How Zumba, in partnership with Perceptycs, turned its app into a $4.5m subscription success

In just 14 months the Zumba app went from 0 to $4.5M in ARR, >30,000 active subscribers and won Revenue Cat’s “Anchor’s Away” Shippies award, for the best new subscription app launch. 

And all this in a time when the fitness app market is more crowded and competitive than ever before.

In this article, I’ll break down how my team at Perceptycs worked with the Zumba team to drive such significant growth in such a short period.

Getting started

The Zumba team knew that they could build the best dance fitness app in the market and leverage their brand equity to kickstart growth. After all, you can’t say dance fitness without saying Zumba! And they did just that.

But a few months after launching, CPAs were higher than projected – and rising – making growth unsustainable.

That’s when I got the opportunity to work with the Zumba team to help diagnose the issue and plot a route forward. There was clearly an acquisition and conversion problem but what exactly was the bottleneck?

After diving under the hood and reviewing almost every metric imaginable we decided to remove some noise – scaling back spend on Google and killing TikTok spend entirely to focus on two things:

1) Ramping up creative testing on Meta

2) Optimizing the new user journey (install to trial start)

Ramping up creative testing on Meta

I don’t recommend that brand new subscription apps push spend through multiple UA channels at the same time in a post ATT world – especially if they are offering a free trial. Of course, you can compare on platform metrics. And yes, with an MMP in place you can compare CPIs and CPTs. 

But without running any incrementality tests or isolating specific channels, you’ll struggle to get visibility on how trials from different channels are converting. And unless you have an unusually high ad spend budget for a new app, it’s unlikely that either you or the ad platforms will be able to learn and optimize efficiently.

Killing TikTok and scaling back Google meant we could focus ad spend on a single channel and run a series of systematic creative tests to identify key messaging, angles and creative concepts that could be scaled.

We 10x’d creative testing volume and after just 3 weeks, drove a 52% reduction to CPA.

But the most exciting part was that we now had solid learnings around what language, messaging and value propositions were resonating in the market. What’s more, we knew exactly how to leverage those learnings next…

We took the winning language and tested variations of it on Zumba’s App Store screenshots, splash screen and paywall. The result: A cohesive value proposition from impression to purchase and improved funnel conversion rates across almost every step.

Optimizing the new user journey

But this was just the tip of the iceberg when it came to improving the new user experience.

The original strategy focused on minimizing the steps it took for users to access content in the app so they could experience value before committing to the app. In practice this meant that most users didn’t see a paywall until they completed their first class in-app.

A strong, logical strategy when you’re hosting world class content in your app. But with 80% of users typically starting a trial after just a few minutes into their first app session, this meant that the majority of new users were never even seeing a single paywall in the app. (In our case only 40% of users who completed onboarding were seeing a paywall).

We brought the paywall forward to the end of the onboarding journey and increased the frequency of paywall occurrences in the users’ first session, which alongside the TOF improvements, enabled us to hit target CPA for the first time since launch.

The cycle repeats

At this point the Zumba app had generated over $1M in revenue and metrics were looking pretty healthy across the board. So it’ll come as no surprise when I say that we started to scale up ad spend to juice the machine as much as we could!.

Initially we increased budgets too quickly and saw a spike in CPA as our winning creatives lacked the resonance required to perform consistently at a higher scale. But fortunately this could be controlled. We leant more into creative diversity alongside creative volume and began experimenting with a range of concepts and formats – UGC videos, non-video UGC videos, native static creatives etc. The goal here was to reach different segments of the market across placements and scale horizontally.

At the same time, we ramped up the frequency of product growth experiments to run a higher volume of paywall and pricing tests. These experiments were a mixture of new hypotheses generated by the Zumba team, new hypotheses generated by the Perceptycs team and previous tests we’d seen success with across other clients.

Our primary goal was to increase the percentage of users taking out an annual plan over a monthly plan. This is something I recommend the majority of early stage subscription apps focus on. With an annual plan there is a guaranteed 12 month ARPPU which means you can easily determine profitability and scale up or down as needed. On the contrary, with monthly plans you are forced to predict monthly renewal rates and then wait for actuals to validate the predictions. This is particularly difficult for new apps who have little to no historical data on subscription retention. 

We saw strong initial success with commonly adopted subscription app paywall tactics like removing the free trial from the monthly plan, increasing the monthly price relative to the annual price, and anchoring the annual price against the now higher relative monthly price.

However, it’s taken time to land on a big win from pricing tests, especially across geos. (Something we are still working on today!)

Across all of these experiments, we were cautious to not become overly obsessed with just hitting CPA goals since a lower CPA doesn’t always mean more revenue or increased profitability. Instead we paid close attention to ARPPU:CAC ratios in order to assess performance.

The race begins

From the very beginning we were pushing traffic both directly to the app and through Zumba’s Web2App funnel. The majority of our creative testing occurred on the web since attribution was easier. Although overall, we spent more on app campaigns as we saw a strong halo effect on ASO which brought down the blended CPA.

But as revenue increased and we crept closer to Apple’s 30% fee threshold so did the desire to see web revenue overtake app revenue! So the race began.

Focusing more on the web funnel had the added benefit of enabling us to run onboarding experiments at a faster rate. Whilst we had ramped up product growth experiments before, the focus was on pricing and paywall experimentation. We’d actually only conducted two onboarding experiments in seven months. 

I know what you’re thinking! ‘Only two onboarding experiments in seven months! Why?’ Well, Zumba’s onboarding completion rate was solid and other areas of the funnel needed more attention. Ultimately, it’s important to focus on removing bottlenecks and not testing for testing sake.

But since our web funnel had only really consisted of a landing page, sign up page and checkout page up to this point, we had a strong hunch that we could improve web conversion with increased experimentation. So we started to test radically different onboarding designs, experimented with the length of our onboarding journey and steered clear of a ‘question first’ onboarding journey in favour of an ‘answer first’ onboarding journey.

By this I mean that the goal of onboarding became to answer our prospects’ internal questions. We wanted to prime them to pay for the app, not just ask questions and collect data.

These tests enabled us to see our blended web CPA drop below our blended app CPA for the first time.

Where to next?

And now with ~$350k MRR, the journey continues.

“The past year has been a transformative journey filled with invaluable lessons, and we’re just getting started. Partnering and maintaining open communication with Nathan and the Perceptycs team has had an incredibly positive impact on our results. We’re at the early stages of growing the Zumba community, with a strong pipeline of improvements to drive acquisition and engagement. Success lies in finding the right partners, fostering strong team dynamics, setting a clear long-term vision, and taking disciplined, focused steps to achieve it.”

Lucy Levy, Chief Consumer Officer, Zumba

Key Lessons from Zumba’s growth

Zumba’s rapid growth highlights the importance of prioritizing efforts on high leverage activities, doubling down on both experiment volume and diversity, as well as assessing performance with the right metrics. It’s not enough to just ‘run lots of experiments’. It’s important to start with a set of clear hypotheses and a refined definition of success to conduct experiments that drive real business value. By combining creative freedom with a systematic testing process, apps can improve the metrics that matter and scale at an incredibly fast rate.