Chapter 2

How to Create Sticky Growth By Starting at The Top of The Funnel

Retention is often thought of in the middle or bottom of the funnel, only after you’ve successfully acquired a user.

But in reality, the retention process starts much sooner, when people first interact with a product.

A high percentage of churning users doesn’t necessarily mean your drip email campaign isn’t working. It can also mean that the people who were brought in weren’t right after all.

The objective at the top of the funnel isn’t just to grow, then, but to balance “quality growth.” As Marissa Tarleton, CMO at RetailMeNot, says:

“Growth is about, “How long will that new user stay with us?” from an audience perspective.”

An ad campaign would be labelled a failure if it generated thousands of clicks with zero conversions. The same holds true for product retention.

Interestingly, publishers, communities, and SaaS apps often use comparable strategies to drive “quality growth.” The individual tactics might differ. They might come at it from different angles.

But when you compare their approach with the ultimate end goal, you’ll see the similarities.

Here’s how products like The New York Times, Quora, and Airbnb break down growth engines to acquire the right users from the get-go.

How publishers use referral source and past behavior to predict subscriber growth

People have been banner blind for years, with the vast majority ignoring anything that remotely looks like an ad.

Ad-based businesses have been dealing with this problem for years now. Unfortunately, they have another problem to contend with now. Since that original study in 2013, the use of ad blockers is expected to rise. 3 in 10 internet users will be using blockers by the end of 2018, according to eMarketer.

And it’s costing publishers billions of dollars. Then, publishers also have to deal with tech platforms continuing to encroach on their space.

“Google, Facebook, Yahoo, Microsoft, and Twitter captured 65 percent of the $59.6 billion spent on digital advertising last year,” according to Nieman Labs.

As a result, publishers like The New York Times are relying more on their paywalls to acquire and keep users.

They introduced the first paywall only in 2011 but have already long passed a millions of digital subscribers since.

Interestingly, the way they’ve used this paywall has changed over the years. The number of free articles someone could access before being digitally cut off has fallen from 20 to 10 and will now continue to be cut to five.

They’re even considering a new model that dynamically changes this number based on non-subscribers get to the site or other factors, like their reading habits, according to a recent Bloomberg article.

Other publishers like The Wall Street Journal, The Washington Post, The Boston Globe, and even Wired are also tightening access to their content.

The Wall Street Journal is cutting off Google visitors. Non-subscribers to the Boston Globe only get access to two articles every 45 days. And Wired will be releasing a metered paywall for the very first time.

Yet, despite all of this tightening, subscriber numbers continue to hit record highs.

Nieman Labs confirms that “any ad revenue declines that result from fewer pageviews are likely to pale in comparison to the revenue gains from new subscribers.”

The first step these news organizations have taken is to continue to refine their paywalls.

They’ve started closing off a lot of the ‘loopholes’ that previously allowed people to gain more access to their content.

For example, The Boston Globe will now require incognito users to sign in before being able to read a single article.

Meanwhile, The Washington Post has experimented with offering a free six week deal to readers who also register by email.

Major publishers have traditionally allowed some ‘soft’ openings for users to get free access to content. The goal is similar to any app: give users a free trial or taste test in hopes of keeping them around for the long haul.

So they’ve been continually refining just how much access it takes to get people to join for the long haul.

One study found that most of the major outlets allow some degree of ‘free’ reading.

Publishers are also starting to use predictive analysis to determine when paywalls should be raised or lowered for each individual.

David Skok, former Managing Editor at The Boston Globe, lays out one compelling use case:

“Imagine a reader browsing the web on their smartphone while on a train heading into work. They click on a link through Reddit and arrive on your news site where they are served a paywall. Using predictive analytics, we are quite certain that this Reddit mobile reader will not subscribe to your website. In fact, the reader may even post on Reddit just how much she despises your paywall. So, instead of wasting our time trying to get that reader to subscribe, what other kinds of value can you exchange with her that could be of mutual benefit? Perhaps it’s an email newsletter signup form that could begin an inbound marketing relationship? Perhaps it’s a video preroll ad with a high CPM to generate maximum ad revenue? Perhaps it’s a prompt for the reader to “like” you on Facebook so that they can help expand your reach?”

The Wall Street Journal, which has always been a subscription-based product, takes a much harder stance on paywalls.

Karl Wells, Vice President of Sales and Marketing for their parent company, puts it plainly:

“Who is your most valuable audience? Is it the people who are paying you, who are are sustaining the future of quality journalism? Or is the people coming to your site for free and not paying?”

For the Wall Street Journal, the product is the marketing. Quality content is the reason people stick around.

Beyond paywall testing and predictive analytics, new technology can also make it easier for these loyal subscribers to continue reading for weeks and months to come, as we’ll see later with The New Yorker.

But they, too, rely on data science that predicts the likelihood of someone subscribing based on previous actions others have taken.

The first rule of sticky growth is to use referral source to predict what people are going to want.

How online communities like Quora achieve “quality growth” through growth engine iteration

Quora recently raised a Series D worth $85 million, giving them a valuation of $1.8 billion. That sky-high valuation comes in part by being the one of a kind.

They curate first-hand expertise vs. Wikipedia’s second-hand approach. And while Yahoo Answers has been around much longer, the content is more often low-quality or out of date.

Right off the bat, Quora is taking a page from publisher’s by looking at where people are coming from, first.

They’ve been building out language-specific sites into Spanish, French, German, and Italian, so people can ask and answer questions in each native language. In the near future, one of their ambitious projects includes allowing real-time translations to be performed on the site.

Their ultimate goal is to use tools or features like real-time language translation to help amplify the built-in network effects. This approach builds on Andy Johns' successful work as Quora’s first growth leader, after making the rounds of both Facebook and Twitter.

A big part of Johns early success at Quora had to do with the early company’s DNA:

“I learned growth hacking at Facebook, then at Twitter I was able to play my hand at organization building and influencing the DNA of the company. But when I started working at Quora, it was the best of both worlds – where the developers and founders gave me ample runway to do the right things. The culture friction didn’t exist.”

But how does this manifest itself? What’s a practical example of how this benefits your company?

Johns elaborated for Mark Fidelman at Forbes:

“Quora had the best technical infrastructure to make quick changes to the code without it impacting the rest of the platform. ‘We just started getting stuff done. We were able to run dozens of experiments at a really fast pace and quickly started producing significant results. We were able to run multiple experiments a day to test, fine tune and optimize user acquisition.”

Despite Quora being a Q&A site, its path to success has looked remarkably similar to publishers. For starters, success begins with a quality product. Or in the case of publishers and Quora, content. Johns elaborates in a Growth Hackers AMA:

“The truth is that you have to always focus on quality. The Silicon Valley is obsessed with exponential growth curves. Those can come at significant cost (e.g. Zynga). So I always pick the company that can maintain a reasonable rate of growth over a very long time (e.g. 10 - 20 years) than the company with 2 years of killer growth followed by 2 years of exponential decay. I think that's the potential for businesses like Quora that focus on quality over quantity. It's about how durable your growth is. Not about how quickly the growth happens, IMHO.”

Pure acquisition and user growth isn’t the goal. Instead, the key to Quora’s success has been a never-ending focus on quality growth -- the kind that sticks around -- by again mimicking the same strategies publishers are also finding success with. For example:

“One of the keys to success was to observe the most active users and study their patterns. Then, create experiences for new and existing users that encourage them to fall into those same patterns.”

Exactly like publishers are doing with porous paywalls that personalize permissions based on referral source or predictive analysis.

Johns doesn’t like looking too far out into the future, referring to six-month+ projections “chaos theory.”

He likes to boil things down to the essential components of a growth engine, first, before creating and running iterative tests that seek incremental bumps along each step.

At Quora, that looks like:

  • 1

    Quora grows by getting people to write content

  • 2

    Each piece of content written gets Quora more traffic

  • 3

    Some of that traffic converts into new signups

  • 4

    Some of those signups becoming new writers

  • 5

    And the process continues

Tests are then designed around each micro-conversion, with iterative growth models being updated monthly as Johns received new inputs.

The trick is twofold for communities like Quora.

On the one hand, you can’t lose sight of who’s coming to the site (and why), because that often has a huge bearing on what they’re going to be willing or able to do.

While on the other hand, you also can’t afford to make changes that lift the first few metrics at the top of your growth engine if they lower or decrease the ones towards the bottom.

That’s the difference between heavy acquisition companies like Zynga or Groupon who shoot up, from those like Quora or Uber that show lasting promise.

How Airbnb and Calendly use customer feedback to fuel growth

About a decade ago, Airbnb initially found traction with an age-old platform hack. They piggybacked on Craigslist property listings, creating a simple integration that allowed users to one-click post a user’s existing listing to Airbnb’s platform.

Years earlier, a little video platform named YouTube used the same idea to pull users from MySpace. They were one of the first to allow users to openly embed videos to other sites.

After Airbnb’s early success, they turned to a Dropbox-like referral system to help increase viral growth from users already on the platform. Not only do referral systems work because they leverage network effects, but also because referrals often bring in people who’re almost ‘pre-qualified.’

It’s the same reasoning behind why Facebook ad lookalike audiences perform better than randomly targeted ones: people who’re most like your best customers are better prospects than strangers.

A referral system is a feature.

But it’s also like a mini growth engine that Andy Johns spoke of earlier. There are a series of micro-conversions that people pass through before becoming a new user.

Jason Bosinoff, Airbnb’s Director of Engineering, shared a few lessons they learned when revamping the referral system to increase user signups over 300% per day.

They started by identifying success metrics along each step or micro-conversion of the referral system.

Here’s what that looked like for them:

  • 1

    Monthly Active Users sending Invites

  • 2

    Invitees per Inviter

  • 3

    Conversion Rate to New User

  • 4

    Conversion Rate to New Guest

  • 5

    Conversion Rate to New Host

Airbnb then adopted Johns’ second approach of building out different growth forecasts based on this engine, using “Good,” “Better,” and “Best” to show what the end result might look like with changes at each iterative step of the way.

After setting up event tracking for each user engagement, they set out to expand support across their web and mobile apps. That way, people could send and receive cross-platform.

Personalized user codes would be generated and sent out to new recipients:

And tapping that button sent each user to their own personalized landing page depending on past behavior.

For example, here’s what a brand new user, who downloaded the app directly from the store, might see on their mobile device:

But if you accept a referral invitation from a friend, here’s what your landing page would look like, instead:

Little experiments like this, to simply increase the number of people who receive an invite and sign up, add up over time. Airbnb also identified “product levers” to see a cause-and-effect relationship play out.

Here are three that Jason highlighted:

  • 1

    We can increase “# of invites per inviter” by offering imports of address books from email services or recommended contacts on the phone.

  • 2

    We can increase “share of active users sending an invite” by improving the discoverability of referrals.

  • 3

    We can increase “conversion to new guest” by allowing referrers to send reminders to invited + signed up users that they have credit available.

They then double-down on each step.

They would try to isolate the times someone was most willing to refer a friend. They triggered the feature at the exact right time, then “segment performance data by entry point” so they could compare results.

Another experiment looked at the copy on individual emails, comparing the results of a self-interested variation with an altruistic one.

The altruistic one won out. The point is that ‘sticky growth,’ or retention, isn’t a single step, hack, or tactic.

It’s the end result of this entire process, from when people first come into contact with your product and through the very first few steps of your growth engine.

And that starts by understanding how users ultimately benefit in the first place.

Claire Suellentrop found this out first-hand while at Calendly:

“The best marketers should actually be working in some support or success capacity before you ever touch anything. You need to live in your own territory before bringing other people into it.”

Her first days on the job weren’t focused on getting users. They already have over 10,000 in beta waiting for her. Getting people interested in the product was the easy part.

The hard part was getting them to stick around.

SaaS economics meant that Calendly needed to retain users over the long haul so the $10/month amounted to something.

So she worked backward, ‘flipping the funnel’ to focus on the most valuable low-hanging fruit.

Claire lined up customer interviews, working alongside both developers and the support team, to understand exactly what your customers experience:

“If a company wants to see the 25–95% profit increase that comes from higher retention rates, it’s crucial to get out of your own company’s bubble and experience firsthand what your customers are actually doing.”

One of those interviews was with Sean McVey, Director of Demand Gen at Virtru. Sean was using the tool Calendly to aid marketing automation.

This was an unexpected use case initially.

So Calendly was able to explore it further and use it as the foundation for new support docs and customer stories to help other potential users just like Sean.

He was unimpressed with “only 25–30% of inbound leads” following-through with scheduling a demo on their old form-based landing and thank you page combination.

Sean tried removing the friction or lag in scheduling, by embedding Calendly directly on the page.

Just this single change resulted in a 50% conversion increase.

But Sean wasn’t done yet. He personalized the form again with a qualifying question to route specific deals to certain sales people (depending on deal size).

In 30 days, the number of inbound leads scheduling a call more than doubled to 61%.

Calendly could then leverage success stories like Sean and Virtus to not only improve the product, but also better understand how to align its messaging with new people.

They’re no longer guessing or throwing words on a page. They’re using words straight from customer mouths that sum up their aspirations, pain points, and end goals.

And they can use these insights to create more “happy first experiences” that will get new users to stick like glue.