💎 Growth Gems #60 - Paid UA and Lifecycle Marketing
Hi there,
This week I’m sharing gems on preventing churn and winning lapsed customer back, onboarding/monetization optimization, measurement and SKAN . These insights come from Michael Stysin, Eric Seufert, Thomas Petit, Piyush Mishra and more.
Enjoy!
🥇 TOP GEM OF THE WEEK
💎 Private relay is not an opt-in by default, even if you’re an iCloud+ user and update to iOS 15. You have to actually go into your settings and choose to switch on your private relay, and only then will your IP and other details be hidden from tracking on Safari.
by Piyush Mishra (Growth Marketing Lead at Product Madness)
at 23:38 in SKAN: a playbook for today and the future
Acquisition: measurement, brand marketing, SKAN
You knew I wasn’t going to miss this one!
David Barnard (Developer Advocate at RevenueCat) had the good idea of bringing Eric Seufert (Analyst and Strategy Consultant at Heracles Media) and Thomas Petit (Growth Consultant) together to talk about Subscription App Trends and How to Grow in 2022.
Below some of the insights shared on paid acquisition.
💎 If you’re a small developer, start by focusing on your most viable channel (e.g. Facebook) and try to send as much traffic as possible to the web, since all traffic you’re going to get to the app will be hard to measure. This means you’ll need to test your web onboarding a lot.
by Eric Seufert
at 28:45
💎 You have to be more radical in the different ad creatives you create, because you can’t deploy and test as much as before. Bringing creative production in-house might give you an advantage because it allows for creatives to be informed by deep product knowledge.
by Eric Seufert
at 39:22
💎 Brand marketing for digital products should elevate the performance of direct response. This means you need to prompt users again with a “follow-up” digital ad.
by Eric Seufert
at 44:45
⛏️ Going Deeper: Eric explained how this is very different from the approach you might have when you’re running brand marketing for a food item, where you know that people will see it again in a grocery store. Eric shared his thoughts on why brand marketing sits within performance marketing in his MDM post The perilous mythology of Brand Marketing for digital products.
💎 When you’re just starting, putting an MMP is not the first thing you need to do because you can still operate the basic networks without: Apple Search Ads has its own attribution, Facebook events can be set up with the SDK, Google events you can add with Firebase, and for influencers you extrapolate the results. MMPs make more sense when you expand to ad networks that use probabilistic attribution.
by Thomas Petit
at 51:00
⛏️ Going Deeper: Eric believes the way fingerprinting (i.e. the current “probabilistic attribution” by MMPs) is going to be shut down by the end of the year (with the next major release of iOS). He shared that probabilistic attribution also cannibalizes a lot of organics. What I’m getting from this is that it can be leveraged now while it’s working, but you shouldn’t rely on it too much (i.e. be careful if your partners/networks don’t have a backup plan).
So if fingerprinting is going away “soon”, then it seems like a good idea to get a head start on understand SKAN.
Easier said than done!
Shamanth Rao (Founder at RocketShip HQ) received Piyush Mishra (Lead Growth Manager at Product Madness) as a guest on the SKAN: a playbook for today and the future episode of his Mobile User Acquisition podcast.
Some good insights on this complex topic.
💎 In the last 6 months, DSPs have started getting much less null values compared to Self-Reported Networks (SRNs).
by Piyush Mishra
at 09:30
💎 Non-revenue events tend to be poor predictors of revenue events. However, if you just base your predictions on revenue in the first 24 hours then if only 3 or 4% of users are actually doing a purchase you’re taking the risk of not accounting for 96% of users. Instead, you might be able to look at a mix of events (e.g. somebody reaching the lobby, viewing an ad, and doing a hundred spins) to see if it can become a better predictor of the likelihood of a purchase (and use it for your conversion value schema).
by Piyush Mishra
at 12:36
💎 Besides the fact that the direct postbacks we can now receive are only for iOS 15 users, the other problem is that we don’t have the matching key. We do receive the raw report campaign ID and ad network ID, but we don’t have the key to match it to an ad network name or an ad campaign name, because that exists only with the networks.
by Piyush Mishra
at 21:59
⛏️ Going Deeper: with iOS 15, advertisers can request to directly receive a copy of the postbacks. This can be done through your Mobile Measurement partner if you specify them as endpoints (there isn’t any user interface yet but seems like they are working on it). This is a welcome change, but Piyush explains that it is still hard to exploit the data because you can’t match ad network ID <> ad network name and campaign ID <> campaign name.
For some networks, if you have historical data you might be able to match some things. But it’s impossible for Self Reported Networks (SRNs) (e.g. Facebook, Google) that create additional campaigns IDs on their end (that you don’t really know about) so they can have some ad group and creative-level reporting.
💎 Facebook and Google were not sending re-download percentages in their postbacks. Now that we are getting the postbacks directly, we can have a directional metric on re-download percentages.
by Piyush Mishra
at 22:20
💎 Private relay is not an opt-in by default, even if you’re an iCloud+ user and update to iOS 15. You have to actually go into your settings and choose to switch on your private relay, and only then will your IP and other details be hidden from tracking on Safari.
by Piyush Mishra
at 23:38
💎 Right now it’s a transition period with different layers of data because users are converting through:
SKAN 2.0 (still 40-45% for Piyush), where you’re getting only click-through conversion,
SKAN 2.2, where you’re getting view-through and click-through conversion,
SKAN 3.0, where you’re getting multiple data points on the last bid and who won and who didn’t and who was the assistant installer, etc.
On top of that, you’re getting direct postbacks only for iOS 15 users.
by Piyush Mishra
at 25:57
⛏️ Going Deeper: Piyush shared that to make this even more complicated, only a few networks are working with conversion schema right now because of probabilistic attribution still being used. As long as that’s the case, it’s very hard for marketers to dive into conversion value schema completely. Also keep in mind that it’s not fair to compare the MMPs and Apple’ attributions: the kind of post-install data that the probabilistic attribution gives you is much higher than what Apple/SKAN gives you.
Onboarding: be radical, simplify
A few more gems shared by Thomas Petit in the RevenueCat panel, this time on onboarding and monetization.
💎 If you’re early stage, don’t A/B test: be radical and make the change directly. If it makes a difference, you will see it.
by Thomas Petit
at 57:20
💎 Changing the whole onboarding/freemium experience is going to have more impact than incrementally testing price. Example: paywall on the first screen or how much you offer for free instead of testing $50/year vs. $70/year.
by Thomas Petit
at 58:05
💎 If you have strong network effects (e.g. Strava, Alltrails) then you don’t want a +10% ARPU in the short-term if it results in too many people churning, as it will reduce your overall growth.
by Thomas Petit
at 58:42
💎 Be more explicit about the action that users need to take in the app (the action you’ve identified as leading to retention and therefore renewal). Simplify the app experience: way too often, apps have many features, which distract users. Dynamic In-App messages are very efficient at pushing people into doing the next meaningful action.
by Thomas Petit
at 1:10:20
⛏️ Going Deeper: for the very fist action it’s pretty common to have onboarding lead users directly to it, without “dropping” users in the app first (cf. Headspace example below).
Lifecycle: preventing churn, winning lapsed customers back
When acquisition is challenging, churn gets the attention it deserves…
Michael Stysin (Founder at Qonversion.io) teamed up with Alice Muir and Paulo Golovattei (Growth Consultants at Phiture) for this Prevent churn and win lapsed customers back webinar.
💎 Mobile subscription churn is not linear: subscribers leave at a higher rate in the first billing periods. This means it’s more impactful to focus on churn in the first billing periods.
by Michael Stysin
at 05:45
💎 For involuntary churn, you want to automate the communication to remind users to fix their payment details on file with Apple or Google. You can do this by triggering the communication using billing issue subscription events tracked server-side.
by Michael Stysin
at 09:55
💎 For voluntary churn:
Detect the cancellation event
Send a push notification asking for the cancellation reason
Offer a discount if the cancellation reason is money
by Michael Stysin
at 10:40
💎 Collect the cancellation reason responses and use the corresponding loss in revenue for each response to identify where you should focus your attention.
by Michael Stysin
at 11:08
💎 Do not neglect the free user experience. During your free user onboarding, provide users with the very best of your free experience. Focus on 1 or 2 “a-ha” moments to promote.
by Paulo Golovattei
at 14:25
💎 During trial onboarding, always guide your users through the most valuable features of the paid experience and keep highlighting the benefits of your subscription.
by Paulo Golovattei
at 15:39
💎 Break-out your lifecycle strategy by the different monetization stages: trial starts, free trial onboarding, trial to membership conversion, membership retention, churn prevention & winback. Test and validate individual hypotheses for conversion at a stage (e.g. trial starts), then create a scaled and automated user journey so that you can focus on the next area of the lifecycle (e.g. free trial onboarding).
by Alice Muir
at 19:22
⛏️ Going Deeper: below a concrete example of churn prevention from Allison Bryant (Sr. Marketing Manager, CRM/Lifecycle at Ten Percent Happier) in a recent Braze webinar. They were seeing a dip in retention rates at the 1-year mark, heavily correlated with the engagement in the last month of the subscription. They created daily email prompts for these users (8-9 emails total) sent a month out from their yearly renewals, prompting them to try the “begin again” course. This boosted engagement by 57%.
💎 Deep-dive into the top 25% most engaged users and the bottom 25% least engaged users to identify the actions leading to each. Then shape the user journey accordingly.
by Alice Muir
at 24:12
And before I leave, something to ponder when deciding where to focus your efforts and experiments right now between probabilistic attribution, SKAN and incrementality:
“Of all the solutions that are out there today, which will survive in 2022?” - Piyush Mishra (Lead Growth Marketing at Product Madness)
See you next time. Stay savvy!
⛏️ Sylvain