Lies, Damned Lies, and Statistics
I was looking over some app store metrics tonight, in particular the breakdown of apps by category, and wondered if you could look at the breakdown of the top 500 paid apps by category and determine which categories were over or under served. The results were pretty interesting.
First, I started with this category breakdown over at 148Apps. The breakdown looks something like this.
| Category | Percentage |
|---|---|
| Games | 16.7% |
| Books | 11.8% |
| Entertainment | 10.4% |
| Education | 9.3% |
| Lifestyle | 7.8% |
| Utilities | 6.0% |
| Travel | 5.4% |
| Music | 4.0% |
| Business | 4.0% |
| Reference | 3.7% |
| Sports | 3.1% |
| News | 2.7% |
| Productivity | 2.6% |
| Healthcare & Fitness | 2.4% |
| Photography | 2.0% |
| Finance | 2.0% |
| Navigation | 1.9% |
| Social Networking | 1.8% |
| Medical | 1.8% |
| Weather | 0.4% |
Of course this is a simplification, but if supply and demand were perfectly balanced, we would see the same breakdown by category in the top 500 paid apps. Instead it breaks down like this:
| Category | Percentage | Abs Diff | Rel Diff |
|---|---|---|---|
| Games | 40.2% | +23.5% | +140.3% |
| Entertainment | 9.2% | -1.2% | -11.4% |
| Utilities | 7.8% | +1.8% | +29.7% |
| Photography | 6.8% | +4.8% | +234.4% |
| Productivity | 5.2% | +2.6% | +101.6% |
| Lifestyle | 4.6% | -3.2% | -41.2% |
| Healthcare & Fitness | 4.2% | +1.8% | +74.1% |
| Social Networking | 4.2% | +2.4% | +130.3% |
| Music | 3.8% | -0.2% | -5.9% |
| Education | 2.2% | -7.1% | -76.4% |
| Business | 2.0% | -2.0% | -50.3% |
| Navigation | 1.6% | -0.3% | -14.3% |
| Weather | 1.4% | +1.0% | +235.2% |
| Reference | 1.2% | -2.5% | -67.3% |
| Sports | 1.2% | -1.9% | -61.8% |
| News | 1.2% | -1.5% | -55.8% |
| Books | 1.0% | -10.8% | -91.5% |
| Travel | 1.0% | -4.4% | -81.6% |
| Finance | 0.8% | -1.2% | -59.2% |
| Medical | 0.4% | -1.4% | -77.5% |
Games dominate the makeup of the app store submissions, so you would expect it would occupy a large percentage of the top 500 apps, but I wouldn’t have expected it to be so out of balance. Before really thinking about it, I wouldn’t have put much thought into the game category because on the surface it seems that category is over saturated. But with games representing almost 1/2 of the top 500 and only 1/6th of the submissions, it may not be unreasonable to expect to crack the top 500 with a game. What about some of the other categories? Sorting by relative difference between overall and top 500 representation might make this easier to see.
| Category | Percentage | Abs Diff | Rel Diff |
|---|---|---|---|
| Weather | 1.4% | +1.0% | +235.2% |
| Photography | 6.8% | +4.8% | +234.4% |
| Games | 40.2% | +23.5% | +140.3% |
| Social Networking | 4.2% | +2.4% | +130.3% |
| Productivity | 5.2% | +2.6% | +101.6% |
| Healthcare & Fitness | 4.2% | +1.8% | +74.1% |
| Utilities | 7.8% | +1.8% | +29.7% |
| Music | 3.8% | -0.2% | -5.9% |
| Entertainment | 9.2% | -1.2% | -11.4% |
| Navigation | 1.6% | -0.3% | -14.3% |
| Lifestyle | 4.6% | -3.2% | -41.2% |
| Business | 2.0% | -2.0% | -50.3% |
| News | 1.2% | -1.5% | -55.8% |
| Finance | 0.8% | -1.2% | -59.2% |
| Sports | 1.2% | -1.9% | -61.8% |
| Reference | 1.2% | -2.5% | -67.3% |
| Education | 2.2% | -7.1% | -76.4% |
| Medical | 0.4% | -1.4% | -77.5% |
| Travel | 1.0% | -4.4% | -81.6% |
| Books | 1.0% | -10.8% | -91.5% |
Weather and Photography apps make up 0.4% and 2.6% of apps submitted to the app store, but they represent 1.4% and 6.8% of the top 500 apps. That’s pretty huge. iPhone users love them some weather and Photography apps! Social Networking, Productivity and Healthcare & Fitness also seem to be under represented in the app store.
At the other end of the spectrum are Books. Books make up 11.8% of app store submissions, but only 1.0% of the top 500 fall in that category. The app store is pretty well set on Books. Seriously, no more. Medical is 1.8% of submissions, but is the teeniest sliver of the top 500 at 0.4%. Other categories are over represented are Travel, Education, Reference, and Sports. Turns out the overlap between sports enthusiasts and the typical geeky iPhone user is a bit smallish. Who’d’ve thunk it!?
So the other thing I was curious to know was: suppose all apps are created equally (they’re not, I know, but just for the sake of curiosity let’s suppose they are) and you submitted an app under each category. Using these ratios, what are the chances that your app would crack the top 500? (we’re going to assume that the 41% of free apps is constant across all categories, so we won’t be competing with those.)
| Category | Odds |
|---|---|
| Weather | 1 in 121 |
| Photography | 1 in 121 |
| Games | 1 in 169 |
| Social Networking | 1 in 176 |
| Productivity | 1 in 201 |
| Healthcare & Fitness | 1 in 233 |
| Utilities | 1 in 312 |
| Music | 1 in 430 |
| Entertainment | 1 in 458 |
| Navigation | 1 in 473 |
| Lifestyle | 1 in 689 |
| Business | 1 in 816 |
| News | 1 in 918 |
| Finance | 1 in 993 |
| Sports | 1 in 1,060 |
| Reference | 1 in 1,240 |
| Education | 1 in 1,719 |
| Medical | 1 in 1,803 |
| Travel | 1 in 2,207 |
| Books | 1 in 4,779 |
If you’re looking to crack the top 100, it’s not just 5 times less likely. The category breakdown is slightly different here.
| Category | Odds |
|---|---|
| Photography | 1 in 589 |
| Games | 1 in 652 |
| Weather | 1 in 846 |
| Productivity | 1 in 871 |
| Utilities | 1 in 1,356 |
| Music | 1 in 1,363 |
| Healthcare & Fitness | 1 in 1,629 |
| Navigation | 1 in 1,890 |
| Entertainment | 1 in 3,006 |
| Social Networking | 1 in 3,695 |
| Lifestyle | 1 in 3,963 |
| News | 1 in 5,505 |
| Sports | 1 in 6,361 |
| Reference | 1 in 7,442 |
| Business | 1 in 18,911 |
| Finance | 1 in ∞ |
| Education | 1 in ∞ |
| Medical | 1 in ∞ |
| Travel | 1 in ∞ |
| Books | 1 in ∞ |
Notice, for example, how Social Networking is 4th easiest to break into the top 500, but drops down to 10th position when it comes to the top 100. Social Networking is popular, but not so much top 100 popular
Anyway, I realize that I’ve vastly oversimplified this analysis, but I think some of the generalizations are probably accurate. And besides, I was mostly just having fun playing around with numbers and thought I’d share. Maybe someone else will find them interesting as well.
Is any part of this data massaged or fabricated?
I ask because if the data set is 500 apps, there should be no odd decimal points. 1/500 = 0.2%
I found it interesting, for sure. Thanks!
There are no odd decimal points – except in the first table, where the data set is not 500 apps.
@dr no, no massaging. I can’t say there is no chance of an honest mistake though. Also, the first column of the first two tables are the breakdown of the top 500 and there are no odd decimal points.
dr: I was wondering that. If 16.7% of the top 500 are games, that means that there are 83.5 games in the top 500. How on Earth did that .5 get there?
201 of the top 500 are games = 40.2%. 81,650 of all 488,152 submissions are games = 16.7%.
@dr The first table is percentage of all applications in the store.
Interesting, but how many apps is a person likely to buy? For instance, one weather app is all I need but I don’t see why I wouldn’t buy a bunch of games. Same for books (although I don’t like books as apps).
So, weather may look like a promising money maker but perhaps it’s because the market is already saturated. Games don’t really seem to saturate because they are used for a period of time, become stale and are put aside, only to be replaced by another game.
So I don’t really know how much your statistics are telling us. Interesting none the less.
John, that top section that you reference is all apps listed on 148apps.com I believe. The second chart is the top 500 which, as Casey points out, has no odd decimal points.
Nice! One fairly simple but interesting element to the “how likely are you to make the top 500/100” question would be how volatile each category is, or how entrenched incumbent apps are. More volatility might make it easier to break through, but also harder to stick around.
Look at the source. He’s using the ratios from a much larger dataset to apply to the 500-app numbers. The 16.7% doesn’t come from “Top Apps”, but from all apps.
I suppose paid apps means free apps with in-app purchases (to remove publicity, add features, etc) aren’t taken into account. Right?
That model is pretty popular and seems to work quite well nowadays. It would certainly be interesting to account for these apps, maybe in a distinct breakdown.
Anyway, thanks for your work and for sharing it. It’s definitely interesting.
Looking at the highest grossing apps in the UK store – another really interesting thing is that it’s not just games, but *kids* games that are consistently in the top 10/25. And the ones that heavily rely on in app purchases are high their (animal zoo’s, Smurfs etc regularly beat out TomTom and others for the most profitable apps in the UK).
Worth looking into the straight app price, vs in app purchase level as free/lower priced apps are making a mint from in app purchases, presumably on an adult’s credit/debit card.
Is the goal of the app makers to make money or crack the top 500. I say this because when you say don’t make any more medical apps. Medical apps are generally more expensive apps. Obviously there not gonna sell as many. You should look at top 500 paid apps by revenue
I am not sure I understand exactly where you were going with this, but in the end, if you are trying to make money at this, shouldn’t you be looking at gross revenue rankings? I don’t think that is what your were looking at in the paid lists.
I can say from experience, that most $0.99 apps have a hard time making much money in the end, where as a quality, mid-ranked $4.99-9.99 app tends to keep itself in the top 10 of a category and makes much more revenue than a low priced app that has to sell 5-10x just to match the revenue.
It would be interesting how the distribution lays out when looking at gross revenue rankings, eh?
re: pj
I have no idea how other people consume apps, but I have several weather apps. I’m still searching for the perfect one. I also have several photo apps, each of which gets used in different situations. Most apps are rather inexpensive, so this is a platform that encourages experimentation in the form of downloading/purchasing apps, rather than sticking to one or two apps per category.
By contrast, I only have a handful of games, and tend to come back to the same 3 or 4 over and over again. App store users are probably as varied in their downloading and purchasing habits as Mac or Windows users, if not more so. All of which makes generalizing user behavior a difficult exercise.
RE: how many apps to we need. I have 4 weather apps already – because sometimes I got a free one first, then later found out about better paid apps, or got specialized weather radio, or app that displays weather on the badge etc. Same with other categories, how many flashlights do you need? I tried maybe 5 before I found the one I like best and use. Alarm clocks? I have at least 5 and have stuck with one that is my favorite.
I have all the apps on my iTunes, and if I see an update for one I already have, can see what the improvements are and if any better than the one I currently have synced with my phone.
I have about 750 apps downloaded, about 250 currently on the phone. That is including a lot of travel apps that I only sync if going to a particular region, and various games that I played and removed later, and some business apps that I only leave on as needed.
I have used like the same two weather apps and sports-score apps since my iPhone 1st gen. Weather is weather – not much changes in how to present it. Sports scores are easy to present as well, and most apps simply do the same thing as all the others. I added one single weather app when we moved to tornado country so I could get NOAA alerts. Thats it, three weather apps in our entire iOS tenure (over 6 devices).
By contrast, my family has purchased well over 200 games. Some we play for a while then get rid of. Some we come back to. Some are just for kids, some for adults. Some suck and get deleted right away.
And finally, I will pay $.99 for just about any app. At $2.99 I start to think about it. At $4.99 it has to be something worthwhile. At $9.99 is better damn well meet a specific need.
I have purchased one $20 business app for work and even though I use it infrequently, when I do it more than makes up for it’s $20 cost. I have yet to find an iOS game worth $5.
I have a family member who makes a very specialized iOS app for the medical industry. Their app costs hundreds of dollars – and has an on-going support and subscription cost as well but is highly purchased and used because it is very specialized to meet a specific need.
This wasn’t really intended to be a very rigorous study, rather just playing around with numbers to see if anything jumped out at me. For the most part I’m not drawing a lot of conclusions from it (except the book category!). Although it could perhaps be useful to reinforce (or dismiss) a theories proposed elsewhere.
I agree that it would be useful to look again using the top grossing apps, and I very nearly added that as well, but it was getting late so I went with what I had. Perhaps I’ll take another look this evening.
Also, there is a lot of volatility in the app store and a more accurate picture could be drawn from monitoring the various categories over say a week or month and taking the average composition over that time. I didn’t feel like waiting a week or a month, so I went with a snapshot using yesterday’s data. Indeed I did a similar analysis a year ago and there were only 2 weather apps in the top 500 at that time versus 7 yesterday.
I relied primarily on appannie.com to gather this information, but also checked a few items over at appshopper.com.
Lastly, thank you all for the feedback (and @gruber for the link). I wasn’t really sure that anyone would notice this post.
I’d look at revenue. That would be more interesting to me. Where am I likely to make money not where are people likely to download (and perhaps ignore) my free program.
An interesting post. Might be an idea to do the same with just iPad apps as that is a different ball game. I created a travel app which shot into the top 100 (and was featured by Apple) yet actual number of downloads was quite small – it seems that getting to the top of the iPad charts would be a lot easier (as there are obviously a lot less apps). Might be interesting to look at how that affects Universal apps though…
And, of course, all things are *not* equal. Take for instance one App (which I won’t name) that is being given incredibly large amounts of exposure on local Hawaiian radio stations owned by Pacific Radio Group. The application sucks. UI is decent, overall, but the content is sorely lacking and performance sucks.
How can a *free* App (a travel guide) possibly be so heavily advertised? Must cost a fortune for all that air time!
Not.
Turns out (I did some research) the guy who developed the App (Rich Ringrose) is connected with the radio station/media group that owns a crap-load of radio stations, and *they* are promoting the hell out of this App.
In theory (and, again, this is an exception), some Apps *might* break the Top 500 or even 100, but not because the “odds” are in their favor based on category, or the excellence of the App, but because somebody with deep pockets or the means decides to promote it at any cost. Just my two pennies’ worth.
As a side note, this Rich Ringrose guy says that developers should start with Android because its users have lower standards and more tolerance for “bugs” than iOS users. Essentially, he’s saying: iOS users *expect* your App to be perfect and Android users do not. Quoting: “there’s no margin of error from Apple and the users standpoint.”
Amen.
Out of curiosity, where did you find your sample of the top 500 apps? I’ve been looking all over for such a thing, all I ever found was top 200.
AppAnnie will give you a list of the top 300 in each category. You can inspect apps individually to see where they rank in the top 1000. It’s fantastically tedious and if I were in any way clever I’d figure a way to automate the process.
Great analysis – I guess its a good thing that I’m currently working on a photography app 🙂
Follow up post here which takes a look at the top grossing categories.
One possible explanation being… that the gold rush in 2009 attracted so many developers into making iPhone games, and now we have so many shitty games in the App Store and not many of them can compete with the useful, productive apps to get into top 100. 😀
There are a lot of books because they’re (relatively) easy to make. You just make the app once, and then submit it multiple times with different content.
Yeah paid apps by revenue would be interesting. Would also love to see a games category breakdown – which game genres are generating the most revenue.
What about the decimals?