The Wrong Way to ASO: Test in Google Experiments & Implement in iOS
The final hurdle every user must cross on their journey to install always sits with creative assets. No matter their journey to your product page (if they even make it to your product page), their final decision is made whilst looking at your assets.
I like to think of an app’s product page as the book you write for your users. Your icon is your book cover and every asset, each screenshot and each video, are the chapters in your story. And just like any good book, the order (and focus, skim readers, we’re looking at you) in which you read the chapters is essential to your understanding of the narrative. And integral to your desire to keep reading and keep engaging.
No great story was written perfectly in one go. Ask any writer and they will tell you that first drafts are relatively easy but editing is hard. Refinement is harder. And perfection takes a team of editors and roping in your most trusted friends and colleagues as draft readers.
When it comes to your product page story, you can do even better than an editorial committee; you can do live market research with real users. How do you know what to write in those asset chapters and how you should structure and order them? You test.
And Google Experiments gives Android developers and publishers the power to do just that built into the platform itself. Google Experiments is free which means you can glean a great deal of data about user preferences for ‘merely’ the opportunity cost of maybe a few bad hypotheses. Great. And now that you have data about user preferences, why not simply implement that in the App Store too? Great.
Um, not quite. This is a bad idea. A very bad idea.
The reasoning is simple and two-fold and goes back to the story you wish to tell your users: different audiences have different preferences (they like different stories) and different layouts demand different user journeys (the order of the chapters are important too).
iOS and Android users are inherently different
This goes without saying. But more than just the ‘duh’ factor is that the audiences are actually very interesting in their differences. It all boils down to the fact that there is only one company in the world offering iOS smartphones: Apple. This means that App Store users all fall under one company’s target consumer block as opposed to the dozens (if not hundreds) of companies offering Android-based devices. In a word? iOS audiences are homogenous.
Looking at the smartphone market in 2018, we see that iOS users represent less than 15% of new smartphone sales. The vast majority of the world uses Android (which would be around 85%) and the distribution of these users varies widely between geos and countries.
The ‘vast majority of the world’ is hugely varied and incredibly diverse. App Store users, on the other hand, tend to be more Western, more wealthy, and more willing to spend.
Looking at the distribution between iOS and Android by country, we see this dynamic clearly play out across the world.
A few of the notable iOS-heavy markets include Apple’s home base of the US, Canada, Norway, the UK, Australia, and Switzerland. Non-Western but wealthy, developed countries include Japan and Saudi Arabia.
These (mainly) culturally-Western developed countries are characterized by higher income and higher socioeconomic status. This is unsurprising. At the end of the day, iPhones are a high-end product which carry a steep premium in the marketplace. (It’s also often tied to conspicuous consumption but that’s a whole different analysis.)
It’s safe to assume that as a high-end product, regardless of the local culture, the higher up a certain society is on the socio-economic ladder, the more likely the outcome that iOS will have a greater market share. However, it’s important to remember that just because most iPhone users are wealthy does not mean that all wealthy people are iPhone users.
According to a US survey, in general,
- iPhone users earn 43.7% more than Android users
- iPhone users spend around double on tech than Android users do
- iPhone users spend 31.8% more time on their phones each day than Android users
On top of that we can see that even though Android has a higher share of global downloads, iOS far outweighs Android in its share of consumer spending.
It’s clear the iOS and Android audiences are very different but is it clear that you should communicate with different audiences differently? I certainly hope so. Remember that book analogy? Well, you wouldn’t market Fifty Shades of Grey to the same audience you would War & Peace, would you? You also wouldn’t market War & Peace the same way to an audience of high schoolers forced to read it as part of their curriculum as you would a geriatric classicist. Marketing is all about finding the right message for the right audience.
If you know your audience likes Fifty Shades of Grey, the messaging you’d use would target them specifically. The phraseology, design style, and overall messaging would be focused on your specific app’s audience’s likes and dislikes to really raise conversions by giving them what they want to see.
Android audiences, on the other hand, are a complete mixed bag. They like Fifty Shades, War & Peace, Archie Comics, The New Yorker, John Grisham, and Terry Pratchett. Not to mention those that don’t even read at all. Finding the common denominator that converts well for the whole audience is tricky and Google Experiments is certainly a great first step. But the message that works for that whole entire mixed-bag audience? You can definitely find a more refined and targeted messaging strategy for the iOS audience.
Think about your app’s specific audience and its platform-specific audiences. A ride-sharing app might find that Google Experiments directed them to creatives that focused a messaging strategy around cost-saving, whilst an iOS test might find users were more interested in benefits like convenience and comfort.
But how can you convey that message and tell that story in a way that’s most appealing? By carefully structuring the plot to make it most enticing.
Different layouts lead to different user’s journey
What assets users see, the chapters they read on your product page, are laid out differently in each store and how users move through each store is different too. Thus the way they consume the story you want to tell them is different and that consumption will affect their internalization.
We already know that iOS and Android audiences are different and we know that the product pages look different because they contain different elements in different configurations. See, they’re different.
Just from looking at the different product pages you can see that certain assets are more prominent in the different stores. Two recent(ish) developments have further differentiated how users navigate each app store. Since the Google Play redesign came into effect last year, it has had a major impact on how users navigate the page and significantly increased the amount of decisive users, those that make a decision immediately after exposure to the first impression alone. And for Apple, since iOS11, the advent of autoplay has vastly increased the number of users who view the video content, as opposed to just being exposed to the poster frame.
Just because an asset set wins on Google Experiments doesn’t mean it will translate to iOS because iOS users explore and view more assets than Android users. 33% more iOS users search through more than just the first impression, 51% more scroll the page vertically and a whopping 145% more scroll horizontally through the gallery. This means that weaker screenshots that sit in the latter parts of the gallery could sneak into a winning experiment simply because not very many users (less than 27%) were exposed to them in the initial test.
Based on data from thousands of iOS and Android tests, we know that users arrive at their install/drop decision in very different ways, having consumed (or not consumed) different asset sets. For example, if you’re testing on Google Experiments a new screenshot gallery it might show that the first two screenshots were able to convert better. Even assuming iOS users would have the same preference towards the messaging and creatives in those first two screenshots (which we showed is a stretch due to differing user preferences but for the sake of the argument, go with us), the iOS user would come to a decision based off all (or more) of the latter screenshots in the gallery.
It’s obvious that not all users read all chapters and you need to make sure your product page story (and the story you want to tell) is still conveyed to your readers, I mean, users. Google Experiment tests for the bare bones version of your product page story; think of it as the novella to the novel. Those who like the novella might not like the novel.
Applying GE results in iOS is problematic, we get it. But how problematic is it really?
We ran the numbers to see exactly. An in-depth analysis of pre/post applications of Google Experiment winners on iOS showed us clearly that the potential impact on the conversion rate for both paid and organic traffic can range between 10%-15%. Negatively. But it’s about more than just CVR; the opportunity cost is high as well. Google Experiments runs in the live environment which means that real users are being sent to the test and the reduced installs of the losing variations reflect the real installs and subsequently lower conversion rates. The cost of running Google Experiments is the opportunity cost of lost users. The other side of that coin is the significant growth that is being left on the floor any time you use unoptimized creatives. The point of optimization is to capture as many users as possible and users that pass through, read your product page story and then decide ‘nah’ (when a different version of the same story would have worked) is a floor leftover. And there’s no 3-second rule in the app store.
Data is only effective if it’s specific. Data sourced from one group is only useful for that one specific group. You can base further research and testing on those insights but you cannot implement them as is. And the risk of planning strategy and making decisions off bad data is exponentially high. Successful ASO managers understand that a core element of success is treating Google Play and the App Store as what they are, different audiences.
As the saying goes: “’Different strokes for different folks”.