MADV 008 - #RANTT & AppSnackingTransparency
The iOS14.5 band-aid is being ripped off. Slowly but surely.
As usual, one of my goals is for you to click on the sources I share. Be nice and reward me by not being one of the 60% MADV readers who didn’t click out on anything last time.
⚔️ Store wars
The 30% standard fee on app stores is under pressure from multiple fronts, and the 15% PR stunt may not be enough to put off regulators. We might be closer than ever to see a change in the fee structure Steve Jobs was so adamant to enforce.
I discovered an edge case of IAP exempted of store fee through Clubhouse first monetization attempt: as per a rule created for WeChat in 2018, donations/tips aren’t subject to commissions if 100% of the proceeds go the creators.
Some new information surfaced from the Epic/Apple legal battle, among other an interesting glance at the review process within Apple.
EU commissioner Vestager: “Apple is in breach of EU competition law .. Apple charges high commission fees on rivals in the App store & forbids them to inform of alternative subscription options. Consumers losing out.” Another useless fine incoming years later, or real action to follow this time?
Benedict Evan’s take in Resetting the App Store: “ it’s extremely likely that apps will be able to use their own payment systems, that Apple’s 30% will change, if not necessarily disappear, and that Apple will have to give third party apps the same kinds of hardware and integration access as its own services”
On non-app stores, Microsoft decided to lower its PC games fee to 12% (excluding the juicy console market) in attempt to compete against Valve. Interestingly, Steam has the opposite logic of the recent Apple/Google change: the fee goes down for the biggest companies (25% from $0m, 20$ from $50m).
I’m looking for apps who are selling both digital and physical goods, thanks to MADV readers to point out to me cases such as TheEconomist, Revolut, Sweat or Fishbrain, but I still failed to identified bundled in-app digital+physical examples that would help a promising early-stage startup I’m advising.
💀 iOS14.5
You may or may not have noticed I avoided mentioning any of it in last MADV, as I felt we were all on overdose and it was best to wait for real, post-launch data. We’re there, and it’s easy to get distracted on tactical level: fighting to get releases approved while using first-party only, digging into yet another ATT prompt optimization test, losing hair & nails trying to figure out what those partial if not random SKAN conversion values compare to our beloved deterministic values… Better than adding to the pile in-depth webinars I participated to (with Singular, Appsflyer, MobileAction among others), here are a few bigger picture perspectives or recent datapoints I found interesting:
iOS14.5 was finally released on April 26th. First week adoption seems slow so far (2 to 8% across accounts). Apple is rolling out notifications differently across version, as shown by recent Flurry data.
I had fun times with David on the SubClub podcast, going back to fundamentals and the role of paid advertising in mobile, especially early stage. We discussed some of iOS14.5 consequences in the last part. As usual, if you can’t spare 80 minutes in such a busy week, head to growthgems for a “blinkist” of the episode (17x 💎 on the #gemometer). Eric Seufert posted a thread completing my comments about UA at early stage
I found MMPs to have some of the best content around the iOS14/ATT changes. But if there’s anyone independent who’s helped developers navigate the chaos, that’s Eric Seufert. His 2h workshop on “iOS 14: How to prevail” for just $199 is a steal, and giving some karma back is just fair for all the countless free resources he’s released. In any case more actionable than this SKAN mechanics amusing description.
“No IDFA, more problems”: here are 5 hard questions about app marketing measurement (by Vungle) which are a good exercise to measure your level of prep. Good luck providing easy answers to your CEO.
Eric debunked the myth that “Mobile advertising spend can’t just disappear”, arguing that “Like a glacier receding into arctic waters, mobile ad spend can disappear: unable to recoup enough money to fund growth, the DR advertiser simply has to spend less than it previously did in the face of a step-change in performance”. That’s a real possibility, although I believe the app economy is more resilient than that, not fully dependent on ads, and not entirely rational (isn’t it why VC money is for?). Taking a more optimistic stance, we might see a V-recovery within a couple quarters and get back on trends, which would show up this way:
To look at one of the few bright sides of SKAN, SANs inventory for app installs is about to grow significantly as app advertisers finally will be able to reach IDFA0 users, a long term frustration I raised in Q4 without much exposure.
Not enough attention is given to alternative measurement approaches. Hopefully we won’t have to rely on those unreliable “how did you hear about us” screens that might be directionally helpful, at times, and in complement of other attribution proxies. I’ve mentioned emerging startups to deal with those headaches, like Algolift or metric.works. Maor Sadra from Incrmntal just released an interesting whitepaper on 5 incrementality methods that don’t rely on IDFA (signup walled). From another of their resources, that pix was very much on point:
It’s great to see Apple iterating on SKAdNetwork quick. With 3.0 incoming on iOS14.6, Apple is opening a window on the unicorn of mobile, multi-touch, but I’m slightly concerned about additional mess of multiple networks getting the same postbacks…
App events maestro Louis Tanguay wrote about a “rotten fruit in a lush garden”
Fed up of iOS14 by now? Here’s some swag for you (h/t Niko)
A research on the meanings of “14.5” outside of mobile led me to interesting considerations, between a heavy Russian machine gun caliber, Barnes’s take on Job14:5, and this karmic number interpretation “14/5 is telling you that there has been a pretty hefty abuse of freedom through being irresponsible and shunning any sense of responsibility.” #food4thoughts
🎪 The ATT circus
We’ll start seeing what those opt-in rate look in practice… There’s hope if we read Appsflyer’s benchmarks (data before April 26). It’s less in the wild, says Flurry, reporting just 11% opt-in and 4% in the US last week (not even factoring restricted devices). Or somewhere in between, based on Kayzen data (20% or non restricted).
As always, be careful about benchmarks… I’ve had a heated debate recently with a friend, on why benchmarks are vastly misleading. Eric stated why you should care about ATT figures, which reminded me of this old but gold MDM article.
Talking about opt-in rates: a large number of users have reported on Reddit, Apple forums & else that they had the “Ask Apps not to track” settings option activated by default and greyed-out (aka, “restricted” ATT status). Although Flurry says “restricted” is just 3-4% of devices, I’m seeing wild numbers in the nature. (Pro tip: If you’re restricted, log out & log back in your iCloud account).
Have you taken a specific look at IDFA availability on iOS14.5 and restricted status percentages for the “non ASA” countries Brazil, China & Turkey? You might get a bad surprise…
…But maybe you’ll get a good one elsewhere. Who would reasonably click on that French or Italian “Ask app not to track” button?! 😂
Besides the figures themselves, it’s definitely interesting to look how different apps are requesting consent. Instead of downloading every app out there, head to Sylvain Gauchet’s awesome attprompts.com and get a full gallery of prompts and pre-prompts. It’s crowdsourced, so go and add the ones you see!
If you browse attprompts.com enough, you’ll notice quite a few are infringing Apple latest guidelines (no filtering, no mimicking, no incent, no “allow” on the CTA). Well, I found Apple’s first example bordering on incent myself…
One of the most interesting consent approach I’ve seen came from Snap, with a look&feel that reminds me the style of the growth.design case studies
Sex at first date? At the opposite of the spectrum, a large number of games are prompting at first launch, with no pre-permission nor context… I want to believe this was tested, and while still counter intuitive to me, I’ve seen quite a few pre-prompt screen tests fail at improving opt-in rates…
As usual, games were quickest to react, with single-digit adoption of ATT in non-gaming apps according to Appfigures (4/29).
Božo Janković took a look at all top100 grossing games 48h after release, with only 42 showing the prompt, and barely 8 with a pre-prompt screen. More stats and tons of examples in his full post.
While apps won’t likely be able to circumvent consent that easily (even in China, where the government said no to CAID), “there will always be bigger fish in the sea. Cell networks and ISPs simply don't need to get device-level consent from users, because they're further up the food chain”, writes Alex Bauer in mobile growth news #25.
🍎 Apple Search Ads updates
As revealed as early as February on the ASOstack Slack channel, Apple has now rolled out its suggestions in search in US, CA & UK. Let’s see the impact on ASA inventory. (Edit: unlike what I initially mentioned, it doesn’t only apply on single keyword queries)
Tech media got wind of the new AppStore ASA placement I showcased 3 months ago (slide 43). While the display ads in Stock & News apps are postponed, this new inventory will imminently be available (May?). I’m told by developers in the know it’s CPM only, which makes me wonder what kind of price we’ll see. I wouldn’t be surprised if a few of us get charged 3-digits…
Many pundits whine that Apple uses “tracking” settings others can’t. To me, it’s all first party so far, hence not “tracking”. The opt-out option for Apple “personalized ads” is much more like Google’s “web & app activity”, also buried settings. than the “Allow Apps to Request to track” that involves third party. I always unconsciously call that option “request apps not to track”, maybe so in order to name it #RANTT. Regardless, Apple was never known to be very strong with targeting and we’re already seeing this live, as for reported for instance by Nadir being advertised Aussie snacks (bonus point for Matej’s reply 🥨).
Also on ASA, the new measurement framework Adservices is now in action (and there’s certainly self-preferencing here), although I’m quite disappointed so far, as I expected significant improvement from the LAT-on inventory reporting.
#Freenemies: According to MobileAction data, the biggest ASA advertiser is…
SearchAdsHQ released very detailed ASA benchmarks, including a flexible datastudio to play, link hidden at the end of the report
💡 Other (non-app) interesting reads…
“Online Pareto”: in digital, 99/1 is the new 80/20. Content creation isn’t an exception in the winner-takes-all nature of many online markets, with 97% of Youtubers not making minimum wages, 98% of Patreon creators below $1k monthly and 98.6% of Spotify artists making in average $12/month thanks to the $0.00437 per stream alms Spotify is granting. Just like in the rest of society, a healthy middle-class is missing (h/t Heartcore)
That’s partly why monetization focus is shifting from advertising to commerce in general, and subscriptions in particular. Some winners took that train years ago, like the NYT: early 2000, ad revenue was over $1b yearly. In 2020, subscriptions made $1.2b (half from digital only, growing 30%) while ad revenue fell down below $400m (another -26%) . Other alternatives to the attention merchant model who won last decade include micro-transactions, virtual currencies, and now NFTs (check this great NFT mansplaining). MADV doesn’t have any of this, but have you acquired any of my bitclout yet?
a16z often has amazing content (have you not checked their landing page on Social? legendary). This overview of user research platforms is another one.
Reforge (now a16z portfolio) also has incredibly valuable articles, their analogy of growth & racecars was on point.
Mistakes running a growth team, by John Egan
I had 3 more sections on the latest subscription apps tactics, web-to-app strategies & ASO news, but those will wait for at least MADV 009 , and the iOS14.5 dust to settle a bit…
Thanks for reading! Sharing is caring.