First, the article assumes each developer creates one app (160,000 * 8700 = $1.392B revenues, $1.392 * 30% = $418mil "profit"). This error leads to further errors down the line.
Second, Apple App Store sales is estimated to hit apx $4.9B 2012 [1]. This means apx $1,470mil in gross income (what the article calls "profit"). The article uses only $1.39B revenues and $418mil in gross income in its calculations as outlined in point one.
Third, this analysis assumes every developer would not purchase an iPhone if they were not developing software. Given that the vast majority of iOS developers do it part time, I think this assumption is highly unlikely.
Fourth, including profit from phones as a % of appstore profits is just wrong. Phone profits are not included as part of appstore profits. Thus it cannot be used as a % of it. It is additional income, on top of the appstore.
Finally, just to give you an idea of how little profits from developer matter... In the last 4 reported quarters, Apple earned a pre-tax net income of $49,000mil. Or $134mil per day. This versus the article's $64mil over the entire year from developers. Somehow I don't think they spend a lot of time or effort trying to get money out of developers.
[1] http://www.mediabistro.com/appnewser/apples-app-store-revenu...
The analysis in the (simple) article, thereby, is actually quite funny, and is part of an overall correct narrative; although, I agree that it was not described in quite the right context, and lends itself to your math complaints, which are technically quite correct.
From this standpoint, using the average developer income does not destroy the analysis (and is in fact required of it): if I were using a similar analysis in a talk, I'd want to speak to the people in the room, to give them a feel for what Apple will make from that sample.
Secondly, it is not problematic to ignore that the developer might have otherwise purchased an iPhone: the purpose of this kind of analysis isn't to demonstrate that Apple is "trying to get money out of developers", it is to demonstrate a comparison in the money made.
However, you are quite right that the math in the article is flawed: thankfully, some parts compensate by being conservative; once you take into account processing fees and other fixed costs (as various others have done), Apple's profit on the App Store is closer to 10%, not 30%.
We then get $4.9b*10%/160k, giving (conservatively, working back up towards 12%) ~$3500/developer. ($300+$99)/$3500 is >10%, which is still an interesting figure to be able to trot about at conferences (especially if you try to estimate the money spent on Macs).
(If I were to try to undermine the math in this argument, I'd probably do so from a different angle: the $99 is not pure profit, as Apple has employees scanning over the contracts and vetting corporate accounts; the fixed per-signup costs need to be removed.)
(Additionally, I question whether the average developer buys the most expensive device Apple offers, or instead purchases an iPod touch to use only for development; and then, keeps the iPod touch for longer than a year. A bi-yearly iPod touch would destroy this math.)
It's also plausible.
Early Autodesk's entire business model was built on top of fleecing their dealer (value-added reseller) channel.
Ditto franchises, pyramid marketing, ponzi schemes.
That said, if you're smart and very lucky, it's possible to make real money playing in someone else's sandbox.
Three years ago, a hobbyist could make an app in their free time, put it out there for cheap or free, and see what happened. If the app was any good, what would happen could easily be that they made a livable income off of their creation.
Today, that hobbyist has basically three options:
1. Pay Apple $99/year for the privilege of showing their creation to the world, and release it through the App Store. Feel their enthusiasm for the project drop as they go through that tooth-pulling process, but expect that it will all be worth it since the App Store will at least give them some good exposure. Watch as it doesn't.
2. Pay Apple $99/year for the privilege of showing their creation to the world, and release it on their own. This is a lot like the old days, except that now Apple is taxing you, and the old tools for spreading the word have largely starved.
3. Don't pay Apple a dime and release the app unsigned. Explain on the download page that the end user will have to edit their security preferences to allow unsigned apps. This looks sketchy to users, and now our hobbyist is suspected of being a criminal.
Consider all the other activities that people do as hobbies that have absolutely no possibility of even breaking even. Some of these require expensive equipment, fees, time. Getting set up to develop for iOS, even allowing for the cost of buying a Mac, compares well to what else you could be doing.
You can tell your customers to right-click on the unsigned executable and press "Open", which looks less sketchy.
The huge customer base with credit card details stored already, one-click away from a purchase and with such low piracy rates as those the App Store offers was UNHEARD of before it.
And that's for the Mac App Store: for iPhone/iPad apps the market didn't even exist _at all_.
There's a reason developers jumped all over the App Stores. "Tooth pulling process"? You should have tried making a living off of shareware in the old times.
Also (2), why pay $99/year? You don't have to sign to the developer program to release stuff on your own. XCode can be downloaded for free IIRC, and there are tons of other IDEs you can use.
And (3), security preferences? Just right click "Open" and allow the app to run the first time. I really doubt "This looks sketchy to users" of that the "hobbyist is suspected of being a criminal".
is that true? i would have thought it's a power law. in which case it would be more correct to say "Very few apps do make $8,700 (and of those that do, few make far more)" since with a power law distribution, no matter what level you pick, the extremes are still extremes (if you see what i mean - the 1% have their own 1% who are the crazy rich part of the crazy rich...)
so does anyone have any numbers?
[it strikes me that, depressingly, you could summarize this as: no matter how good you are, getting an order of magnitude better is still hard.]
"Global Apple App Store revenue is set to increase to $4.9 billion in 2012, up from $2.9 billion in 2011, according to an IHS Screen Digest Mobile Analyst Commentary from information and analytics provider IHS (NYSE: IHS). This means that nearly half of the revenue generated by the App Store in its five-year history will be earned this year alone."
Those are some pretty strong numbers. Given that the number is projected, even with a generous error rate of +-20%, any revenue from developer fee will be eclipsed by 4.9 billion.
While the author's estimate that apple generates 15% of its revenue from developers is inaccurate, it has to be said that the entry fee to get into iphone / ipad development is high. I would know since I'm currently developing a product that works across iphone / android / web. Here are our costs thus far (excluding phone-device prices ).
Apple:
Mac Air 1200$
Developer account 99$
DUNS number 200$
Server certificate for notifications 175$
Android: Developer account 25$
<For development - Use the mac or boot an old clunky win laptop with Win-XP on it>
Web: 0$ so far
Developing for apple devices is costly. For the same 1200$ that I could spend on mac air, I can get 2 powerful Win-7 laptops - ~ 500-700$ each [2]. Device fragmentation on android does require you to test on more devices, but there are services[3] that allow you to work around that problem. Besides, when netflix can release their app by testing it on a subset of 8 android devices [4], I don't see why that strategy is not good enough for others.
The author does have a point, but one that should not have been expressed as a percentage of apple revenues.
[1] - http://www.isuppli.com/Media-Research/News/pages/Maps-and-Pa...
[2] - http://www.dell.com/us/soho/p/vostro-laptop-deals
[3] - http://stackoverflow.com/questions/1852248/is-there-an-andro...
[4] - http://techblog.netflix.com/2012/03/testing-netflix-on-andro...
[EDIT] - Formatting
There is no requirement to spend 1200 on a mac to begin development.
There is no requirement to get a DUNS number to develop on the App Store.
As far as a server certificate goes, there is no $175 certificate for notifications. However, if you are running a web service it would be sensible for you to acquire a SSL certificate, regardless of what platform you are developing for.
Developing on "an old clunky win laptop" in particular is far from zero cost.
But as a developer, your own income is obviously not limited to what you can make on your own apps, and since diving into iOS six months ago my income has far exceeded any investment i've made in the platform. Mostly through making apps for other people, by my own app itself has very nearly recouped my total apple investment to date.
This is somewhat tangential to the author's main point, but my take is that there are many angles the author has failed to think of, and it's not a fair analysis of either apple or developers.
Where's the back-of-the-envelope calculations on spending $500-$1500 for a Mac and $99 for the Dev Account and turning that into a lifelong and highly valuable skill? It might be more money than $800 + $25 for the same on a Windows PC and Android, but it's not a huge difference when you compare the output. What other hobbies can be turned into a lucrative career with a soft cap of $2000 and 6 months to a year of practice?
(The talk was recorded, and I believe has been posted in various places; it also goes through why tiered pricing models for things like bandwidth work the way they do, as many people who are confused by the App Store also never worked that out for themselves either.)
"I've done the math before, and this post is wrong" isn't very compelling when the post in question shows the math behind its conclusion. What would be most compelling would be to point out the flaw in the post's math, and also to show us your own approach.
So you can take the post as right on the basis of the math in there, or even -more- right on the basis of saurik's calculations.
I think "The talk was recorded, and I believe has been posted in various places" was basically "you can find a citation if you look but I don't have one handy, sorry", which is fair enough for a brief HN comment; I, at least, still find the data point somewhat useful.
Compared to Amazon where the hardware exists to give users access to content.
Explains some of the price differences between the devices.
