Fascinating visual projections of the iPhone apps market data
Truly fascinating: both the data and its visual representation by Trefis. via theNextWeb [if you read via RSS, we recommend you jump to the our blog]
AppsFire.com ★ blog ★
discover and share apps on ★ fire ★
Filed underdata
Developers, don't panic! Not 80% of iPhone apps are sucking
Wow! we did not anticipate that our distribution graph would generate such a huge echo on the media and blogosphere - from ReadWriteWeb to PCWorld. We felt compelled to follow up with some thoughts.
We'd like to bring some clarity on the confusion that our data has spread over many blog readers out there
- The 80%/20% rule is pretty universal to start with, and it’s no surprise to find it here again
- 80% does NOT correspond to the number of apps that are not downloaded, but rather corresponds to the number of apps either not downloaded nor kept on the iPhone (eg i download something and uninstalled it). The difference is important (although we suspect not huge in %)
- It does not mean 80% of the apps are Crapps (self expression), or unused (we don t measure usage) it means that most of the apps are not found.
- Does that mean you can't make a buck by being in the long tail? No. For example we have discovered that with the right pricing some apps with a few thousands install can make some very nice money. It depends what your ambition is, and how useful your app is
- Because it will always come down to that: a useless app has indeed very little chance to make it. And in those 80% there are some. let's not be blind. Users can spot them and sanction them
- Useful yet poorly marketed apps will also fail : wrong pricing, wrong merchandising, wrong promotion, wrong pre-appstore teasing....
- Timeline: we don't have the data and maybe other app analytics company (or Apple) could help with that, but in those 80% some apps have been launched and either not updated, or simply discontinued. We're not sure which part of the cake we're talking about, but definitely some apps are counted in cumulated and should not be included. Let’s not forget that a bunch of pre-3.0 apps no longer work on most devices; apps that were already in the long tail in the 2.x days have sank deeper by now.
- Actually what really matters is the number of apps that are actively maintained. It is definitely NOT 100k
- Our sample is composed of a few thousands users: yes we are a young service, and yes our sample is biased. We fully disclose that fact. But bear in mind that most if not all iPhone research that are published out there are based on similar if not smaller sample size that are extrapolated. Here we save the extrapolation because we give you 100% real data of what is actually installed on people's iPhone. (and not the apps that are sitting idle in the iTunes library)
- We count on the intelligence and sense of judgement from readers and bloggers to understand the context of that data. Like we said in a comment to one of our reader: what matters is not so much the data at a precise moment, but the trend over time.
- This is why we 'll publish regularly an update of this data set so we'll see how right,or wrong we were.
- Btw the graph looks a bit different if you double click on it by category. We'll release soon data going in that direction.
- Our observation is that as we grow and progress it is harder to detect new apps installed and we are confident that the 20/80 rule will apply in the future
- What matter is the rotation of those 20k. It is easy to understand how this will change over time. New great apps will come and replace some others.
Bottom Line: While a few posts using our data conclude that it is not worth trying your way in the app store, we believe this is precisely the opposite. In fact, more than ever, there is a growing need for great if not awesome apps. Plenty of room there! As far as we are concerned, there is a massive potential to re establish some equilibrium with the way people find and install apps (worth having). A few startups and big companies are currently working on solving that issue.
AppsFire is one of them.
Short thought: Are free apps really free?
What a stupid question. no? Well maybe not. If you got an app for free from the App Store, you probably noticed that many of them serve advertising. Like in the example below
Now if you are running your app on a data network (which is not wifi), those ads are served dynamically via the internet, meaning live.
This simply means that by running a free app, you are using your data package. Many times this does not really matter because the app requires you to connect to the web in order to be used (for example Last.fm), so the incremental usage of that free app is not really critical.
But if you thought your app would run locally and that could use for free, you're wrong. A free app with advertising will use your data package.
Wait! if you have an unlimited data package, then of course you're safe and you probably won't mind. But if not: mind your bills
The only problem, which is more general today, is that there is no way to know what apps are consuming in terms of data. That would be a tremendous information
For which either Apple, or Mobile operators will have to open their data a little more.
What they should do is either allow developers to have access to that information (splitting generated by the app and usage generated from the ad serving) or give general access directly to the user. Making money from advertising is a great thing, but consumers should know what it cost them.
For now mobile ad formats are rather simple and static, but the day those formats will become richer (like on the computer) the data required to serve them will become bigger.
The same way we know from our operators which phone number generated minutes of call (and therefore part of our bill), we should get the same from apps.
Agree?
Looking deeper at the ★ Fart ★ Factor
One of the striking thing to observe in the App Store is the number of Apps that are mmm how to say it...Fart related
The interesting fact to observe is that a very basic app (not sure which one really kicked the movement) started off a real wave of apps from games, to music playing including the Fart factor. And lots of them are paid apps! [and very lucrative ones]
We were curious to find within our user base how this was spreading. We found that 5% have at least 1 app related to Farting. When you think about it this is a lot and there is something probably deep to analyze about this kind of self expression that has a strong social element too (hard to imagine you can enjoy this for yourself only). Since our database in biased we imagine that the penetration ratio of this category is much high among mainstream users.
This is probably a category Apple won't help you discover. But we decided to create an AppMix below to give you an overview of the type of Apps that you can find in this "Category"
We just hope that your iPhone screen really do not look like that
credit picture
Facebook vs Twitter on the iPhone: Compete says something different. Why?
Compete lead a research about how users use their smartphone. There is some interesting data in that report. Some of which we questioned (if this is the same report).
Another piece of information that stands out is that about 70% of users use Facebook from their iPhone, which is pretty comparable to our own research (to be precise % of iPhone owners who have installed the iPhone app).
Facebook is hot among iPhone owners: 71% of iPhone users report accessing Facebook from their mobile device, 37% listed Facebook as one of their top three most utilized apps and 18% claim it's their favorite app.
However Compete declares that only 26% only use Twitter from the iPhone. To be honest this is a little surprising and if the method is declarative i think it is worth digging a little deeper.
From our report 60+% have installed a Twitter app on their iPhone and from our declarative poll close to 65% (more than 1200 respondents) use Twitter from their iPhone (more than Facebook)
So what is it going to be? The answer has probably to do with the way the sample was built (Compete sample was made of 100+ respondents who own an iPhone). Our sample is probably biased by the nature and age of our service, but we believe the biais is not that strong (confirmed by the fact many of our stats match others reports built with serious methodology).
I'd love to hear more from Compete on that one.
Twitter wins big time in usage on Facebook from the iPhone
A little update on our research. We issued a poll to understand if you guys are using more twitter or facebook from the iPhone. So far a little more than 700 respondents have participated.
The result is crystal clear: iPhone owners use more Twitter than Facebook on Their iPhone: 69% are using either more Twitter or only Twitter.
This is of course a small sample (although 700 answers start to have some meaning). What would be really interesting, would be the actually aggregated data usage for both services provided by ATT (or any national operator in deal with Apple). They probably know a lot more...
update 24/09/09: we passed the 1000 respondents. The trend is the same. One reader added a smart comment: since Twitter status can be synced with Facebook or some apps enable double publishing the Twitter usage and Facebook usage could be correlated but the prime access goes to Twitter
Compete: Most iPhone owners have 10+ apps... Yeah right..
According to a declarative survey on 104 iphone owners Compete find that more than 70% of iPhone owners download less than 20 apps.
Numbers don't add up. This means that the large majority of iPhone owners has downloaded at best 700m apps. When the reality is that more than 1.5 billion apps have been downloaded.
Their sample is very small (104) and the methodology is declarative (who really remembers how many apps are downloaded).
Our own report (indicating that on average an iPhone owner has 65 apps) is maybe not perfect and has an obvious bias, but it was based on a real picture (1200 users). And Compete's survey is i believe far from bringing a true market picture. (what about also paid vs free apps? )
In our next release we ll publish a distribution tree that will bring a more accurate picture.
ReadWriteWeb has more
Check it: Some ★ important precisions ★ and complementary stats from our report
Given the heat our report has generated (btw Top presentation of day on SlideShare...), we thought it would be important, before some readers get a heart attack, to publish some complementary precisions and also unreleased data. This was our first shot. And first shots are what they are. First shots. So here are some important things to remember.
On the average amount spent per device
The data that was mostly discussed was the average spent per device: $80
Many readers/iPhone owners do not identify themselves with this number. Well, let's look closer..
This was calculated by dividing the cumulated value of all paid apps that we found on the device by the number of devices that we track. This gives a simple picture but does not help get the reality of the how different users spend really, and indeed, there is a bias from the Appsfire that we suppose is made of iPhone enthusiasts and power users. However, note that we do not track apps that have been uninstalled, so the bias is perhaps neutralized in some ways...
To be totally precise we should have published a distributed graph of value spent (how many have spent more than X and less than Y). We'll do that in our next edition. However digging deeper we found that one of the most "representative" value spent per device is $45, by trimming away the outliers and looking at median figures. Some users spend lot more and some lot less.
On the average price of paid apps
The $1.56 average price per application that we mentioned was the average price across the 15k+ apps that we track; it should be noted that free apps were included. Again, simple average. To make the picture clearer, excluding free apps from this calculation yields an average price of $2.87 per 'paid' app.
By popular request, we dove into the data to find out the average prices spent on a per device basis (within our Appsfire dataset). And this is what we found: average price per application (all included) is $1.31 while average price per 'paid' app is $3.46.
Bottom line
Like we said in our previous post, this data will gain depth and meaningfulness over time especially as our user base grows.
------
For the record, here is the original report
How about some ★ real data ★ on iPhone owners?
update: read also those complementary results and precisions
Today we are releasing our first "State of the App Store" market insights, specifically about Apps and iPhone metrics many companies, journalists, developer and bloggers are curious to know.
And it has a unique approach compared to the many market research you have read here and there. Why? Because our research is actually based on the analysis of what is really installed in someone's iPhone/iPod. Ad Mob for example released a great piece of research based on a sample of 1000 iphone owners. But this research is based on declarative data. Flurry has also some great data but is related only to apps that have installed their analytics code and can't provide a full and accurate picture about what's exactly installed in a given device.
What we have done exactly?
We used a sample of 1200 appsfire users (meaning that we know precisely what each user has) and took anonymously (we insist on the privacy respect), a picture of what's exactly installed on each single iPhone. We then aggregated this data to produce a unique overview about what can be found "on average" in an iPhone.
You can view below the actual data and the conclusions (or here if you read in RSS)
While many debate about the actual size of the Paid App Market (remember it is different than the size of the iPhone app market), our preliminary research drives us to think that Admob is not far from being in the right range. It will be interesting to see over time how this evolves as our methodology sharpens and app prices change.
Some important reserves should be taken into account
- We are assuming that the price of each app has not changed since we do not know when each app has been installed and purchased. Since the average price of each app is decreasing over time, this means that our data is understating the reality
- Like for all research, the data gain meaning over time when it can be compared and analyzed consistently. This is only our first release, we ll issue more data soon that will be compared with that one
- Sample representativeness : this is an area that will get better over time. We need first to make a split between iPHone and iPod Touch users. Next releases will include that. In addition, we need to make sure the windows/mac ratio is more clearly addressed. We released our Windows version after our Mac version. A biais could appear there. It will disappear over time
- Over time we ll provide a better picture of this data per region and country, since its local App Store has its own catalog.
update: the news has been covered by TechCrunch
update 2: we made the home page of Digg. you can give us a vote if you wish
