Wellness app

Overview:

We were approached by a wellness app that had been spending in the 6 figures pre-ATT…..

Until they were hobbled by ATT.

Since iOS 14.5 started rolling out, their profitability dropped – driven mainly by incomplete attribution as well as crippled optimization on key platforms.

Over the next few months, they lowered their investment to keep pace with their lower profitability – and very soon they were spending well under 20% of their peak spend.

And with the incomplete metrics post-ATT they werent even sure if they were profitable.

Having read about and seen RocketShip HQ’s expertise in post-ATT marketing, they approached RocketShip HQ to help unlock performance and scale.

Could their downward spiral be reversed?

This was a low-down of what Rocketship HQ achieved for them:

How did we attain these results? We combined product-focus, rigorous creative testing, channel expansion – and a relentless focus on incrementality. Here is a summary of the key strategies that worked in sync to drive this growth:

1. Evaluate product health to ascertain scale-readiness.
2. Optimize SKAN to maximize signal
3. Unlock creative and messaging opportunities through rigorous creative research
4. App store optimization to improve conversion rates on store listings
5. Granular keyword structuring to squeeze the most juice out of Apple Search
6. Scaling on Meta with our progression-based testing framework
7. Conquer the limitations of UAC’s automation with our Branch Out strategy
8. Cast a wide net with paid content channels
9. Geo expansion to maximize reach(profitably)
10. Running a tight ship to ensure profitability is in sync with operational complexity
11. Measuring the unmeasurable: using Media Mix Models to understand true impact of multiple channels

Here are a few takeaways that helped us unlock scale that can hopefully help you unlock performance and scale as well:

1. Evaluate product health to ascertain scale-readiness.

We started off our engagement by understanding the scale-readiness of the product – because if the product metrics are not strong enough, no amount of marketing can help. Some of the metrics we looked at were:

• ✅Trial to paid conversion rates
• ✅Retention of paying users
• ✅Monetization and LTV

2. Optimize SKAN to maximize signal

If you’re looking to scale on iOS, you want to be mindful of SKAdNetwork, which is Apple’s protocol for privacy-compliant measurement after the introduction of its App Tracking Transparency policy in 2021.

As we performed an audit of SKAN on the account, we realized that the only events that had been mapped in the conversion schema were paying subscriptions – and because paying subscriptions were happening 1 week after install, these were not tracked by Apple’s system of timers which typically expired within 24 to 48 hours after installs.

Indeed, as most of this app’s growth had happened pre-ATT, the absence of a signal-maximizing conversion schema had been a big factor holding them back.

We reworked the SKAN schema to replace it by one that we’ve seen work for 99% of subscription apps. This would let us track trials(which should be a key but obvious metric for subscription apps); *and* a couple of other metrics in addition, so we could evaluate the quality of trials in addition to trials.

The exact schema we used and the reasoning behind it are in the full case study.

3. Unlock creative opportunities through user research & competitive research

When this account had been managed in the past, there had hardly been any creative updates. While the account had scaled its spend pre-ATT, it had been accompanied by a deterioration in performance due to creative saturation.

Yet, a spray-and-pray approach to creative wouldnt work – gone are the days of making a million variations and dumping them into ad sets to surface winners.

The post-ATT world called for different strategies – and we took a deliberate approach to dialing in on messaging, via our user review framework.

As part of our user review framework, we hit the app stores – and did a deep dive of app store reviews to understand why users were downloading the app, the problems they were trying to solve – and the benefits they were getting from it.

In addition, we did a deep dive into our competitive ads database to understand the competitive landscape. Our competitive ads database is a living, breathing archive of ads that our team notices ‘in the wild’ and chronicles and tracks on an ongoing basis. All key members on our team have an OKR to track and analyze a certain number of ads in our competitive ads database.

Based on our deep dive, we prepared an analysis of the key elements that constituted ads that we saw ‘out in the wild’. Some of these were:

1. ‘Satisfying’/ASMR footage
2. Listicle type ads: ‘3 signs….’
3. Speaking to different pain points: sleep, anxiety, stress, relationships
4. Speak to ease of use
5. Soothing/relaxing footage and sounds
6. First person testimonials
7. Memes

We used these learnings to inform both the ad creatives as well as app store screenshots.

4. Improving conversion rates on the app stores

Once we understood the key user benefits, problems and motivations, we reviewed the app store screenshots and icon of this app – which had not been updated for a long time – and whose conversion rates were well below benchmarks for high-performance subscription apps.

We benchmarked the app’s store presence against key comparable apps. Based on our insights from the user reviews, we ran a series of tests for screenshots and icons using Apple’s Product Page Optimization that cumulatively led to a 20% improvement in conversion rates.

The other big opportunity that unlocked conversion rate improvements was with Custom Product Pages – using which we ensured that high performing ads led to custom product pages that mirrored the aesthetic and messaging of the ads.

The exact CPP framework we used is in the full case study, which you can download for free.

5. Granular keyword structuring to squeeze the most juice out of Apple Search

The first channel we looked at was Apple Search – mainly because the near-deterministic attribution makes it easier to evaluate than other channels.

We first reviewed the keyword ecosystem of the product using ASO tools, and thematically grouped the keywords into categories. We then structured the keywords into thematic campaigns optimized for conversions and performance:

Discovery: this was a search match ad group, with the goal of ‘discovering’ new keywords that would eventually go into the core keywords campaign.
Brandkeywords
Competitors
Corekeywords

Ad groups based on sub-themes(also informed by pain points and benefits surfaced during our user review process)

As we reviewed the full funnel performance(cost per trial) by keywords and keyword themes, we were able to turn off underperforming keywords and themes(turns out competitive keywords dont work too well) – and double down on high performers.

We also turned on CPA optimization for high volume core keyword ad groups. Even though CPA in Apple Search entails optimization for cost-per-install(and not per trial or downstream events), this led to a drop in CPIs and CPTs(cost per trials).

While Apple Search had limits to scalability as it is an intent-based channel – we were able to improve the cost-per-trial by over 25% for non-brand campaigns.

6. Scaling on Meta with our progression-based testing framework

With the foundation of the above steps, we began to scale on Meta, which had been the most important channel for this account.

A big component of scaling was to anticipate creative saturation – and to do this, we would have to surface new creative winners and replace those that were getting saturated.

We had the option of choosing from different testing frameworks – and we chose our progression-based testing strategy to minimize the cost of testing.

We tested all new creatives in a tier 3 geo, moved winners from this test into a round 2 of testing – and only moved proven, validated creatives into business-as-usual campaigns.

and yes: more details about our creative process – and also how we tested our testing process are in the full case study which you can download for free.

7. Conquer the limitations of UAC’s automation with our Branch Out strategy

With UAC, we ensured that events from Firebase were flowing into the UAC account – as we’ve seen a significant delta in performance with Firebase as compared to using an MMP alone.

Although UAC is often spoken of as a set-it-and-forget-it channel, relying completely on automation completely misses the opportunity to granularly target specific user segments on the platform.

Sure: you cannot explicitly target interests or demographics in Google UAC – but our Branch Out strategy allows you to ‘tell’ the Google algorithm about the kind of users you want to go after.

With our Branch Out strategy, we essentially treat inputs to UAC ad groups(descriptions, display ads and video ads) as being similar to search keywords, so we select our descriptions, display ads and video ads so as to target the users that are responsive to the right messaging and value propositions for your app.

and yes: more details about the Branch Out strategy, including the exact account structure, in the full case study PDF.

8. Capitalizing on UGC and virality to scale TikTok

Once we had nailed Meta and Google, and were actively scaling on both channels – it was time to start looking at other channels to expand our footprint.

We started actively testing TikTok(which is something we recommend once a product is post $75k a month in spend – with exceptions of TikTok-native products).

With TikTok, the key element of success is TikTok-native creatives. We leaned into our key creative structures – along with a testing framework specific to TikTok.

More details of our TikTok creative structure in the full case study.

9. Cast a wide net with paid content channels

As we started to scale TikTok, we began exploring native content – Taboola in specific. The key elements of nailing Taboola were:

1. Targeting: we targeted lookalikes and contextual topics/interests.
2. Creative: we combined clickbait-y headlines with a CTA calling out the free trial.
3. User flow: we started with an ad to app store flow, which worked well enough that we stayed with it.

Taboola yielded strong performance but had relatively limited scale – but any scale was good scale as long as it yielded performance.

10. Geo expansion to maximize reach(profitably)

As we started to see scale in the US, we started to evaluate geo-expansion. To ensure we expanded intelligently and with minimal risk, we were cognizant of the possibility that we could see low cost-per-trials but very high cost per paying subscriber(or low ROAS).

To anticipate this, we rank ordered our key geos by conversion rate from trial to payers. We also calculated downstream LTV of each geo to understand our most and least valuable geos – following which we set up a tiered campaign structure across our key channels to target groups of geos that had similar LTVs.

11. Customized dashboards and rock-solid processes to ensure profitability is in sync with operational complexity

With multiple geos and channels, we’d not only scaled spend but also introduced significant complexity in our operating structure.

What was critical at this level of scale was to ensure the complexity didnt trip us up – and that we continued to grow, and grow effectively.

To this end, we set up a customized dashboard with key acquisition metrics that were available at the touch of a button – segmented by channels, geos and time durations.

This saved a ton of time that used to be spent on ad-hoc analyses and pulling data from disparate spreadsheets.

In addition, we set up a rigorous operational and review cadence that allowed us to be on alert for any significant changes in spend and performance. This, in addition to our Slack-first, remote-first team structure allowed us to react quickly and effectively to changes in the marketing and product configurations.

(Details of our operating cadence are in the full case study)

12. Measuring the unmeasurable

A big challenge we faced when scaling on multiple channels in a post-ATT world was that none of the channels’ metrics were comparable to each other.

Apple Search uses Search Ads Attribution API
Meta and TikTok use SKAN
Google UAC uses Firebase(and bypasses SKAN)
Taboola uses fingerprinting(er… probabilistic)

How do we compare these different channels’ performance when they all have apples-to-oranges metrics?

We made use of Media Mix Models using Facebook’s Robyn open source protocol once we were well past $100k in monthly ad spend – to evaluate the ‘true’ value of each source.

At times this led to surprising insights. We noticed one of our bigger channels was not incremental according to Robyn – and when we cut spends by well over 50%, we saw almost no change in blended CPAs.

While that was the biggest win, we used Robyn reported true metrics regularly to calibrate and level-set our performance.

In summary, this was a low-down of what Rocketship HQ achieved for them:

• ✅ 4x increase in scale
• ✅150%+ year 1 ROAS
• ✅12+ creatives tested per month

It took a village – and numerous moving parts to unlock this growth:

1. Evaluate product health to ascertain scale-readiness.
2. Optimize SKAN to maximize signal
3. Unlock creative and messaging opportunities through rigorous creative research
4. App store optimization to improve conversion rates on store listings
5. Granular keyword structuring to squeeze the most juice out of Apple Search
6. Scaling on Meta with our progression-based testing framework
7. Conquer the limitations of UAC’s automation with our Branch Out strategy
8. Cast a wide net with paid content channels
9. Geo expansion to maximize reach(profitably)
10. Running a tight ship to ensure profitability is in sync with operational complexity
11. Measuring the unmeasurable: using Media Mix Models to understand true impact of multiple channels

All of this is to show how it’s STILL possible to unlock tremendous growth, even in a post-ATT world.

It’s not easy – but by applying the rigor of full-funnel strategies, deep user understanding, channel expertise and focus on incrementality, it is absolutely possible to grow month on month, even in a post-identifier world.

It’s what we run day to day in our agency Rocketship HQ.

Want to Learn From Our Team Directly? Click here to join our newsletter Mobile Growth Lab to get insights from the trenches of mobile growth – and be the first to know when we release our in-depth courses.

Want Our Team To Run Your Ad Campaigns? If you’re spending at least $50k a month, Click here and we’ll talk.

Want our team to produce high-performing creative?
Find out more here:
Ad creative production services
UGC production services
TikTok Creative Exchange