Discover more from Growth Gems
💎 Growth Gems #87 - Ad Monetization, Product Growth, and Creatives
This is a free newsletter where I share the key insights from the best webinars, podcasts, panels, and talks on growth. Try it below👇
This week I’m sharing gems on:
These insights come from Mike Taylor, Jason van der Merwe, and Felix Braberg.
🥇 TOP GEM OF THE WEEK
Creatives: swipe files, memetic, efficiency
A lot of growth practitioners and experts don’t share anything publicly.
Out of the ones who do, if I had to name the most underrated one, it would be Mike Taylor (Co-founder of Vexpower - also underrated).
So, I got excited when I saw that Shamanth Rao (CEO at RocketShip HQ) had him on the Mobile UA Show episode How to build a creative process based on meme-mapping.
It’s on creatives, but the principles behind them are more profound. I had to mine it!
💎 Build a Swipe Files of all the examples of success in a specific domain you want to conquer and tag the recurring patterns. This works great to spot recurring themes for ads that perform well (or don’t perform well). Then, you can decide what to copy and when to innovate.
⛏️ Going Deeper: Mike shares more about why in this essay. Swipe Files work for ad creatives, but also very well for many other things (onboarding, lifecycle marketing, etc.). This is the same theme as “being a student of the game”.
There’s no wrong way to start; you can improve it later—Miro board, Airtable, etc.
💎 Gaming companies are at the forefront of creative testing. One reason is that there is no underlying substance (vs. selling shoes, for example). Gaming can be very lucrative because of great margins, but there is also a lot of competition, so it has become a marketing game.
💎 Borrow memes from a category or thing that your customers like already. Example: Monzo bank, which targets mostly students, made their card orange which was very different at the time. It started because one of their designers had bright orange Nike shoes, which was also in line with their target audience.
💎 Post-ATT, you should only “actually” split test your most existentially important things. Keep it for big conceptual differences, the one or two very important things. Don’t waste your testing slots. They used to call this performance branding. Example for a digital bank: splitting the bill with friends when traveling might resonate with students but not with someone that mostly eats with their wife. Cheap currency exchange might work well with both students and couples that travel, but not for other groups.
💎 When you have a lot of internal stakeholders that want to give their opinion (it’s on brand, it’s good, their opinion has been heard), you need to optimize everything else in the process around that one bottleneck. In creative production, this translates into having a “creative lake”: fill up your reservoir with already-approved modular creatives. Example: you shouldn’t say, “I want to show this ad” but instead, “Can you approve these 6 stock photos, these 7 illustrations, and these 20 copy lines”.
💎 At a big scale, the difference between paying $100 or $500 for creatives is nothing. Something that might seem inefficient, like having 4 people working on the same task, might be efficient in light of your bottleneck.
Product Growth: testing principles, team setup
There are a few podcasts for which I never skip an episode.
The Sub Club podcast by RevenueCat is one of them.
Strava is often touted as an example of a subscription app with a great organic growth loop.
In Lessons From Building a 70-Person Growth Team, Jason van der Merwe (Director of Engineering, Growth at Strava) shares growth advice and his experience with David Barnard (Developer Advocate at RevenueCat).
💎 When starting a growth team, don’t over-index the impact/importance of specific tests. First, figure out the process for going through the complete a/b test cycle (idea, hypothesis, running the test, evaluation).
💎 Find a service like usertesting.com and have your team watch videos of customers going through the onboarding.
💎 New user activation tends to be where you can get the most wins early on: connecting people who showed intent to come to your product with the value of your product.
💎 Early on at Strava, they had an “Experiments weekly” meeting where they presented all the growth experiments. There used to be a lot of debates and exchanges, but as the team grew, it became less valuable, so they decided to split it into separate meetings for each growth team.
💎 They used to do a quarterly “Experiment Day” where the team could run whatever experiment they wanted: no buy-in needed. Often it ended up being experiments that were not prioritized because of the evaluated impact. Example: follow back button after someone follows you, that translated into the biggest uplift in follows.
⛏️ Going Deeper: Jason also shared that some of their best wins have come from engineers 6 months out of college. For them, it pays off to make room for everyone’s ideas.
💎 Try to understand the impact on WAU (Weekly Active Users) over a year that moving a metric like 2-week retention might have. Strava tries to model this impact and uses hold-out groups for better comparison.
💎 The power of copy is huge. Sometimes, images in product settings seem to be distracting (maybe because they could look like an ad). Copy, no matter where it is, has a huge impact (e.g., “Continue with” vs. “Sign up with”).
Ad Monetization: unity bug, Amazon Publisher Services, AdMob
In two & a half gamers session #52, he shared several gems valuable to readers monetizing through ads.
For advertisers…Don’t forget you’re on the other side.
Enjoy the acronyms 🤯
💎 If you have too many placements on the Unity dashboard, you get latency on the server side because they get latency on their side too.
💎 InMobi placements perform better on banners compared to their bidders.
💎 The AdMob billing algorithm is based on force-feeding placements for the first 3 days to drive learnings, which gives you a revenue spike.
💎 Unity had an over-reporting bug where they were capturing all the high-value placements in the waterfall, which led to reporting 10x the earnings. No actual revenue was earned, and it messed up AdROAS campaigns. You can file a complaint with Unity to get compensated.
💎 Amazon Publisher Services (APS) bidding algorithm works by matching people with their Amazon ID. They prioritize onboarding publishers that overlap with their customers (primarily games that older audiences play - the assumption is that they have a higher purchasing power).
⛏️ Going Deeper: Felix wrote an entire article about this here. Since then, it seems that Amazon has been reaching out to some game developers.
💎 You can also make direct Transparent Ad Marketplace (TAM) deals with DSPs, which changes the commission to just a few cents. Top DSPs on APS: Xander, ShareThrough, OpenX, LoopMe. Mobilefuse, Somato, Rubicon, IndexExchange, InMobi, Pubmatic. If you’re starting, don’t run any deals and figure out the top DSPs on the Monetization side, then reach out to them to do an IO deal.
💎 Amazon Publisher Services currently doesn't support the video end cards, which are superimposed by ad networks and highly impact CTR.
🤔 My 2 cents: how many advertisers thought they did something amazing to improve their CTRs? No, it was just the end card change on the publisher side…
💎 Experimenting with end cards on networks like Unity/AppLovin makes a big difference on the UA side in terms of CPIs and purchases. Example: different visuals.
⛏️ Going Deeper: they talk more about this in two & a half gamers session #33. Apparently, Liftoff relies heavily on very aggressive end cards.
💎 AdMob is finally dropping its multi-call limit policy on third-party mediation (3 AdMob placements per waterfall if you don’t mediate with AdMob), which was aimed at forcing you to their mediation. If you did not respect this rule, AdMob would throttle your inventory by 80%. You can now expose your whole waterfall to AdMob demand.
⛏️ Going Deeper: Felix also wrote an article on this three-placement limit on his blog.
He also shared that to test this, you still need to get whitelisted on Google bidding and should do the rollout over 15 days:
Put it on 5% of your inventory for 5 days, and then if results are positive,
Roll it out to 40% for 10 days, and if it looks good, promote it to the whole app.
And before I leave, here is a quote on creative testing (e.g., testing first in India) from the webinar Winning Creative Strategies for Games
“Test your test process” - Rhiannon Price (Head of Performance Creative at The Sandbox")
See you next time. Stay savvy!
How to build a creative process based on meme-mapping on the Mobile User Acquisition Show
Lessons From Building a 70-Person Growth Team on the Sub Club podcast
two & a half gamers session #52 - Admon gossip corner, Unity expensive bug, APS, Google 3 placements on the two & a half gamers show