The payment argument is nonsense

In response to my HTML5 apps argument a few people came back to how the payment thingy is missing from my idea, and how it will (apparently) be worthless because of that. I’ve been thinking about that a lot in the past few days, and I’m increasingly of the opinion that the payment argument is nonsense.

Sure, everybody who does iPhone apps, or who’s glancing cursorily at the mobile market without trying to gain in-depth knowledge, currently believes that the App Store concept is going to be a huge success because of the opportunity for developers to earn some money. But they’re just wrong.

I did some back-of-napkin calculations and found that, macro-economically speaking, iPhone app development costs money right now. And yes, an individual developer can strike it rich, but that’s getting rarer and rarer. I do not want to build a new app ecosystem based on arguments from developers who just want to take a gamble in the App Store roulette. Gamblers’ arguments are not real arguments.

Deconstructing payments

How do you, as a developer, get paid for your work on an iPhone app? (Or an Android, Symbian, Windows Mobile, BlackBerry or whatever app; there’s no fundamental difference between the various platforms.)

The first and most obvious way is just getting paid by a client for creating the app. Now that’s a payment model I can believe in, because creating apps for others is what developers do for a living.

Secondly, developers can create an app just for the heck of it; because they have a good idea they’re itching to try out, or just to gain some experience in building apps. Although a developer may put a € 0.79 price tag on the app, the point here is not making money, but developing for the sake of developing. No payment model here, but this, too, is something that’s not going to go away any time soon. Developers want to develop, after all.

Calculation

It’s the third model, developers creating an app and expecting to get rich from the proceeds, that garners the most interest right now. And it’s here that the fallacy of the argument lies.

Recently Tomi Ahonen published an interesting calculation. He estimates the development costs of an app at $30,000 (6 weeks of work at $125 per hour) and multiplies that by the number of apps in the app store: 150,000. That gives a total development price of $4.5 billon for all the iPhone apps that have currently been written.

Now one could object to those numbers. Should we say that a developer who develops an app hoping to get rich “pays” $30,000 for his app? Let’s say he doesn’t, and let’s furthermore estimate that 50% of the apps fall in this category. (I’m making up these figures, but I want to significantly downsize Tomi’s numbers before proceeding with the argument.)

Besides, 6 weeks of work might be a bit much for an average app. For your first app, sure. For a high-quality graphical app, sure. But not for the average app. So let’s say that developing an average app costs not $30,000 but $20,000.

So after significant downsizing we arrive at a total iPhone app store value of 20,000 x 75,000 = $1,5 billion.

Now the total revenue of the iPhone app store in 2009 is $780 million. 30% of this goes to Apple, so the worldwide app owners have earned $550 million, or slightly above one third of what they’ve invested.

Sure, the $780 million is per year, and might conceivably grow, while the apps that have been written for $1.5 billion will remain available year after year.

So let’s say the app store as a whole breaks even after three years. Breaks even! No profits yet. And the calculation already includes all apps that have been written basically for free in the hope of future wealth.

Then there’s the problem of updates. Updates also cost money, so there will be maintenance payments if you want to keep your app in the store year after year. How much? Shall we say 20% per year? (I expect it to be higher, given OS updates and such that have to be accomodated, but let’s again give app developers the benefit of the doubt.)

So in order to keep the 75,000 apps that people pay for in the store for three years, the cost per app is not $20,000 but $28,000. 28,000 x 75,000 = $2.1 billion, against an income of $1.5 billion. In three years, your average app will win back about 70% of its development cost. Not exactly a get-rich-quick scheme.

Striking it rich

Still, until now we’ve taken the macro view. The point is that individual developers are not necessarily bound by macro-economic rules. If a developer spends two months of his time on an app that ends up in the top 25 and sells millions of times, yes, then he has definitely made a healthy profit.

But how often does this happen? In the early days of the app store it did, but what are the chances right now?

All the stories I hear are about how hard it is to attract attention to your app without being a well-known brand, hooking up with a well-known brand (which graciously allows you to work for free in exchange for you hooking up to its brand value), a huge marketing campaign (which presupposes a large company with a well-known brand), or an outrageous amount of luck.

Yes, you can get lucky. But right now the reasoning of developers who want to get rich by developing apps mostly resembles the reasoning of the inveterate gambler who goes to Vegas one more time because he can feel that this time it’s really really going to work.

HTML5 apps vs. app store apps

I said that HTML5 apps, which are web-based and far more interoperable already than any native apps, will oust native apps in the long run.

The counter-argument is that the HTML5 app route doesn’t allow developers to get paid. That’s true — for now. (I expect the mobile operators to start offering a payment system in the next two to three years.)

Because developers aren’t getting paid, the argument continues, they won’t want to develop HTML5 apps.

Not the run-of-the-mill developers who just develop what they’re paid to develop. If a client wants an HTML5 app they’ll build an HTML5 app.

Not the developers who develop just for the sheer joy of developing. They’ll take a decision based on technical arguments; mainly superior UX vs. superior interoperability.

The argument applies only to the gold-seekers that develop apps in the hope of striking it rich. Yes, the HTML5 app ecosystem doesn’t offer something they consider vital and that means they’re less likely to switch to it.

But what percentage of developers are we talking about here? 30%? 20%? 10%? Frankly I feel that even 10% is a rather high estimate.

Besides, the number of strike-it-rich developers will only decrease with time because they’ll find out that striking it rich is more tricky than they think, and they’ll run out of money and become a for-hire developer who now and then develop something for the heck of developing.

The payment argument applies only to a small and shriking minority of mobile app developers, and it’s exactly that minority that has a gambling attitude. And I’m not really interested in arguments coming from gamblers, because they believe whatever they need to believe in order to satisfy their craving.

The payment argument does not say anything about the big and not-so-big brands who want a mobile presence (regardless of profit) and are willing to pay developers for building an app.

The payment argument, in other words, is nonsense that only appeals to gamblers. I’m going to ignore it.

Comments

1 Posted by Andrew Hedges on 11 March 2010 | Permalink

If your guesstimates are close to correct, this is the most convincing argument I've seen yet about why Apple's App Store (and, I suppose even more so for other app stores) is not viable in the long term.

Big brands will still find value in being linked to Apple and taking advantage of their distribution eco-system, but for indie developers, the choice is clear.

Viva HTML5 apps!

2 Posted by ppk on 11 March 2010 | Permalink

That's what I think will happen. Native apps are only interesting for companies who want to invest heavily in having a mobile presence - and even for them an HTML5 app may be more interesting.

(Of course devs who just want to develop stuff may build anything, and that's perfectly fine.)

3 Posted by Matthew on 11 March 2010 | Permalink

Have to say I agree with you about app store monetisation being a lottery nowadays. But I think 'getting rich quick' is not the only part of the attraction.

The idea of a centralised payment system is also, I believe, important. First, there's a perception (probably correct to a degree) that where you have a single, centralised entity handling payments, people are more likely to trust it with their money than a horde of minor app-makers: and so are more willing to spend money on apps!

Second, it saves developers from having to worry about handling payment methods & schemes themselves: the app store does all that for them.

However, neither of these is a fatal problem for HTML5 apps.

We just need an 'HTML5 app store'.

Would it be very technically difficult to do? It would need a solid payments backend. It could use OpenID for tracking users' identity across multiple apps and OAuth for authorising apps to charge users' app store accounts. A way to send users to a version on their platform's native app store might be good too.

I suspect the greatest difficulty would be getting someone on-board who's big enough to sell it to the mass market. But the 'HTML5 app' idea has so much potential that really shouldn't be impossible...

4 Posted by ppk on 11 March 2010 | Permalink

1) Payments will be handled by the operators, which will be significantly more user-friendly than even the app store. Just one "Yes, I want to buy" click, and everything's arranged for you. No logins or credit cards needed.

2) What people will pay for will not be the apps themselves (who ought to be transferrable via Bluetooth anyway), but the content. Get app for free, pay for the content it gives access to.

5 Posted by John Boxall on 11 March 2010 | Permalink

"I expect the mobile operators to start offering a payment system in the next two to three years."

While this is possible (most operators have been dabbling in this sort of thing since premium SMS) - do you believe it will have _any_ impact on the web development community?

Personally I don't have much faith in carriers' given their track record of getting developers onboard with their programs.

Is there any reason that would change in the future?

Any payment solution is going to come from the web!

6 Posted by ppk on 11 March 2010 | Permalink

It won't have much of an impact on web devs who don't work for a start-up, but it *will* have an impact on their clients.

Operators are failing spectacularly in getting web devs on board right now, but I feel that if I keep on talking to people at Vodafone, in a few years they might actually get it.

Operator-based solution is fundamentally more user-friendly than any web-based solution because it *doesn't require a login*! Or a credit card.

7 Posted by JP on 11 March 2010 | Permalink

Interesting that you wrote this today. I just wrote an article about this earlier today. You are so right! It's a losing proposition. All of the app stores should just be used as channels, and yet that is only temporary until HTML5 really picks up steam.

http://bit.ly/iphone-loser

8 Posted by Thijs on 11 March 2010 | Permalink

I'm waiting for a apple app store like website/application that is platform/device/operator agnostic but bundles the selling & advertising of HTML5 apps in one convenient place. One place to buy & (re-)download your HTML5 apps for all your devices. Mobile & otherwise.

If you let the operators do the sale of HTML5 apps what will happen to your app when you switch operators? Where do you re-download your apps?

As long as there is not a single big place independent from the mobile operators where you can buy and (re-)download your HTML5 apps I don't think this will go anywhere fast.

9 Posted by Thomas Aylott on 11 March 2010 | Permalink

I think you're mostly right in general.

But, I believe that the gold rush for iPhone app development will eventually die out. Which will eventually make it economically viable for long-term quality apps to break through since the get-rich-quick low quality apps will die out.

However, I do believe that the technology behind "HTML5 apps" will only become more important for both HTML5 apps themselves and iPhone App Store apps. Appcelerator Titanium, and things like that, make HTML5 app skills more portable.

There certainly are still usecases for developing Obj-c iPhone App Store apps (especially for the iPad). But the majority of apps can and should use a lot of "HTML5 app" technologies in one way or another.

10 Posted by girlie_mac on 11 March 2010 | Permalink

Recently I was talking with some iPhone users (non-developers) who do not really use web apps.

They said apps are avail at App Store and easily found what they want, or just randomly download something interesting.
Also they don't mind paying for apps because they are cheap and it's just like buying some candy bars - $2 a pop or so.

I personally have developed web apps and a "native" app using Titanium. I was amazed by the number of download the app has got.

So I really feel that regular users probably don't care how the apps are coded, but how easily obtainable matters.
As other guys already have commented, we need some sophisticated HTML5 app store/directory.

11 Posted by Armando Sosa on 11 March 2010 | Permalink

"He estimates the development costs of an app at $30,000 (6 weeks of work at $125 per hour)"

Unless, of course, that you're stuck in the third world, where very competent developers are getting paid $125 per week! In that economy, little $1,000 in profit is a small fortune.

I don't know how that is relevant to the discussion though, just wanted to point it out.

12 Posted by Nick Kleinschmidt on 11 March 2010 | Permalink

The reason there's so much contract work for iPhone apps right now is that other people are convinced they're going to get rich selling the apps later. An area without decent payment services is going to have scarce contract work. The first class of developers may not care about payment services, but they're going to go where the money is.

13 Posted by Enrique Amodeo on 11 March 2010 | Permalink

Good post, but I think girlie_mac is right. The more easy to find and install an application is, the more users will install it.

The install procedure is as important as the find and buy point. Talking with non tech people I realize they conceive an app as an icon in the desktop that they can launch with a single tap. They don't think HTML web apps are apps because they need to open a browser to reach them. So the HTML5 app must be installed locally to be on equal foot with native ones.

If in the future, to find, buy and install locally in the desktop an HTML5 app is as easy as a native app, then HTML5 apps for the mobile will success.

On the other side most people I know (techs and non-techs) don't ever pay for a mobile app. They always search for the free ones. But this is perhaps a cultural issue specific from my country.

About the costs of developing and application, perhaps in other european countries 125$ and hour is common, but here the standard fee is 60$

Go HTML5!

14 Posted by Jimmy Whales on 11 March 2010 | Permalink

I see the whole mobile billing / mobile payment issue resolving itself over the next year. There is going to be too much demand for developers not to correct this and open up new ways of freeing payments of an Apple monopoly. Regardless, marketing/branding/reviews are the only methods that will get apps noticed, otherwise did they ever really exist?

15 Posted by Peter on 11 March 2010 | Permalink

It will be interesting to see how the Google Apps Marketplace will factor into all of this. If that proves to be an effective model, they may expand it to include other types of web apps.

http://www.google.com/enterprise/marketplace/home

16 Posted by Marcelo Emmerich on 12 March 2010 | Permalink

Very interesting read this is. I am a software development consultant doing contract work for a carrier, where this is a very hot topic. I see a bright future for HTML5 Apps, however the following things have to be sorted out:

Copy protection. Are HTML5 Apps copy protectable at all? The app is the source code. If it is not copy-protectable you can't charge for it. Making HTML5 Apps copy protectable would imply a significant effort for carriers, OEMs and software vendors.

Network APIs. How will carriers avoid being degraded into bit-pipes in the future? Maybe network APIs (i.e. billing) is the answer.

Payment model. Shall HTML5 Apps charge by download, or by usage? Maybe using a subscription model? Or should HTML5 Apps not charge the end-user at all? At the moment I would tend to believe the latter option is the only viable one for several reasons:
- people are not willing to pay for using "websites"
- payment method fragmentation. How will you charge users? Paypal? Google checkout? Network billing API? Roll your own?
- UX break. People generally dislike entering payment information on mobile devices.
etc.

Ok, I ran out of space here :)

Marcelo

17 Posted by Maciej Aebkowski on 12 March 2010 | Permalink

First of all -- it’s not the developers who want to get rich. It’s the companies that hire them. So I think you’re wrong saying that not a lot of devs are coding for profit.

Also -- not all websites profit. Does that mean that only gamblers make websites for profit? Not sure about this.

Ok, but my main point is -- why fight with the AppStore? Can’t we just develop "HTML5 Apps" and put them in the AppStore? Sure we can. And we also can publish them for free (or donationware, or paid accounts) on the web.

18 Posted by Julian on 12 March 2010 | Permalink

I can see the point about making a living. But what about pocket money? The person who makes an app (HTML5 or otherwise) for the love of it, in their spare time, but wouldn't mind a little bonus, will not be valuing the hours as dollars, but as joy and entertainment.

19 Posted by E.Casais on 12 March 2010 | Permalink

For an experience report about developing an application for the iPhone, including figures (costs, sales, downloads) see

http://www.streamingcolour.com/blog/2009/04/27/the-numbers-post-part-2

At $32000, the total cost is very close to what T.Ahonen was suggesting, and hence the conclusions about the profitability of application stores for individual developers appear to be realistic.

20 Posted by two-bit-fool on 13 March 2010 | Permalink

I totally disagree.

Arguing that there isn't a lot of money in native iPhone apps, doesn't make HTML5 apps a better option...what are the economics of HTML5 apps?

Anyway, I tried to stuff my rebuttal into a comment but quickly ran over the 1,250 character limit. My full response can be found at...

http://tinyurl.com/ycxsnb8

21 Posted by Henri Sivonen on 14 March 2010 | Permalink

Billing via mobile operator might work for apps that are only relevant to a country or a city. And things with that kind of local-only relevance tend to be site-like and not app type.

For apps with cross-border appeal, it seems that Paypal, Google or the like are more probable monetization gatekeepers that mobile operators. (I doubt telcos can agree on how to federate their billing on the Web as opposed to calling a phone number in a couple of years.)

22 Posted by Constantine Vesna on 14 March 2010 | Permalink

How many iPhone apps got even at all? With 5 apps per user per month, 100.000 apps in store and power law, we can safely assume that over 80% of those apps are failure.

@two-bit-fool
I think PPK's point is that HTML5 apps will give you the same money as iPhone app ;) Which is around zero u.s. dollars.

@Marcelo Emmerich
Agree, there is no copy protection for HTML apps, so - only client-server solutions can have a business perspective, either it will be subscription or one-time payments.

23 Posted by Michael Kozakewich on 15 March 2010 | Permalink

$125/hour? How many developers expect to make $260,000/year?

There's a curve of experience-vs-rate, and I'd expect a large portion of the app developers to make far less than $125/hour.

Also, you assume offhand that HTML5 apps don't yield any profit. There are a great number of HTML apps that use a variety of methods to find some revenue.

In other words, things are a bit more complicated, and while App Store apps sound like they're just about breaking even (on average), HTML5 apps sound like they could cover their costs, too.

Perhaps the larger issue is to convince the public to pay a proper price for the work that goes into these apps.
And to not spend $30,000 on the development of the less-beautiful and less-useful apps out there.

24 Posted by Robert Schultz on 22 March 2010 | Permalink

The iPhone app store is like the American Dream.

Not everyone makes it but everyone THINKS that they can make it.

It's this idealistic view (which you label as gambling) that drives so many people to the platform.

The very act of so many people trying to obtain success cause a large number of apps to be created which creates a very healthy and robust environment which attracts even more people.

So I have to disagree with your assertion that the "payment arguments are nonsense".

25 Posted by David on 10 April 2010 | Permalink

Good post ppk.

I've been working for years on a way of monetizing browser apps. It's entirely feasible.

My solution involves around combining the development model from open source and the revenue model from proprietary (best of both worlds).

The system can also monetize content and back-end services. It seems impossible, but trust me it works...

More info at http://lmframework.com