Over the past eight months, we’ve increased our conversion from download → paying users by 7.6x. We’ve run (many) tests, and here are the three things that drastically improved our conversion at Alpe Audio:

1. Putting a paywall upfront
2. Overhauling the freemium model
3. Changing paywall/free trial.

Brief context on Alpe: Alpe Audio is an audio learning platform where you can master topics from A-Z, through audio courses. Basically Coursera meets podcasts.

Putting the Paywall Upfront

The first change was to move the paywall right after the bare onboarding. We didn’t feel comfortable doing it. In fact, the opposite. It felt super uncomfortable. But it was one (of many) tests we ran. It really worked. Putting the paywall upfront led to an immediate 3x in free trials started and didn’t impact the conversion from trial to paid. Naturally, some users pushed back against having a paywall shoved in their face. We did allow an easy way to close it and discover Alpe Audio for free, but it’s not the ‘default’ or the first thing a user sees and the pushback was legitimate. However, the overall numbers were just much better, including the drop-off in free users. This was an easy change that felt wrong but turned out to be the right thing to do at our stage. Being at the right stage is important. We might re-visit this again later on, and even reverse this when we have a larger funnel with more users and better onboarding. In that event, it might be better to have the paywall appear later, but with a small user base, this led to immediate 3x improvement, which is big. When this compounded with how we changed our free trial paywall flow, the results get even better. But we still care about our free users. This is why we also changed our freemium offering.

Changing our Freemium Offering

Changing our freemium was the largest change we made both in terms of UX & business model. Much more challenging than shifting screens around. Let’s start with why we did this. We weren’t happy with the value we were offering and thought there was much more to offer.We offer a free tier for two reasons:

1. Business-wise we believe that learners should be able to experience value before pushing the buy button and committing.

2. Values wise we believe that learning shouldn’t be gated entirely based on money. Even though Alpe Audio is a steal and offers incredible value for money, some folks can’t afford and we’d love for them to be able to career upskill because of something they learned on Alpe.

   
   

Based on these two principles we did some market research to see what similar products offer. The result was to offer the first one or two lessons of any given course for free and then put a paid gate. Lots of apps do this. But we found that this doesn’t work. Users have no progression and nowhere to go. They know it’s gated and so from the get-go put in less time and effort to try out the product. They churn = higher CAC. So we added in some more value: a few free courses and a randomly chosen free daily lesson. Why? So that learners could also access some courses and experience the full Alpe Audio experience as well as enjoy the random selection a free lesson can offer: Shakespeare, finance, product management etc. But it didn’t work out that well. We interviewed dropped users and found that:

1. The experience of seeing most lessons locked 🔒 was a turnoff. It gave the impression that there wasn’t anything to return for — even if that wasn’t in fact the case.
2. Random free lessons are fun and a dopamine boost (‘what will it be today??’) but it’s not enough to keep users coming in vs. their podcast or music app of choice. Competing with ‘free’ is hard.
3. Free courses are meh. They have to be a perfect fit for what a learner is looking for. Not usually the case.

So what did we shift to? A Zoom model: 40 free minutes of learning a week. Any course, any lesson. The entire library. The hypothesis was that this will increase engagement across the large number of Alpe Audio users who haven’t subscribed yet. 40 minutes is a good amount to give away: you can learn a lot from doing 40 minutes a week of anything. Fulfills both of our reasons for offering a free tier.

How did our shift of freemium perform?

To test we measured: listening time, retention, conversion from free → paid, and top of funnel bounce rate. This change made a huuuuge difference. The results were great:

• High free user retention
• 10% conversion to paid from those who skipped the upfront paywall
• 2x increase in median listening time per user

Median listening time for nonpaying users *doubled*. Off a base of hundreds of users this is meaningful. More users listening to more content.

Retention for the free users over 30D grew 70%! Users who utilize the 40 free minutes have flatlined so far at 35% retention vs 8% industry standard. High retention = people reaching the paywall more often BUT after they’ve experienced the full Alpe value. Which leads to higher conversion.

Conversion free → paid is the key driver for our business. The key to this is getting more people to the paywall and getting those people to convert. The users hitting the paywall after 40 free minutes converted at a whopping 10%. That’s a big number for us, especially since these are users we used to lose.

Bounce rate is always a concern, and changes in the funnel can cause self-selection somewhere up the funnel which → higher %% off a lower base. So we had to make sure that this wasn’t the case, and indeed, the bounce rate from top of funnel stayed steady.

Changing the paywall

Last but not least, was changing our paywall and free trial funnel, which led to a further 37% increase. We’d tried a few small tweaks to the paywall over the months, but when we saw a case study about Blinkist and how they changed their paywall flow, we knew we had to give it a try.

Which of the below paywalls performed better? Option A, on the left is simpler and communicates the value you get. Option B, on the right, deals with the common user fear of forgetting to cancel but doesn’t communicate the value you get as a user.

Option A vs B for paywall design

I highly recommend you read the original post to understand the why, but the results were quite clear: option B outperformed by 37%. This 37% was on top of the increase of putting the paywall upfront, and helped mitigate some of the trust issues that putting a paywall upfront creates.

Increasing conversion from free to paid by 7.6x

Bringing it all together, these three changes increased conversion by 7.6x, which has reduced our customer acquisition cost. This hasn’t been an easy process. To get here we’ve run many, many, many tests that didn’t work — which is exactly why we wanted to share this with the rest of you openly: so that you don’t have to waste your time and resources.

Enjoy!

First published here