It's Time to Build Up Subscriptions Too

A subscription entry strategy for happy users and healthy revenue

As I shared in previous posts about the secret power of onboarding paywalls (intro paywalls) shown right after installing the app, proactively approaching users early on is extremely important for revenue growth. But at the same time, this can be a poor (even unpleasant) approach from a usability standpoint, and it often leads directly to a drop in user retention. That's why it's a surprisingly difficult strategy that requires careful navigation through meticulous A/B experiments.

On the other hand, triggering a paywall when a user attempts to use a specific premium feature is a relatively less unpleasant approach. This is because, from the user's perspective, the context makes sense. "Ah, I need a subscription to use this feature." This approach doesn't negatively impact user retention, but on the flip side, it doesn't do a great job in terms of subscription revenue. That's because to use that feature, the user has to go through quite a few steps to dig into that context. As a result, while the conversion rate relative to paywall impressions is excellent, the problem is that the raw number of paywall impressions is simply too low.

Couldn't we leverage the strengths of both entry points?
I wanted to find a sweet spot that preserved the user experience while also boosting conversion rates.

In the "deep within context" approach that provides a good user experience, the "deep" part isn't what matters. Being placed in the right "context" is the core factor. It was only considered "deep" because users were diving relatively far down to find a specific feature. So, I decided to try building up a solid context on a "shallow" level.

Building up the context

We can't just blindly move the location of all premium features to a shallow depth to achieve this. That's because even in a "shallow context," the important thing isn't the "shallowness," but the "context." Therefore, to plan out the right context first, I took a look at why users were visiting these deeply buried premium feature screens and what made them want to use them.

Among various hypotheses like "out of pure curiosity," "needing a stronger alarm," or "because it looked cool," the idea of "needing a stronger alarm" seemed the most persuasive. Users who needed a stronger alarm were likely browsing through various features until they reached the deeply buried premium ones. That made me think we should go out to the shallow end and find those users who need a stronger alarm.

It was still undecided which user cohort to target or how to appeal to them, but one thing was certain: it was crucial that this be provided not as a short "nudge" message, but as a "series of flows," so that the user genuinely feels they are placed within a "context."

So, I decided to recycle the "premium onboarding" flow that was already being provided to subscribed users. Since the gist of the premium onboarding at the time was also "making existing alarms stronger through premium features," the goal of the flow was similar to our current objective. Thus, this planning became an experiment to place users in the context of strengthening their alarm by offering a sneak peek of the post-subscription "alarm strengthening process" before they actually subscribe.

The method was simple. We'd show a dialog to a specific user group saying, "Your current alarm power is low, try strengthening your alarm," and run the above strengthening process only for users who accepted. Then, when they hit the final CTA, "Strengthen Alarm Like This," the paywall would pop up.

The effect of the build-up?

1.4x increase in subscription conversion, with no drop in retention!

The effect of the build-up was incredible. In a way, we were adding a brand new subscription funnel that didn't exist before, so an incremental increase in subscriptions was naturally expected. However, the two crucial things were how much it would increase and whether retention would drop, and the variable deciding that was "how appropriate is this context?" Seeing the conversion rate increment jump up significantly without any drop in retention, it seems the series of steps where users directly strengthen their alarm power served as a solid context for them to be in. It even took up the 2nd largest share among all subscription conversion entry points...!

Advancing the build-up

This experiment was provided to new users who met specific conditions. At this point, I thought the question "Don't you need a stronger alarm?" might actually be a more suitable context for existing users who have used the alarm app at least once, rather than new users. After estimating the impact this way and that, the target audience was large enough that I expected a sweet impact on par with that of new users. I plan to immediately follow up by recycling this same build-up for existing users. Unlike new users, existing users already have somewhat established usage habits, so we anticipate they will react more to the context of strengthening their alarms using new features (increasing conversion rate), and we expect far less concern about retention dropping.

Also, looking at the drop-off rates by funnel during the build-up process, there was one specific segment with an unusually high drop-off. If we improve the funnel so that more users reach the paywall—the final destination of the build-up—wouldn't the impact of this build-up grow even larger? Likewise, after calculating the impact, I felt confident that we could achieve an even bigger incremental conversion rate. I plan to run this as a separate follow-up experiment as well.

Couldn't we leverage the strengths of both entry points?
I wanted to find a sweet spot that preserved the user experience while also boosting conversion rates.

Through this new attempt, I think I've found a vague answer to the question above.

The core here was building up a context that solves the user's problem (needs) and approaching them first at the right time.

At this point, how attractively it's structured without feeling forced is important, and depending on the target audience and timing of exposure, it can easily become an even more attractive context. Because of this, we'll need to pay attention to continuous optimization for those variables.