What the analysis is missing, in my opinion, is sufficient emphasis on the age of the iOS App Store. Not only has the $5 billion has been paid out over four years - meaning even a 1% app won't necessarily replace a decent job in a Western European country or the US - but also that much of that 1% has probably gone to apps that are no longer in existence due to changes to the approval process, the developer agreement, the API, or simply the maturing marketplace - i.e. 650k represents the current number of apps in the marketplace not the total number of apps which have offered over the past four years.
That's not to say the analysis isn't a useful back of envelope calculation. Because what it shows is that the economics of the app store probably favor developing many lower quality apps in places with low cost of living - places where a few hundred dollars is significant income.
I have one of those 1% apps (barely). What the survey didn't ask me was how much I spent on the thing. I spent over 50k on marketing and maybe 10k total on other stuff like graphics. It also took 3 yrs to pull in that money so way below typical wage in a western country.
Another point, which trails from yours, is the percentage of spending over time. I would estimate the the iPhone/iPad early adopters were more apt to pay for an app than your average user today.
The article is not clear on this. But the data from the image is unrelated to the stat of 5 billion. The data was compiled over 7 days from 252 respondent 1 year ago. From that data, the median life time revenue of an app is about $3,000 and the distribution on revenue is a long tailed one. The OP is also misleading in that you can't assume a uniform distribution in each bucket. Quite likely the same uneven long tailedness would appear at each level.
The article is pretty clear. The income distribution data in the image is extrapolated across the entire App Store.
"Before reading the rest of this post, bear in mind that these numbers are not necessarily accurate, and may not be representative of other categories in the App Store. However, if they were representative, the back of our napkin would give a very different split from the default assumed position:"
1% of apps (6,500) would be sharing 36% of revenue
($1.75bn), i.e. an average of $269,230 per app
19% of apps (123,500) would be sharing 61% of revenue
($3.05bn), i.e. an average of $24,696 per app
80% of apps (520,000) would be sharing 3% of revenue
($150m), i.e. an average of $288 per app
Sorry you're right, on that aspect I was mistaken. But the caveats listed in the thread and the weakness of the assumption of a uniform distribution in each bucket makes it difficult to draw any meaningful conclusions beyond what the original article had. That is, I don't think the mean is useful at any level.
the article's main message that being in 19% is an easy target is based on the mistake brudgers discovered; namely 25K in 4 years, or <10K in 1 year is barely a good target for any company or indie developer.
that's a presentation trick from Apple, mainly not telling how much developers get in current quarter, but rather bragging about a rolling number. the blog writer was a victim of the apple trick of rolling the revenue given to the app devs back to the beginning of the known history.
That's not to say the analysis isn't a useful back of envelope calculation. Because what it shows is that the economics of the app store probably favor developing many lower quality apps in places with low cost of living - places where a few hundred dollars is significant income.