💎 Growth Gems #56 - Revenue Growth, Paid UA and ASO
Subscribe to get the Growth Gems newsletter every other Tuesday👇
This week I’m sharing gems on keyword optimization, creative testing and subscription apps. These insights come from Ryan Kelley, Eric Crowley and the Traplight team.
🥇 TOP GEM OF THE WEEK
💎 As there are more privacy implementations put in place, you might not be able to use custom audiences anymore (or costs will rise). You might have to rely on creatives (theme, orientation, duration) to do the targeting for you. Benchmarking these now on your custom audiences allows you to get insights on what to use on more automated or privacy-focused campaigns (AAA, SKAN).
💎 Especially post-ATT, it’s important to focus on organic ways to attract customers, instead of just pumping ads through Facebook. Spend more time defining your niche and finding ways to find your product organically. Examples:
Bike app partnering with trails associations, who then act almost as a marketing partner. Strava is doing something similar.
Pray.com partnering with the NFL for people to submit prayers for players in close games
Gregg partnering with nurseries to have flyers about the app when people buy new plants
⛏️ Going Deeper: check out the Consumer Subscription Software Insights 2021 report by GP.Bullhound where you’ll find more insights and examples like the ones below.
Revenue Growth: breaking the subscription “ceiling”, locals vs. tourists
💎 You don’t see a positive net revenue retention in subscriptions (vs. SaaS, where it’s standard). But some apps do manage to “break the ceiling” of subscription. This is relevant if you’re already at a huge scale: focus on core users to find out what could make the experience better for them. Examples: Tinder with single IAPs to get more relationships, Fishbrain with fishing gear, an app for camping adding a marketplace to buy used tents, etc.
💎 Looking at the average LTV value is problematic because there are in fact different profiles of users: “locals” and “tourists”. You want to understand who the “locals” are (highly engaged users) and find ways to bring them more value. Using the LTV of ”locals” is also more compelling to investors.
Example: walking app where some users end up walking everyday and love it, others drop off after a month or two and others sign up but churn immediately. The LTV of these 3 groups are very different.
⛏️ Going Deeper: once you’ve broken your cohorts into “tourists” and “locals”, you can understand the acquisition channels they come from: if your subscribers churn off right away, you might be wasting money. The “locals” LTV will also look better to investors. You can find more info about tourists vs. locals on page 18 of the report and in this article on the RevenueCat blog.
To give some additional perspective on this: Ron Schneidermann (CEO at AllTrails) was sharing in another episode that one of the big (successful) bets they took was to focus on more “casual” segments of their users. So you also need to keep in mind the growth potential of who you’re focusing on.
Paid UA: creative testing
💎 In all the creative tests Traplight ran:
The share of AN rewarded video grows as video duration increases
FB Feed has the opposite trend, with shorter videos gaining more spend than longer videos
FB In-stream video peeks at 15-17.5, because 15s is the maximum unskippable ad length for mid-roll ads
💎 IPM grew as duration increased but only on AN Rewarded video, not on all placements. This is because it is the longest unskippable placement (30s) so most people who watch actually stay for the entire ad (making them more likely to install).
💎 Facebook’s algorithm has thresholds for spend on ad placements based on resolution and duration. Example: landscape and > 30s will get more Audience Network Rewarded video placements, 4:5 gets no spend on Audience Network unless it’s > 15 seconds.
💎 This means you can identify your placement and duration “sweet spots” for creative production. This can also be used in campaign management, to leverage Facebook manual placements for optimal campaigns but also influence automated campaigns with the help of creative production.
💎 Traplight uses creative modularity for quick iterations and testing. They put together the elements below (they have this in different resolutions)
Standard beginnings and endings, already rendered
Several content blocks, already rendered with sound transitions and fade-in opacity
💎 Traplight learned that the more core the product they advertise, the more consistent the behavior is across channels (Facebook, UAC, etc.). Casual games might see more differences in behavior/performance but for Midcore/Hardcore games it’s easier to rely on legacy data.
💎 As there are more privacy implementations put in place, you might not be able to use custom audiences anymore (or costs will rise). You might have to rely on creatives (theme, orientation, duration) to do the targeting for you. Benchmarking now on your custom audiences allows you to get insights on what to use on more automated or privacy-focused campaigns (AAA, SKAN).
💎 Since creative modularity is not a good fit for “narrative” videos:
First, learn the sweet spots for duration/orientation
Then, leverage the insights for the production narrative-driven videos
ASO: keyword optimization
💎 Keywords at the beginning of the string get more weight. The more important a keyword is, the closer to the beginning of the app’s name (or subtitle) it should be.
⛏️ Going Deeper: Ryan also shared that fewer keywords means more weight for each keyword. So if you’re in a highly competitive market, sometimes a better strategy is to use less keywords so that the weight goes to the focus keyword(s). The order matters too, because the algorithm can’t read backward.
💎 Keyword duplication can really hurt your ranks. Keyword repetition is ignored for scoring and the algorithm ignores all instances of the keyword except for the last one. Example: same keyword in title, subtitle and keyword field.
💎 The algorithm remembers, so update your subtitle often. If you get 30k installs per month and remove your subtitle it will remain indexed and become evergreen because Apple will associate those keywords with the subtitle. You can keep updating your subtitle every 4-6 weeks and rotate in keywords to try and update them.
💎 Focus beats performance. Where you put keywords (title, subtitle, keyword field) makes a huge difference and is the most important part of the ranking algorithm (more than new downloads and ratings).
💎 For some apps, not having a subtitle can help with rank: all of the weight of the subtitle goes to the name.
💎 Keywords with low search volume and CVR will hurt visibility. Keywords in the keyword field that get fewer installs will lower an app’s “quality score”. Remove keywords with low search volume and low CVR, even if it means having less than 100 characters if you can’t find other keywords.
⛏️ Going Deeper: I think it’s the first time I hear that not using all 100 characters for the keyword field or not having subtitles might work better. That’s an interesting take, probably worth testing.
💎 Don’t split keyword phrases between name and subtitle, otherwise Apple will weight that keyword phrase to the lowest keyword in the phrase. The name has more weight. Use the full search term in the subtitle.
💎 Eliminate the “stop” words: Apple filters out some keywords altogether because they are too generic.
💎 The first 1.5 screenshots are for conversion, not promotion. You want to use your screenshots to explain the solution, not to sell your product (people don’t want to be sold to).
And before I leave, to sum up Ryan’s presentation:
“Most people completely overthink ASO” - Ryan Kelley
See you next time. Stay savvy!