has subscription to a museum.

These are all things that are very rich for us to target and layer onto the second point of or the second tier of that wedding cake.

So, we have the basic demo NGO, we put behavior and psycho on top of it.

And there's one last section that we can generalize across all of these platforms and that's placement and devices.

Placement is a little more complicated, so we'll go into that in a bit.

But devices.

We are all living in a mobile-first economy at this point especially for search engine optimization.

We're calling it mobile getting 2018 because Google has said they will index mobile sites first.

And in short, what does that mean? It's just that like people are consuming content on these things and these things are getting better bigger and faster.

That's like the gist of it.

So, two additional proof points.

Facebook in September started prioritizing positive mobile experiences and that's advertising.

So, if you really want like, let's say, we go further down, we read between the lines a little bit.

If your site does not load quickly on mobile, then you will be charged a premium or your ads will not be prioritized over organizations that do have properly optimized mobile web sites.

So, we're already living in an age where both your advertisements, your social, your search, everything digital marketing related is benefited or suffers based on your mobile experience.

It's a big thing that we're focusing on Kickstarter now.

We're trying to get all of our pages to load under two seconds on mobile.

It's just an industry benchmark that we know of.

And in the same way you can measure social media.

Using like third-party web analytics and app analytics you can measure this as well.

I have a quick screenshot at the very end.

All right, and then placements.

This goes into, here's the TLDR.

This is Facebook.

It's like you can show ads on the feed, right rail, in messenger.

I still have not seen an ad in messenger, so, I'm not sure where these things are going.

In stream and so forth.

So, within each network there are so many individual placements as well.

So, that's where it's like all right, we have figured out who these people are, what they value.

Now we just need to show it to them at the right place, at the right time.

And that should be full circle and really comprehensive for your initial planning for these social media.

So, here's a concrete example.

Last year we launched a new product.

It's called Drip.

It's a subscription product that really it's for the longevity of creativity.

So, Kickstarter allows for individual projects and a lot of people launch multiple projects.

However, what if after your first zine you need to publish editions 2, 3, 4, and so forth? Drip allows you to have subscribers on pledge and support you months a month, like a dollar a month, $15, and so forth.

There's an entire industry for this but we want your product here.

So, I've already talked about if you have high prospect quality for Kickstarter you know we're a PBC.

This is our mission.

We are the world's largest community of people dedicated to bringing creative projects to life.

That is why we exist.

So, we have like such great success metrics like billions of dollars pledged to projects and like hundred forty thousand successfully funded projects, millions and millions of back various, millions and millions of individual transactions to projects.


So yeah, putting this into place.

This is how I think about our funnel.

We have, all right, we need people to hear about us and crowdfunding in general.

Super fortunate that South Park has taken care of a lot of that for us.

If you haven't watched the episode please don't.

No, it is humorous.

So, there's top of consideration about it.

But then after that one of my big challenges is figuring out how do we get people to understand Kickstarter as a PBC? Do we need to for a transaction? Not in most cases, but for retention and getting people to be loyal to the brand absolutely.

So, we might as well start earlier than later with our messaging, evaluation, and then getting people to that where if they do want to support the creative economy, they think of Kickstarter first.

That is the goal.

It's like what success looks like.

So, here we have the overall audience, and 2 I've replaced as creatives, 3 – active users, 4 – monetary participants, so, they've pledged, and so forth.

And then here.

So, I've just adapted it really slightly.

These are meant, these frameworks are meant to be templates for you.

So, when you do get the deck feel free to like slice and dice it however you want.

So, this is what we did.

We have these videos of creators, these really successful vibrant creators.

We have them for every single category.

Our categories are dance, arts, music, film and so forth.

So, we have videos for all of them.

However, video content doesn't necessarily convert users.

But that's fine for us because we're assuming that we need to drive people through prospect quality as quickly as possible.

That is the objective here.

And based on how they interact with these we'll retarget them.

We'll just, we'll help them through the process.

But we need a little bit more behavior data because we're already jumping into the second tier of this wedding cake.

So, first is Julia Nunes.

Ah, man, I love her.

I've loved her since before I joined Kickstarter.

She has launched multiple projects.

She started out as a YouTube star doing ukulele covers.

Since then she's toured with Ben Folds, Weezer, and so forth.

Not the point of this lesson, but check her out.

So, I'm gonna show you two videos one of Julia and the other one of Lucy just because it'll put things together.

So, here's the ad for lowest prospect quality to drive them through.

Did anyone catch the proof points in there by the way? So, we have the topic, we have the parted user funnel, and then we inject it.

Oh, wait, you can make money here? That's great.

We support creativity.

That's awesome also! And they are strategically placed throughout.

So then, here's another one.

This is Lucy Sparrow.


I don't know if anyone went to this.

She opened a bodega in New York.

And everything was made out of felt.

It sold out so quickly.

She's super awesome.

[Music] So, if someone watched more than three seconds of it we would then send them to different parts of the website that are meant for them.

So, we would make individual landing pages as well.

And we're actually launching a whole sub domain called creators.


Com that has all these individual entry points.

There's hundreds of pages we have planned to supplement these because we want to trade creators as the unique individuals that they actually are.

And to me that's like aligned with like respect and dignity for your users.

We're not spraying and praying because it doesn't work and because our users don't deserve that.

So, let's just like jump further into the funnel.

Now we have content.

Social media and content marketing are so hand-in-hand.

They really can't exist without each other.

So, here's a skill share that we made.

So, this is Stephanie.

She is our former creator, creator engagement manager.

And we can see here each of those pieces of information they're not that long.

But those are all the pain points that you just go through.

So, what we're trying to do is eliminate the barriers to the right people immediately.

And then we follow up after that with individual posts about all of those pain points built out.

So, based on the content they skip to they're like oh, wait, I want to self-select into finding your project story.


So, if they do that then we should send the more content about that because they've already expressed interest.

And this is where the data really comes into the user behavior.

And we are just in the second tier of that cake.

So, we are tier 1 and tier 2.

After that you figure out the device and the placement in addition to it.

So, there's like a whole world in addition to this.

Message to address their concerns is super highly segmented.

And I think there's a good point to make about structuring your team as well.

This is how people used to hire digital marketers and social media marketers.

You would have like creative social person.

This person's like this looks great.

I Instagram all of my food.

I want to put this out there.

And then you would have a data team that measures it and gives them reports, and then a technical team that would make the pages, and so forth.

However, nowadays You need to have it all together at once.

And I think that all the proof points and all the things we've been talking about so far, sorry that was one-sided, the things I've been talking about so far, so sorry, this is this will be interactive later, support this because you need the right content to the right people the right times and you need the analytical scale.

So, you slowly test out what works for them as well as a technical skill to deliver to them at the right times.

So, content really needs to work with social media because it exists there.

It really is a property that you own even though technically you don't get all the data you could get if it was on your website.

So, this isn't I'm gonna say age-old saying for digital, which is not really an age-old saying for the world.

But content is king and distribution is queen.

Even further than that.

Jonathan Perelman former SVP at BuzzFeed said: well, distribution wears the pants in the relationship.

What I've heard recently though is that context is the royalty overall.

I think this should resonate now after we scaffolded to this.

It is that regardless of the content or the distribution, I mean, those are givens, you need both of them at this point.

So, like let's stop having arguments about which one is better Let's figure out overall for each user and for each objective what mix of both is necessary to put the right content in front of the right person.

Because it's contextual placements at this point.

Section one – done.

All right, we definitely have time for questions afterwards, but does anyone have any questions since this is like a natural breaking point.

-What did you call the slide where you had total user based consideration evaluation compared immediately addressable audience.

What was this system called? It's, it's a.

I don't remember.

I can actually find the exact name for you because there is a proper term for it.

I've been referring to it as a user model, sorry.

-I just wrote down loyalty loop and I wasn't sure if that was actually the right thing.

So, there's one overall.

That's two concentric circles and that's a Harvard Business decision, customer decision model.

And then there's the other one which is the user model.

But there's actually a name for that, you can definitely dig that up.

I learned it because I went and bought all of the research books in the library for the design, the UX and design courses here and it was in one of them.

Knowledge is power.

[Laughs] So, alright, the next step then is planning.

I have templates and stuff for you, but I want to go over this idea of training your user base.

So, I think that.

Okay, virality.

When I say like let's make this viral a lot of people laugh.

It's a term that got really dragged through the mud, but I think there's a lot of merit there.

I think we just used it wrong and there's a bad mentality around it.

We had assumed, we looked at examples of.

Remember the guy who took an iPhone and shredded it in a blender? Within like a day millions and millions of views and we considered that viral.

And yes, it's viral growth but I don't think that's a standard use case.

It is like super kitschy.

Great, but I equate it more so with actually getting a virus.

So, the flu has been rampant lately.

And if, let's say, Sarah has a flu and she coughs in my mouth.

I don't immediately start vomiting, right? It incubates.

And I think that's the missing part of the perception is that to make social media viral and a hit by reality it needs an incubation period.

That's, that's the reason why I just gave that example.

And we only associated virality with viruses that were immediate.

And people would give up.

And they didn't really understand that this planning and the execution, so I'll go back one slide, the execution for social is a lot of sweat equity.

It's like you are on all the time even if you have paid.

Because paid is less and less viable over time as it's being oversaturated with other advertisers.

So, let's say and assume there needs to be an incubation period here.

What does that actually look like? So, take everything from section 1.

We have the way we want users to value us and be trained.

So, older example is throwback Thursdays.

When throwback Thursdays first started we were like hey, this is kind of cool.

And then advertisers and marketers started planning weeks ahead.

It was like hey this thing is going to happen every Thursday.

We should really go do it.

And now there's a bunch of other, this is where the key word is thematic, there's a whole bunch of other thematic ways to engage with users.

And I think that's regardless of all the algorithmic changes and so forth.

You can apply any kind of digital channel and so long as you are delivering the right content to the right people when they expect, it's great.

It's like no one really likes to be surprised on digital anymore.

Usually at this point if it's surprising it's something that has gone horribly wrong in the world.

Instead we have this virality which is mass times velocity.

Throwback Thursday is an easier one to show and look at.

But, let's say, it's email marketing.

We have a newsletter that goes out every week on the same day at Kickstarter called Projects We Love.

And people actually love opening and looking at it.

And they know when it's going to arrive and at what time.

And then let's look at the skim.

The skim sends their digests at the same time to the same users every single day.

You know when it's going to happen and they know when you're going to read it.

And one of my favorite examples is actually Jimmy Fallon "Thank You Notes.

" Jimmy Fallon is so talented.

So, we, that's like we don't don't all have that already.

But he does this bit every single Friday called "Thank You Notes.

" And there his users know that it's going to happen.

And it ends up being trending immediately across multiple social media channels.

So, this social media virality is great because it takes time to build, but during that incubation period it is actually compounded growth every time because the same users are participating.

As more users participate it starts trending on the platforms locally, nationally, globally, and so forth.

So, you're actually just taking the same amount of resources, same amount of effort every single week, and it grows.

I mean, it's like the magic formula.

You do need the right content to the right people and use the frameworks in the user journey model.

It's actually virality I think is fairly simple it just takes time and repository of content and an understanding of your users.

But the formula is basic.

Get them do that.


So, I just I recently did a meet-up for the.

I always forget, but it's like the the group of formal journalists.

And it was like 150 of us.

We did an experiment where we did #ShriTips.

Shri is actually the former Chief Oigital Officer of the Met.

And that's where I got that example.

And we all used the same hashtag at the same time and it became number one trending in New York immediately.

And it was like, oh, so, a hundred fifty people could do this right now.

And like, it was a great proof point that it's not like, it's not this fallacy of virality where it's like, oh, now we need like a million people to watch this right now.

If you kind of just jumpstart it.

So, we're doing stuff like that at Kickstarter as well because if you want to jump start it and use that ,you can always pay to jump start it.

I had not been a fan of boosting posts before.

But it's something that I've grown to like because to jump start a post if we think it's great content, it's going to the right people, and we can really get it viral, then why not throw fifteen, twenty dollars behind it.

It's, these are things that over time aren't going to break your budget.

So, what we're doing overall is that we are playing with the algorithms.

We know that there is just this velocity, there's this velocity formula we need to play with.

So, every morning I just boost a post for ten dollars, and we've already seen the compounded growth effects.

So again, it's super simple.

And it's just trying new things and testing.

And at least for now it's still fairly cheap.

My assumption is like with everything marketers will eventually ruin it when we all start doing it.

So, get on it now while you can.

Alright so the next step is more so the how to, how do you then arrange your strategy around this? So, we've definitely thought and ideated or hopefully you've ideated ways that you should slice your users, your content, and structure your KPI, your key performance indicators.

Here are tools that might be super useful for you.

First, listening tools.

I use all like most of these for my personal brand, let alone like for organizations.

And a lot of these are freemium.

So, you can listening, well, just show who is mentioning you and give you individual feeds based on who's mentioning you, any key words, or hashtags you want to monitor as well.

It's really popular among people who use customer support for social media just as a great way to see like is this something I need to address right now.

I just sit there and hit refresh until someone actually messages me.

So, mention is Mention, HootSuite, Sprout Social.

These are all freemium products where you can start using it, but then once you hit a data cap or you want additional features they start, you have to pay for it.

Since this is like a workshop for startups I'm mentioning all the freemium stuff.

I have list of tools that do it all.

These are tools that are like thousands of dollars every single month to use and pay for.

But they'll help you with not only listing, but publishing too.

So, scheduling these posts.

Let's say, you know the thematic content you want and your red bow.

We did the documentary about the BMX biker and the musicians, and all that kind of stuff.

Because BMX biker, I don't know how to ride one.

And we were like alright, we already know which channels you should go to, at what times, and all that kind of stuff.

Save yourself the hassle, schedule them.

I'm in a publishing tool.

Then measuring.

We I didn't put it into this deck, but if you're interested I also used to teach the Google Analytics lesson, and it has whole section about things called UTM codes.

When you share things to social media, it doesn't, and people click to your site, it'll show that you get traffic from social, but it doesn't necessarily differentiate your efforts from everyone else's effort.

And you need to use codes within query strings to actually track this.

It's actually fairly simplistic, but I'm just giving the TLDR, so you know that it exists.

It's very important.

And then I can send you additional information afterwards.

Here's a use case.

I share something to Kickstarter's Facebook.

And you're like hey, that's super awesome.

I'm gonna share that to my followers.

And a whole bunch of people click from yours and mine.

And let's say, we have ten transactions.

How do I know which ones came from your post or my post because it all comes from Facebook? You can add keywords to the URL that really you're tagging it, so you know that it's yours versus theirs.

And that opens up capabilities in Google Analytics and property measurement tools.

So, Google Analytics isn't the only one and step 4 of this, we are in step 2 of 4, is about measurement tool, so, I'll go further into it.

But Google Analytics and these web properties are super impactful, but you do need to tag them yourselves.

But you always have the native on analytics platform.

It's Facebook has Facebook Insights, Twitter has its analytics platform.

And they'll give you a whole bunch of really great insights.

It's just not necessarily connected to everything happening on your website.

So, you lose part of the user journey.

Alright, so next.

Putting it all together.

I'm gonna hop out of the deck for a second, but there are things called content calendars and editorial calendars.

It's a way to organize what you post.

And just, it's like a lean marketing strategy.

So, I have a link here to templates for you.

Again, it's in the deck, but feel free to follow along.

I just made a bitly link and that is a zero not an O bit Li.

John Chiang 0 9.

So, I'm gonna hop onto the deck real quick and show it to you wherever I put it.

All right, here we go.

So, there's three templates in here, and they're meant to also scaffold just like this lesson.

The first tag on the first tab is a way for you to assess the content that you have right now and whether there's any gaps.

So, on the left we see all the stages of the user funnel.

We then see content types that we think would work for our users.

I give you the kitchen sink.

No one really needs all these types of content anyway.

So plug and play what works for you.

And then for these content types what content do you already have? If you don't, what do you need to create? To whom does it go to? Or who is it for? Because your audiences are diverse, just, and unique, just like we are.

And then whether or not in gated.

Here is whether or not they need to exchange information or money to receive it.

So, is it free or is it free with an email address? Huge difference there.

I think of content marketing in this kind of way as bargaining with the user.

Like, I'm like, hey, you should all give me your email address.

And you're like no, I don't really want to do that.

I'm like, what if I will only give you this lesson deck if you give me your email address.

Oh, you're like okay, so, this deck is worth that.

Everyone has a way to quantify the worth of their email address and their inbox.

So, we need to figure out whether or not it's gated.


This is something that's again more so for content marketing, but it helps you understand.

For the content are you addressing all the KPIs necessary? So, going back to that one minute exercise.

You've already talked about the top, we've talked about your value propositions, how that fits into your social media strategy.

Next step, do you have the content that's necessary for the distribution? So, step two, tab number two allows you to not only say alright, we have different types of content for people, but does it achieve the objective? Because at the end of the day we need to achieve the objective here.

The third one is the actual social calendar.

So based on the content name you have the date that you post it, the time that you'd want to post it.

The time is not necessarily because users would read it at that time.

One of my favorite tweets is Childish Gambino asked everyone if there was one historical figure you'd like to revive, who would it be? And the top response was Instagram's chronological feed.

[Laughter] So yeah, if these feeds are like no longer chronological, but the post time just helps you understand your organizational process.

UTM's here, so this is your tracking.

It's noting the tracking.

So, when you do go and measure it, all you have to do is input it to see the results.

And then I have just different channels broken down for the content.

And then this links to any folders, whatever you want.

Again though, this is made for more so startups.

If you use Sprinkly or SpreadFast they have their own CMS's, content management systems, so you can actually calendar and everything within there.

This is more so for either just a thought exercise or if you don't have thousands of dollars to spend on this every single month.

Yeah, if you don't fit into those two use cases this is great for you.

All right, so we do not have an activity specifically for this, but I do have the templates for you to play around with and enjoy however you'd like.

So, that's phase two.

We did research and planning, now about the sweat equity and always being on.

I love this.

I stole it from one of my friends who's an account exec at Twitter.

She came to speak to one of my general assembly classes and used this information.

I think they use it in their like pitch decks.

So, Twitter and I think all of social really can be categorized as live, public, conversational, distributed.

Well distributed, per say.

The live part is something that we haven't talked as much about.

But there are.

I think that if you aren't doing a live content then you're missing out because it doesn't cost any additional money.

And really like if you, if anyone tells me I can't make content because I don't have the equipment.

I'm gonna call BS on that because this takes better video than my SLR.

This iPhone 7 plus takes, yeah, product placement.

It takes 4k video.

So, a lot of stuff that I post and, I post about is actually just taken on a device anyway.

So, live video is great because you can then extend the reach, but it still exists in your profile and on your account.

So, you can boost it and reuse it at any time.

This really started in like 2013.

Facebook did a video a live video with Condoleezza Rice.

So, that was the first news case.

And they did a Q&A with her.

So, it goes into the third pillar which is conversational.

And people interacted back-and-forth with her.

And then every single week after that they put a little bit of money to boost it.

And they had their community manager be on and respond to comments as it went.

Because the content didn't get stale it was great in live at the time, but people still wanted to interact with it.

And it doesn't require the same big lift as it did initially.

They're getting a lot of mileage out of it.

And so, I'm seeing a lot of head shakes.

So, choir preaching.

I get it.

Okay, great.

[Laughs] And then that brings us to conversational.

A while ago Reddit invented this thing called AMAs.

They're Asking Me Anything's.

And like Bill Gates, he was one of the first people to do it, and Obama has done a few at this point.

But it's where someone sits on social media and others can ask this person questions and receive answers in real time.

So, we do this with our creators as well.

And I've done it in other organizations.

And Twitter, they're just called Twitter Chats.

And a key strategy for when I was at MakerBot and we were trying to get k-12 educators.

There's a hashtag called #edchat.

Every Tuesday afternoon k-12 educators were asking each other for advice about teaching math and science in the classroom.

I was like wow, this is great.

Hundreds of thousands of people that can be addressed right now with write content and the theme, the Khan, and the virality is already set up to go.

So, having a place and understanding, combining this with the thematic, giving people the voice, and having conversations back and forth.

If we want to, just like, this is not necessarily an edge case.

But it's great for customer service, too.

I can bet that most organizations you receive customer service tickets that can be prevented through proper education.

So, these live chats can definitely be great for that, too.

Like, you give your community manager set of macros or a set of common responses that you would just email the people anyway.

And instead you allow them to answer these questions live with users on social media.

So, there's so many great use cases here.

I want you to follow along for this activity.

So ,we know there's four pillars right now: live, public, distributed, and conversational.

I want you while we go through this use case to think of one example, to note one example from KLM.

So, KLM, the airline, did this brand campaign a while ago, but I think it's still super relevant.

And they do for these four pillars very well.

So, it's the "Happy to help" campaign.

It was not targeted at the US.

So, if you hadn't heard of it then it's probably because you were living here at the time and not traveling in Europe.

So, they did this happy to help campaign where they targeted anyone upset with airlines and then gave them a positive customer experience.

That was it.

Simple, premise, really-really great.

So, they looked at travel conversations to figure out when they should activate the campaign.

When are people not only flying the most, but what are they complaining the most? Because you can bet like, alright, maybe I would have thought more so in the winter because there's a greater chance of delays cancellations and so forth, but this is great.

The data and validates my assumption That's what data is there for.

Prove me wrong.

So, then they researched.

We have the places where they did the research noted here, by the way.

What the major keywords were? Because they still need to find the conversation.

So, they know the conversations happen.

Then you need to find the conversations here.

So, they have the key words, they have the timing, they're getting ready to go.

They do the rest of this stuff: targeting, content, and so forth.

So, this is the actual branded campaign.

Again, four pillars.

Note one example of each, please.

[Music] We are proud of our customer service.

[Music] But you have to fly KLM to experience it.

That's why for one week we decided to help out everyone, even people who don't fly KLM.

To do this we set up a special 24/7 #HappyToHelp control center at the Amsterdam Airport Schiphol where we track down travelers who are in need of help.

Of course we found many people with help from our staff all around the world, but most by constantly scanning Twitter, hundreds of travel related search words.

Most problems could be solved in a tweet, but some needed a bit more attention.

We had a speedboat taxi ready in the Hudson River to beat New York traffic.

-Yeah! It's a lot better than being stuck in traffic.

A car in Amsterdam to pick up passports.

And a motorbike in Hong Kong to track down forgotten items.

We gave Eva a wake-up call for her early flight.

-Good morning Eva, it's your wake-up call.

Provided a comfy bed for passengers who got stuck at the airport overnight.

A speed tour of Amsterdam for people who had a couple of hours layover.

Sang a lullaby for Sean who couldn't sleep.

[the woman is singing] Honeymoon transfer to the gate for Tom and Kelly.

And got Jill in touch with her boy band crush to get her through a delay.

-Oh my God! And much, much more.

[Screaming of joy] -You're going to be a flight attendant.

It's been amazing.

We helped out thousands of people flying other airlines all over the world with millions following what we did online.

So, to everyone who doesn't fly KLM.

Have a safe trip! And to everyone who does, remember, we're always #Happy to help.

If we all had a budget to do something like that.

That's not the point.

It is to show you what you can do.

One actual item is just figuring out when people are complaining about things.

We call it a con question campaign.

I don't, I think it's kind of like shady to do it really aggressively.

But there's two examples that I think are really practical.

One, I was complaining about my Verizon service because it was really spotty.

And so, I was like that Verizon service is really spotty.

This is like a New York thing.

Is there anything I can do to do it? Sprint responded first.

And it didn't make me convert to a Sprint user, but I was like, wow.

This is really great.

And it didn't actually seemed like it was an automated reply either.

It probably was, to be honest.

That's how it was so quick.

But I was like, this is really nice to have a response from someone.

You like really care about me.

The second thing is that in my social media marketing class here at NYU we did a similar kind of exercise over an entire lesson.

And one of the students.

So, their use case for the semester is that they are trying to market a fictional headphone company.

So, they were looking competitors and one soon was like, wait.

We took a look at all of boses social and after 5 p.


Eastern time they stopped responding to customer service.

Customer service like Tweets.

I was like alright, well, let's find a way to tactfully do this.

But that's, like and they start again at like 10 a.


the next day.

That's a huge gap.

Like, so what is the opportunity that you can activate there? And also they don't do Sundays from what we were able to see.

So, it's stuff like that where I don't think it's like, it's not like calling out your competitors.

But when there's a user need and it's not being satisfied what can you do to support it? And I bet that if you are a headphone company, which I do not mark it, like that's where, that's one of the ways I would start.

Maybe that's when you would host your Hashtag Twitter Chats, maybe and so forth.

So, two actual things afterwards.

People love behind-the-scenes content.

There was a great lift.

And let's say you're like yeah, John, but who cares about brand lift? Every single one of these people who participated can be retargeted.

And now you have information about the user intent and how closely they are aligned with a brand.

So, it asked prospect quality.

You now have measurable prospect quality with millions and millions of people that you can retarget with different content, different CTAs, and to achieve your different objectives.

And I think that's another thing that's missing.

Social does get a bad rep and it's like, why would we pay for awareness? Well, we don't, you don't really have that much information about a users.

Or maybe you've taken just straight-up ads as far as you can go.

So, let's use organic to like really delight them and get them further through that funnel.

Alright, so how is this live? I know these are all like no brainers, because they explain them pretty well.

But who wants to recap for me? Not sure if we necessarily need the microphone and for this one, but.


-It's live because, is this working? It was just recording, okay.

That'll be interesting to have on a recording.

-It's live because they had the team in the office -constantly searching through the tweets as they were being posted.


-I would say it's probably because it was all being done in airports.


-Someone else will have to pick that one up.

And then -conversations was because the folks who they targeted and the individual acts were probably tweeting about what happened after the fact.


Anyone wants to pick up distributed? Yeah, over there.

Also, right on time.

-I think the distributed and the conversations kind of go together.

Like -with the happy to help hashtag was an easy way to distribute.

-But then also like they had things, they alluded to -things within the spot of like, I don't know if it actually was projected up on a wall of a building.

-But like, they had examples of how they were distributing it live.


So, there's long tail statistics.

There's a great book by Chris Anderson called The Long Tail.

Essentially it just says there is a curve where the tail never hits zero.

And one of the main examples he uses is Barnes & Noble versus Amazon.

So, does anyone know how many titles a standard Barnes & Noble carries, like the big ones? They have like a cafe and all that kind of stuff inside.

It's like a hundred, he says.

So, I'm not an expert at books, this kind of stuff.

That makes me sound really dumb.

I'm totally literate, by the way.

So, it's like 110 to 150 thousand titles.

And oftentimes these are just the best-selling books.

What he talks about and what he broke out with in this book is that Amazon makes, oh my gosh, Amazon sells more books and games more revenue than Barnes Noble just by looking at the books that sell once per month.

The cumulative amount of revenue buy all the books that sell once per month are greater than all the things that Barns and Noble sells for the top 100 thousand titles.

So, it's just an example, a concrete example of this long tail.

So, for distribution, and bringing ii back, and campaigns like this it doesn't matter if this person has like 20 followers because you're trying to make a campaign that reaches a whole bunch of people to really build, and so, I'm gonna reuse this term again, but like compounded growth here.

One of my favorite campaigns is actually Purina.

They used to do pimp my dog.

And they would find people posting pictures of their dog, download it, draw on top of it and repost it.

And it was super delightful.

And it didn't take them very long to do, it's like just sketches.

At this point you don't need to download it.

You can just like do it right on there and repost it.

So, there's all these different ways that you can address that long tail of users.

And I think that this is a campaign that did it fairly brilliantly all by with gigantic budget.

All right, so here's what I'd like you to do for a follow on.

Please, go to Facebook.


We are going to set up a campaign together over the next 20 minutes.

And then we'll have time for Q&A.

You do not need a business account to set up ads manager.

I think some of you might end up being prompted to do that anyway, just ignore it.

And Facebook does not ask for your payment information upfront.

So, you don't need to actually do that.

If you are in mobile this might be slightly tricky for you.

But I think you can do it.

I have faith.

You seem technically savvy.

And then I have a prompt for you, and then we'll walk through this together, but I'm going to show you the prompt first.

slash ads, slash manager.

So, we now are all going to work for a fictional travel organization.

And the prompt is that I want you to use Facebook ads manager to make a target audience.

I'm using all of their cool functionality.

So, we are trying to sell an upcoming trip to London to young singles.

A lot of this language is intentionally ambiguous because it's meant for you to interpret beautifully.

So, we're headquartered in New York, but they're leaving from SF, Seattle, and Austin, and it's not cheap.

That's, those are the criteria.

That is what you have to work with.

So, 50,000 isn't necessarily the golden number, but it's a healthy number for an exercise.

So, you can learn how these different variables and features within Facebook ads work together to increase or decrease the overall target size.

So, let's do that together and solidify this knowledge.

All right, so I'm just using my personal account because probably not great to show you my Kickstarter, there is actual data.

So, if you just signed in you'll be sent to this screen by default.

If you did not just sign in do what I did and hit the green button that says create.

Facebook and every ad platform is objective first.

And if you weren't following on that's totally fine.

You're, you can work with a partner, you can ideate this as well.

That's why I chose this.

You don't necessarily need to know it all.

Facebook also has, actually some organization published the targeting parameters and features of Facebook, so, I didn't do that.

I didn't link to it, but I can send it out.

So, alright, if you're on the screen AdWords does this now, Pinterest, Twitter, Reddit.

It makes you choose your objective first.

Why are you choosing to spend money? Is it brand awareness? Reach? Are you trying to do this consideration ad? Hey, look, it fits right here.

Or is it for a conversion? There's some really cool things here.

For, let's say, engagement it also allows you, I'm going to click back and forth, you don't necessarily need to follow this, but you can just get event responses.

So, choose events that you you want to promote.

So, this helps with the placement and creative types.

This is what I'm saying about like email has a single placement It's like you have one inbox and that's about it.

Whereas with these ads, and we're already diversifying on this because we can do event responses, there's something called lead generation here where people can give you the email address without leaving the Facebook experience.

You can do messages, and so forth, and so forth.

I'm just gonna choose traffic because, it's in the middle, top of middle, there's no particular reason.

So, choose whatever you, whatever you want.

We're going to have the same basic features.

Enough to complete this exercise.

So, let's dissect this.

We have this screen after you choose your objective, and it spits out three columns.

Left column just tells you where you are, so you don't lose yourself.

Middle is where you actually input information.

And the right shows the reach and the impact.

So, on the right there's two modules.

The first module is potential reach.

And that means based on the targeting criteria, how many total people actually match this? On the, right below it there's reach for daily results.

That is dictated by your budget.

The more money you are willing to spend, the higher the reach will be.

So, we only want to look at one of those two.

We want to look at the top one.

You want the potential reach to be approximately 50,000.

This is just a number that I made up.

It's not like a Facebook guided number or anything.

I just find it helpful for education.

So, as we scroll down the wedding cake, bottom of the wedding cake: age, gender, location, we've got demo and geo.

So, right here we can choose like New York City as a DMA, or a demarcated marketing area.

We can do radius targeting, zip code targeting.

For an AdWords you can target by borough in New York which is awesome.

Age, language, and so forth, whatever fits you best.

Then we have behavior and psycho which is right here.

It's called detail targeting.

So, this is the second layer of the wedding cake and why we look at the wedding cake first.

All right, let's just look at actually fitness.

Fitness is a huge one.

So, I just typed in fitness in an auto populated, the word fitness over and over again.

But it notes whether it's an interest, a behavior, an employer, a job title, and so forth.

So, there's multiple ways to target people under the fitness category.

The main one I want you to look at is behavior.

It is because in the description this is actual consumers in households that are known for likely buyers of fitness based on actual purchases.

And it shows you the data.

It's supposed to show you the data sort, yeah, here we go.

And this is data provided by Oracle data cloud.

So, you can differentiate the quality, perhaps, of the targeting data used here.

So, this is just fitness.

Now think about everything else.

If this is new to you, mind exploding.

So, what I would like you to do now going into the exercise is actually just feel free to type things out.

But when you hit browse on the right it separates out demo interest behavior in more categories.

I would like you to explore behavior and learn what's available there.

Because it's not until you learn what you have access to you that you can fully use the features, segment your audience, and put in all the rest of these things.

So, if we look at purchase behavior.

Again, you don't have to be following along exactly with this right now.

I just want you to know where it exists because it's gonna, this is gonna be your own choose-your-own-adventure.

Buyer profiles.

Look at that.

Based on your financial information you have an affinity group.

Information from the same third-party data sources.

So, on Facebook and Google Analytics, and AdWords, and so forth you'll see the same nomenclature.

So, gadget enthusiasts dream living healthy and fit, and so forth.

I love that they have the data sources and all that kind of stuff, too.

So, oh, pet products, that's me, dog owners, okay.

As you choose these Facebook has "and" and "or" functions.

So, it gets kind of hairy here because based on what you choose you can see this drop to 1.


And then if I choose a language it ought to drop even further.

So, it's 1.


However, if I add in additional dog product, should have gone back there, add in an additional behavior.


It's just not significant enough.

Let's say fitness as behavior.

It increases because it's an "or" function.

So, if you add multiple variables within the same category it expands the audience size.

Or if you yeah, so, add variables across different categories and then it decreases.

So, this is I think it's just a good use case to figure out because it's like oh, all of a sudden your impressions are increasing, but your click-through rate is decreasing.

Look at the potential size.

Perhaps, that you have all the right ways to target people, but they're layered on in the incorrect ways.

This is specifically a Facebook Ads problem.

But the read we look through Facebook Ads is because it is a really good representation of the other platforms.

Facebook was first to market into this stuff.

It has arguably the most powerful technology.

And if you go into Twitter ads, Pinterest ads the UX, the UI and the features and the data sources are very similar.

So, even when I teach week over Facebook in depth and then students can really navigate and find their ways and become experts on all these other channels themselves by taking this information and applying it to the other channels.

So, that's also what I recommend to you.

So, you have detailed targeting here which is the middle of the wedding cake.

And then look, you have the top of the wedding cake – placements and devices.

I'm padding my own back for scaffolding properly.

So, Facebook will default to automatic placements.

There's different mentalities around whether or not you should do auto placements or edit the placements, and then the camp where you should edit your placements because look at what it would auto to.

It automatically places you on all Facebook, all Instagram, all audience network, all messenger and to me people, people are different on Facebook and Instagram already.

Then different types of people interact on Instagram stories and the feed differently.

And in addition to that based on where you want to target the placement you probably want to use different creatives.

You don't want to just use one generalized piece of creative and blast it out.

Instead because like, I showed you the category specific videos, those might be served best on different platforms.

And on Instagram I might want to do a 12 second cut which we actually do.

Where as on Facebook on desktop will target and do a 42 second cut, which we actually do.

So, yeah, human attention span of about seven seconds, I don't know.

On Instagram the average video duration length is 2.

4 seconds.

So, it's way lower than that.

So, we're looking at, all right, thanks Facebook, but not so great for me.

So, choose where you think this is best.

Right column.

So, this is going to kind of contradict what I just said, but in the spirit of not over refining and prematurely optimizing data scientist Donald Knuth famously said and I'm quoting another famous data scientist that "premature optimization is the root of all evil.

" So, if you, if these placements are all new to you, you don't necessarily know what will and will not work well.

And maybe throw some dollars and try it all now and just see what kind of sticks.

So, that's number one.

I'm just gonna deselect Instagram, deselect audience network, deselect messenger, and just let Facebook be itself.

So, that leaves one last part of the wedding cake and then I'll take questions.

And we'll do the activity and there'll be a good time for me to take questions.

Here it has oh, this is interesting.

Why did it do that? Feeds, instant articles, suggestions.

So, then it has devices.

So, oh, I see why.

I'm gonna take these out, just do the feed and see what happens.

Well, this is weird and I'm like hey, let me show you this thing works.

And then it's like hey, this thing's not working the way I wanted you to.

So, here you can choose the device and placement it already breaks down, Android, iOS, feature phones and there's desktop usually.

A great feature that I love using is only when connected to Wi-Fi.

Because it not only tells me that they have a proper connection, but it tells me that they're probably, they're probably stationary.

This is great for video content especially because a reason why people turn from watching videos and long-form content is because they are in transit.

And you can make, you can make arguments for both, but this is where the placements and the devices go much further than I had alluded to before.

There's Wi-Fi on Android, not Wi-Fi and Android, and further and further down.

All right.

So, that is enough to complete this exercise.

I would like you to play around with it, get a potential audience reach of 50,000 and play with whatever you want.

Hint-hint it does do you household income if you want and you can splice and dices however you want.

You do not need to target every location if you want.

This is the fun of it and why it's an ambiguous prompt.

Please, raise your hand if you have questions.

I'll definitely walk around.

I know we have at least one question in the audience.

So, a couple follow-up things.

First is that this is not necessarily specific to Facebook.

You can do this as you learn new channels and I encourage you to.

This is actually what I give to new employees on my team.

I'm like, hey, go through Facebook and make an audience.

I should actually make it an interview question or interview homework.

Alright, idea, great.

[Laughs] Moving forward at that.

I had a question about advanced segmentation, which is the last step of this.

In Facebook, in Facebook and Twitter you have things called custom audiences and tailored and look-alike audiences.

And we'll look at it together here.

So, it's like this little thing right here.

I'm gonna go through the process really quickly and I'll give you the TLDR.

So, custom audiences are people that you intent like, individual people you intentionally want to target.

And you can do it.

The most common use case is you have a list of email addresses, you upload them and Facebook matches real Facebook users with these email addresses.

So, you can just target them.

That's the super duper manual way.

We automate it at Kickstarter, so we have an adequate automation platform called Simon Data.

And what it does is it automatically updates all of our custom audiences and suppression audiences as well.

Because if people are participating we don't want to continue showing them ads.

So, you have suppression audiences and active audiences for these things.

So, this is the kind of retargeting that I was talking about.

This can be retargeting, it can be one-time engagements.

So, for website traffic I can say website visitors to a cart.

So, I want to do website visitors who went to the cart, but did not get to the Thank You page.

Let's say, these are the actual URLs.

This is now an abandoned cart campaign.

Every single person on the website who added something in the cart, but, or went to the cart, decent difference there, but did not hit the thank you confirmation page should get an ad about coming back.

We're in the prospect quality and conversion probability is this person.

Far right like, this like, this took us like two seconds to do together.

So, that's one example.

Another example, if we create a new audience we can do the customer file, but also engagement.

So, let's say, that this is what I did.

Someone who viewed three seconds of the video 10-25% and so forth because it aligns with higher prospect quality, not necessarily conversion probability, but we we will get them there.

We know what the next step is.

And then you have things called look-alike audiences.

So, this is super cool.

Both of these, by the way, come with premiums.

You will be charged more, but it's because Facebook knows it's so valuable, it is really valuable.

So, let's say we have this custom audience of people who have bought our products.

We upload all the email addresses of people who have bought this one specific product.

Once Facebook matches them, they will then create a whole new audience that looks like, really behaves like the audience that you're pointing to.

So, let's say again, like, I'm interested in the NFL, dogs, and like Prada.

If we'll find everyone else who is interested in those three things and make an audience of them.

It's like a really boiled down example.

And which is great though because the conversion rate tends to be higher here.

And even though the prospect quality is lower we are assuming that we can get them through that funnel a lot faster.

And we can get them to identify with a brand faster because they behave like people who currently do.

So, this sliding scale is the degree in which they look like your custom audiences.

1%, so this is not the exact map, but let's say 1% is 3 common characteristics, 2% is 4 common characteristics, and yeah, and so forth and so forth.

So, it ends up expanding over time.

These are common tests by the way.

So, AB test, multivariate test.

I'll test 1% look-alike versus a 3% look-alike versus 5% look-alike because I don't necessary need.

There is a lot of magic happening behind here.

So, I don't necessarily know how closely aligned these people are.

So, if you already use a look-alike audience I do encourage you to for the same custom based make a 1%, a 3% and like adjust them from there, and test to see where you find diminishing return or where the conversion rate is lower or higher.

So, those are examples of advanced segments that are totally actionable.

So, this last segment is just looking at screenshots of [Laughs] reporting tools.

But first, I really don't watch a lot of South Park, but I love this example.

South Park has this episode where the boys find out all the underpants are being stolen from the town.

So, they go to an underground cavern to find, its a bunch of gnomes.

And they ask the gnomes for their business model and the gnomes show them this.

Step one is collect Underpants.

Step three is profit.

And they just keep repeating this every time.

The boys are like what? Like, how are you gonna make money? And I feel like this is what's missing.

Especially because I've harped a lot about like brand awareness is really important because we can retarget and all that kind of stuff.

But it's incredibly difficult to measure it across all digital for two main reasons.

One, we have cross device attribution losses and cross channel attribution losses.

As a primary example, if you are using Google Analytics and you are looking at your web data.

If I visit from chrome on my phone and I visit from Chrome on my desktop I would be counted as two individual unique users because the cookie has dropped on a browser, on the browser.

In addition to that from each individual channel you don't necessarily know it's the same user if someone's converting from Twitter and Facebook.

I mean, there's definitely ways where we strengthen the accuracy and the attribution reporting here.

There's inherent flaws.

Which means, we got to use multiple tools to properly measure.

So, there's native platforms I talked about within Instagram, within Facebook, within Twitter where you get great information about your users and performance.

What I recommend is you supplement that with some sort of property measurement tool such as Google Analytics, Mixpanel, Chartbeat, Adobe Analytics, Heap.

They for the most part do the same things on the free level, so native platforms, property analytics, and that's actually about it.

I recommend you learn SQL and then connect the individual API's of these channels to a data aggregator for a centralized database.

What does that actually mean? Because your data is distributed across multiple channels what you can do is through coding get that information into a central repository that is the database that you own and then visualize it with the tool.

We use Looker at.

We use Looker at Kickstarter.

I also use Looker here at General Assembly and used Looker at Stack Overflow.

So, it's a really popular tool.

All it does is it sits on top of your database and allows you to run SQL queries to populate tables and visualizations.

So, that is definitely kind of the future that we're heading into because there's proprietary data everywhere and you need access to it in multiple places.

So, native analytics and Facebook Ads, native analytics and Twitter, and then a screenshot of where to find social media network referrals within Google Analytics.

And that's all I have.

So, yay, 15 minutes exactly.

[Applause] Does anyone have questions? Thank you for this.

-Thank you John, that was fantastic.

-Yeah, thanks for.

-We have time for, you know, 10 minutes of Q&A.

Might one ask, just raise your hand and I will give you the mic so we can record your question for the film.

-Hello again, how would you streamline this for the recruitment of Alpha and Beta testers because it's a unique market.

I love that, yeah.

I mean, it's really just top of the funnel.

So, I have mentored at Chobani and TechStars.

For a lot of companies I've recently been injected with some cash.

And the first thing I always look at is whether the funnel is leaky.

So, with alpha and beta testers, for example, you, let's say, I mean it's a really great way to expand the overall pool of people to email.

So, the funnel ought to be where do you want to collect information and collect alpha and beta testers? And then what happens afterwards? Because not all of them are going to be quality testers.

So, you need to have a characteristic identifier.

So, perhaps you collect email address along with like current profession.

Maybe that doesn't work for you, but it's something like that.

And then based on that make an email drip campaign.

You can do this for free in MailChimp where an email drip campaign is a predefined set of messages users would receive by email.

And as those people go through it's usually like confirmation is number one, valley props to the organization is number two, products is number three, product overview, number four is like a user story or brand property like Brandon Stillman, number five is like a call to action.

So, based on how they perform in that you get that data on a user level and you can then inject those email addresses into paid social.

So, that's one way to stream it.

But I think the most important part really is just lead generation.

You're getting all these great people and you want to activate those you want, but you also want to keep them.

You want to be top of mind for all those that won't be.

And ideally it's like you have a million email addresses and you can just pick and choose a 10% that actually participate.

I mean, it will require maybe one, two, three landing pages at most.


-When you're making, starting a startup,and you have like a limited hiring budget -who do you prioritize to hire like to do some of this work that you can't do yourself? It depends.

So, a lot of my answers are it depends to two things, actually.

So, I have a link to the growth hackers conference in San Diego that just happened.

And so, I spoke there in a panel, but more importantly Joanna Lord spoke there.

She's a CMO of ClassPass and they published her slides.

She did a 25-minute talk about hiring for growth.

And it's the answer to this question much better than what I'm going to answer.

But it really depends on what you need.

So, for some organizations I'm like you can probably go on Fiverr and have someone set up like analytics for you.

Whereas for other organizations I was like, stay away from that because they were gonna screw it up and then you're going to have to like fix it at a greater cost later.

So, I would I mean, it depends on the level of operation as well as length and complexity of the funnel.

For like b2b it tends to be much simpler.

So, it's like great.

You can just do basic templated set up for ECOM as well.

But for like most things in between or if you're trying to do like a content or an engagement model you have to work a little bit more outside the box.

The caveat or the exception of the rule is whether or not you're using a website builder like Shopify, Squarespace, Wix.

It completely simplifies your life if you're you want to be a small to medium size business which is what they're built for.

Once you scale past like a million users or something you're going to need to either upgrade or just build your own thing.

So, those are the variables I would use to give you the proper answer.

Definitely look at Joanna Lord's slides.

You can just look at like growth hackers conferences, a blog post that has all other slides.

The be VP of growth at Eventbrite also has like a really great one about like how he made growth at Eventbrite.

"made growth" Yeah -So, where could, where else could we go to like start learning like more of these processes, like the more intricate details? -We like went over an example of Facebook, but like the more the multi-channel stuff -for like a cheap price point.


-I know generates General Assemblies is like pretty expensive.


So, the education on this kind of stuff is spotty.

Not necessarily because the subject matter experts and all the kind of stuff don't exist, but because it changes so often and it's very cost it's not cost-efficient to change your curriculum so much.

Overall, okay, so there's like a stomach issue in education.

So, there's some Udemy classes that are great, and it's because they can operate fairly nimbly.

I don't want to talk poorly about other work, so I'll just talk about the good ones.

So, Coursera I think has some good ones.

Udemy, I really like Udemy.

I would look at the published date of course to see if it's published within the last year because or around the biggest algorithmic changes.

And then what I did is that I used to bike from Ridgewood to Chelsea every day for work because I used that hundred dollars that I would have spent on a MetroCard to test ad channels.

So, do that I would say.

Like, there's no better way than experience it yourself because I mean, most of this literature either already exists or it has yet to exist, so, you can go write it.

That's my standard answer.

Yeah If you want go advertise my dog's website.

She needs it.

I'll give you twenty bucks, I will do that.

But seriously it is a really good use case.

She's great.

-Any other questions? Alright, thanks for coming! This is great.

I mean, you all are great here.


Source: Youtube