The fundamental question regarding mobile app featuring is: how can we measure it? Featuring may be a significant driver of an app’s success, so we need a deterministic way to measure it, same as any other key metric.
When it comes to your own apps – it’s not rocket science: you simply measure the downloads and the revenue uplifts. The proper analyst will do it easily. And if he knows how to predict LTV for your apps considering the traffic quality, your apps’ k-factor, he’ll provide you with the best estimation you can get: how much monetary value the featuring brought.
But what if you need to investigate featuring of others’ apps?
I’m sure that for many of you it’s unclear why we might need it. So I tried to make up the list of the main general reasons:
- What is the actual role of featuring for your business? Can featuring be considered one of the pillars of your business’ economy?
A clear answer to this question may lead you to some changes in you business-plans and priorities:
I constantly meet people that believe that proper platform relations and, as a result, featuring – may turn their business from unprofitable to sustainably profitable. Some other people claim that featuring today is a neglectable factor. So who is right? Can they both be right? We’ll try to sort it out.
- The second question is: what featuring support do your competitors receive? How can we measure it?
To answer this question we need to find a way to effectively monitor and measure featuring cases of competitors’ apps, and the apps of your own. Our tool solves this task totally.
- What are the best ways to maximize platforms’ support? What is the reasonable amount of effort that is worth investing in relationships with platforms?
Sure, our tool is not a magic crystal ball to answer questions like this. But answering the first question about the role of featuring for your business may help to better answer this particular question.
Okay, let’s discuss the method we suggest for featuring measurement. The challenge of measuring featuring is mainly complicated by the fact that being featured often implies having a lot of simultaneous featuring placements on different platforms in different countries and via an unclear number of feature types. Here’s a typical picture of a random large-scale featuring:
On the 6th of October Angry Birds 2, had 867 different featuring placements. Impressive, isn’t it?
How can we estimate the cumulative power of all these placements?
We clearly understand that the amount of installs and revenue that featuring may generate depends on many factors, and first of all, on the app itself. We came up with an approach and find it the most objective and generally applicable: we measure a downloads uplift that every unique featuring placement brings in average, combine it with the data about average LTV for each country and each platform, and also take into consideration the duration of a featuring case (the impact of saturation effect).
Saturation effect means that the longer featuring lasts, the less impact (downloads and therefore revenue) it generates day by day. The daily effect of everlast featuring tends to zero.
Let me give you a simple example: let’s assume that a topmost featuring banner brings in Brazil twice as many users as it brings in the USA. The same time we know that in average a Brazilian user brings 20 times less revenue than a US user. Combining both numbers, we estimate that having that topmost banner in the USA is on average 10 times more useful, than in Brazil, so the featuring score that we show for the USA will be 10 times larger than the Brazilian one. Makes sense?
The real math is way more complicated, but the approach is the same. I want to emphasize it ones more: we aim to predict neither the amount of money nor the number of downloads that featuring generates. What we do is calculate an abstract numeric score that shows the level of support that an apps get from platforms, the desire-ness of the given featuring placements. That approach, in turn, allows us to apply math and perform analysis of featuring of any kind: calculate total/average/median or whatever featuring score for an app, or for a group of apps —and do it across any geo regions, time periods and platforms.
For example, here is what a top featured publishers chart looks like on DataMagic.rocks. We see the total featuring score for the whole September 2018, for all the apps of each publisher. So now you understand what super-large publisher score in our system looks like. Sure thing, you are able to select any geographic region, store or time period at this screen.
Also, we have made it possible to investigate featuring history of any particular app.
Have a look at the example of how can we check out the by-country decomposition of featuring support an app had.
Important note: we don’t analyze personalized featuring of both stores, as it’s a completely different story and does not relate to the intentionally managed featuring in any way. However, “Recommended for you”-like sections definitely make up a significant part of organic traffic, nevertheless we cannot estimate it as we know nothing about apps that get featured there: it’s a totally personalized story.
Though we do know that there are platforms’ guidelines on how to make your app appear in these personalized collections more often.
Featuring distribution disproportion.
Our team prepared several examples of featuring analysis. Here is the most complex and interesting one.
We decided to find out: do platforms distribute their featuring capabilities across genres fairly enough (according to their market share)? Is there any significant disproportion? To give a better understanding of what I call disproportion, let’s assume (with randomly made up numbers) that racing games get 5% of all iPhone downloads in China; meanwhile they get 15% of all the featuring capabilities of the iPhone store in China. That’s what we call a 3x positive disproportion. Or vice versa: if a genre gets 15% of all downloads in a country and gets only 5% of featuring power, we call it a 3x negative disproportion. But why do we rely on downloads? Why is it not a revenue-based calculation? Let’s check both downloads-based and revenue-based versions of this analysis.
Let’s start from iOS.
The size of a circle means its featuring share in every particular country. The color stands for disproportion. Yellow is 1 (no disproportion), dark-red means negative disproportion of 5 and more. Merry-green means positive disproportion of 5 and more.
The sum of featuring-shares for each country equals 1. Same for downloads-shares and revenue-shares for all the graphs that we will see. Actually, it’s a bit less than 1 as we only selected 20 of the most significant genres, so a minor part of shares isn’t present.
Here we can see that some genres seem to have a great negative disproportion when we base on revenue: Dating, Casino, MMO Strategy suffer from featuring distribution disproportion worldwide.
To be honest, it’s not quite clear for me why Dating apps and MMO Strategies suffer that much from disproportionate distribution of featuring, but it’s the fact for both platforms. The conversion rate might be the reason (very bad conversions from a featuring banner views to installs).
We also see some evergreen columns such as “Hypercasual” and “Action: arcade” as these genres generate a miserable share of revenue but receive quite a significant share of featuring. And it’s surprised me: what economical reasons make iOS do that? COnsidering the fact that these games actually DO have significant revenue, but it’s not iAP revenue (we only consider iAP revenue on our graphs), its ad revenue that totally passes the platform’s bank accounts.
Also we see some big green circles that mean that some genres in some countries get totally disproportionate amount of featuring support. For example, MOBAs have positive disproportion approximately equal to 7 in Germany, UK, France. I dunno why, the editorial team guys in these countries must be sincerely in-love with this genre. I don’t think that the conversion rate for MOBAs in Western Europe is better than it is for MMO Strategies. But, love is love.
MOBAs in Europe
I won’t list all the interesting cases now: you’ll be able to look through this research later if you want, I’ll provide you with a link at the end of my speech.
If we base our disproportion calculations on downloads, it looks a little bit more fair (less bloody-red and juicy-green colors), but most of the same cursed genres still look red: Dating apps, Casino games and slots. MMO Strategies look a bit better, but it’s explained by a very small amount of downloads-share these games have (0.5%-1%). Also now we see red circles for hypercasual games as they make up a huge amount of downloads in accordance with their nature and despite the fact that they get great support from platforms the disproportion is still negative.
We also see a red column for “Messengers and Communication” apps. But the reason behind redness of these apps may be different. The thing is that some of the most famous apps like Facebook, WhatsApp and Messenger get featured all the time, people see ‘em each time they open the store, so the “saturation” effect makes the Featuring score for these apps much much lower than it could be. We’ll talk about it a little bit later.
Let’s have a look at the Google Play Store now.
We see that the situation doesn’t differ a lot, though MMO Strategies seem to suffer less. We also see some giant circles of ultimately popular and revenue-generating genres in Asia (Party battlers and Action/MMORPGs). It means Google Play editorial teams do more for those who bring revenue, though the disproportion remains negative when we base our analysis on revenue.
When we look at downloads-based graph it becomes even more clear: Google gives a bit more help to those who bring revenue. It’s understandable.
So, now we got closer to answering the question that we had in the very beginning: “What is the actual role of featuring for your business? Can featuring be considered one of the pillars of your business’ economy?” – it highly depends on the genre and the countries you bet on. And sure, it depends on your business’ burnrate. And some other factors that we can’t discuss now as I only have half a hour.
Featuring score – is a tool that we give you to help understand platforms’ featuring policies, the benefits featuring may bring to your business. Then, you are to make decisions.
That’s it for the main topic. But I’d love to tell about some interesting findings that we came across while preparing this talk.
Privileges for the largest publishers
[Slide: Privileges to the largest publishers]
We’ve done research regarding this: do the biggest publishers receive a lion’s share of featuring? To cut it short: it’s totally so on Google Play:
Monthly publisher score here is an average cross-platform monthly publisher’s score in 2018. Publisher score is our made-up metric that equals to a publishers revenue (excluding platform fees and inclusive taxes) + downloads for the specified period. So in the second column, we see an average monthly sum of revenue and downloads for 2018. Those who get 20M and more publisher score a month I marked green. Those who get in the average 1M score and less per month, I marked red. All others have colours in between red and green according to their publisher score.
I want to underline it here: we take into consideration the “saturation effect”, it hugely changes the picture! A lot of highly visible placements are held by apps that have been holding these placements for ages and therefore have weak featuring score according to our system.
The things are quite different when it comes to iOS. On the very first page, we see a lot of red and orange lines. Sometimes it gives rise to the question: “How the hell did these companies get that huge featuring?”, but we’ll discuss it in the next sub-topic.
The second page looks quite like the first one.
What does it all tells us? Google definitely supports those companies that bring revenue. This logic is easy to understand. What does Apple support? We can’t answer this question yet. But we’ll return to this topic later.
“Hung” featuring phenomenon
We’ve already described what saturation effect is. So, assuming that platforms are interested in advancing their stores’ explorab-ility, platforms should want to rotate featuring placements as quick as possible to maximize the novelty effect and the mentioned explorab-ility. However, we noticed a significant amount of apps keeping their strong featuring placements for months!
Here you see that in the Days column all the values are equal to 62. It’s because we’ve started collecting information about featuring quite recently. So as we started collecting it all these apps have already been featured and haven’t changed their main featuring placements for more than 2 months.
The “effectiveness” jumps up and down because some apps sometimes get rotated between some additional placements, that lead to effectiveness growth. But the main placement remains stable all the time.
If we compare the overall platform featuring score considering the “saturation effect” and without it, we’ll see that “saturation effect”, in other words, the low rotation speed of apps between placements fades away 80% of platform featuring capabilities for iOS and 89% for Android.
Phenomena of super-featuring of super-low-performers
We also noticed a number of very weak (in terms of metrics) apps getting outstanding featuring. The regular case looks like this: a no-name publisher has been doing poorly, having something like 5K downloads and $5K in revenue per month for a long period of time. A tiny publisher, no better than the others.
All of a sudden, it gets outstanding featuring: best placements in the largest countries: USA, China, Japan. And as a result of this featuring the publisher gets something like 10K installs and $10K revenue that are miserably small numbers considering the featuring scale. It’s hard to imagine that platforms had expected some great performance from these apps as they had seen their metrics before the featuring. So what could be a reason for these phenomena? ….. I believe many of you would love to have that key to the hearts of editorial team members.
Here we set our filters to show the most unsuccessful apps. But one can easily find cases when a totally average app gets a 2M featuring score. There are a lot of cases like this on iOS. By the way, we see that the vast majority of the given cases refer to paid apps (they have zeros in the downloads column as we don’t estimate the number of downloads for paid games at all as we’ve always been considering this segment of mobile market to be very tiny; Apple seems to think different.
This phenomenon is actually somewhat upsetting for a number of developers who invested a huge amount of effort to create great products and received nearly no support from platforms. It’s, actually, all the cases we could find using the same filter thresholds, as we used for the iOS’s table. If one digs deeper he can find some apps that had 1M-2M featuring score support, but there are only a few of them and they are mostly unsuccessful launches by major publishers.
So, here the behaviours of the platforms differ significantly.
We only started collecting featuring data in July 2018. So we don’t have a long history of featuring yet, and our estimations of featuring scores are a bit rough. But as we collect data, the estimates’ accuracy progresses day by day.
DataMagic has made this tool free to use. So, please, everyone: you’re welcome to try it! Research the market, watch the competitors, make better business decisions. And enjoy the magic!