What happened to WAP?

NOTE: This is a restored version from this archive.

From 2000 to 2002, I was working at Ubicco, a Fi System spin-off dedicated to mobile phone services. We developed an application server and framework that was supposed to make your life easy when you wanted to build WAP / iMode / cHTML / SMS applications.

Back in those days, I was pretty active on the xml-dev mailing list, so when on 11/03/2002 someone asked “What happened to WAP ?”, I could not do otherwise than write a rather lengthy reply. I’m posting it here because all what I wrote then still stands time, even if things changed a lot in the industry (I would never have thought of the phone-with-camera thing back then). So, let me publicly pat myself on the back !

First of all, even if WAP is considered obsolete and said to have disappeared, the number of phone models supporting WAP is steadily growing in Europe (NB : I’m French and I work in a company who sells a multi-access (including WAP) application server, also a WAP Forum Associate Member). A majority of the phone you can buy at any retail store supports WAP 1.2.1. The handset renewal period is estimated to 18 months by analysts, so we should expect the proportion of deployed WAP phones in Europe to raise dramatically.

But the fact that WAP is not as successful as expected is blatant. Here are a few reasons for that :

1) Too high expectations.

WAP has been marketed a “Internet on your phone”. A simple demo of a WAP phone could easily prove you that it was quite limited.

People lured by the marketing fluff into looking at that WAP thing were easily thinking they were taken for fools when seeing the real thing demonstrated.

Yet even with very low bandwidth, graphical resolution and ergonomics, a device could be successful, the French Minitel is an example for that. Even a crappy device can be successful if there is a killer app available for it (the Minitel killer app was notoriously adult chat rooms).

2) No killer app.

The “Internet on your phone” meme was probably a killer meme, here. There was no real understanding of the fact that WAP users were using their phone while mobile, and that if someone would wait for 15 minutes until they reach a proper PC with high speed Internet Access (heck, even a modem is high speed compared to WAP’s 9,6 kbps), they would not go through the hassle of using a WAP site that was a bad translation of a badly designed Web site.

Developing WAP applications require a non-nonsense approach of usability, both for the service itself and its ergonomy.

I don’t need to buy books from a WAP phone, I can wait 15 minutes to get to my home/office PC, or if I’m in town I can directly go into a bookstore. But that didn’t prevent an online book store from having a project of WAP e-commerce site for which we have been consulted.

Moreover, your site usability had betted be finely tuned and tested, because if you force your users to struggle with your site to get what they want, given that they pay the communication fees, they will leave your site and you’ll never hear of them again. Anyway (sic - this sentence was never finished)

A lot of time has been lost developing the wrong services with the wrong apps. This turned away many potential users. The lack of users then discouraged further interesting services from appearing. Technology misuse has killed WAP by reinforcing the chicken-and-egg effect.

3) Costs and revenue model.

WAP is charged to the user on a per-time basis. Depending on your mobile services provider, it is charged in or out your time credit (when charged out of the time credit, the costs can be quite high in France). Per-time charging + low bandwidth = unsatisfied customer.

Plus, all the money goes to the mobile services provider. I don’t know any European operator that shares its benefits made on WAP airtime. This is bad since as the user is charged on a per-time basis and the bandwidth is low, the use of advertisement to generate revenue is not possible.

This is to be compared with i-Mode (NTT Docomo, deployed in Japan but coming in Netherlands and Germany) which at least proposes a shared revenue model. Content providers are therefore encouraged to develop new services, and as they compete for users, the quality and originality of services is increasing. Without advertising, (sic - this sentence was never finished)

4) Bad timing, first strike.

It is not easy to verify this, but WAP phones may have suffered from the fact that they came on a saturated (or nearly saturated) mobile phone market.

In Europe, the number of mobile phone owners was stabilising ; the interesting offers that provided a strong rebate for mobile phones if you subscribed to a mobile operator was maybe appealing, but it needed you to change your phone number.

So, to buy a WAP phone at an interesting price, either you had to be a first-time customer (which is less common on a saturated market), or to change your mobile operator (you had to churn). Changing your phone is a strong churning incentive, but the volume of churn generated by this incentive is small compared to the growth rate of a non-saturated market.

I think the problem here is that WAP allowed companies with an Internet experience to enter the mobile phone market, without preparing them to its specificities. On the Web, a browser is a relatively easy to replace piece of software. With WAP, your browser is embedded inside a 150-300 euro device that is competing with the one a big number of people already have in their pocket. This makes the market very different and very reluctant to technological evolution.

5) Bad timing, second strike.

The timing was bad with regards to the rise and fall of the Internet startup bubble. WAP arrived on the market when the so-called “new economy” was booming, and lots of company were created to ride the WAP wave. Then the bubble burst, and there was a strong disillusion feeling. I think the WAP baby has been thrown away with the new economy bath water.

6) GPRS and UMTS

A lot has been promised for WAP, and people were deceived. A lot has been promised for GPRS and UMTS, too, but the general availability of these technology went from “extremely soon” to “somewhere in 2006-2007 (for UMTS)”.

Combined with the need to buy a new device, this rapidly caused people to admit the general saying that “WAP [was] dead” and wait for the next generation. This seemed reasonable when GPRS was promised for 2002 and UMTS for 2003-2004, but this slowed the growth of the WAP phone base, and effectively contributed to making sure that WAP wouldn’t be successful. The irony is that GPRS and UMTS will suffer from the same problem as WAP did, namely the renewal of mobile phones.

7) Technical problems

I put it as a last item. People may scorn the WAP Forum in their vain attempt to redefine a whole independent set of network layers, but they were truly successful. WAP does work. WAP 2.0, especially over GPRS (if we finally get it) will work even better.

It’s true that the fact that the WAP presentation language WML was very different from HTML has slowed the development of WAP applications. But it would be ridiculous to expect that this would have been different if the presentation language was HTML or XHTML. The point is that the particular ergonomics of the phone (screen size, navigation capabilities) forces developers to design an application specifically for WAP browsing. Even if the presentation language was HTML, there would still be a need for specific development for phones.

WML is not a complicated presentation language, and its XML root make it easy to write well formed and valid WML content. Even if real WAP phone have different behaviours and bugs (by having written an adaptation layer that handles 60+ WAP phones, their specific screen sizes, behaviours and bugs, I can tell it), the situation is much better than if we had 60+ phone models based on a loose HTML subset.

WAP 2.0 supports XHTML Basic for basic pages, and the full WML 2.0 DTD is an XHTML DTD, with lots of specific tags and attributes, which is why I laugh when I hear “standard XHTML”. By including extension capabilities, XHTML standardizes the fact that nearly anything can be XHTML. Anyway, WAP 2.0 is fully XHTML compliant, but this won’t mean that specific adaptation won’t be needed, believe me.

That is the real technical problem. It is not due to WAP itself, but simply to the fact that developing an application for a WAP client is the perfect use case to check the flexibility of already existing presentation frameworks, MVC design patterns et al. More than often, our customers realised that taking their already existing web site and building a WAP application with the same content and logic was more or less hopeless.

One should therefore not point finger at the WAP Forum and tell them “WAP is dead because you wanted to build your technology alone without involving the W3C or others, so you’re paying the price of your arrogance”. Whatever the technology used, be it WAP, lightweight HTML, XHTML or cHTML (in i-Mode), development for phones require a specific application with dedicated ergonomics, special navigation mode and so on.

This, combined with the fact that lots of already existing web sites or web application are badly written, mixing presentation with business logic, made the cost of developing a WAP application too high. Add a bit of market specifics, especially the impossibility to have a decent revenue model due to the short sightedness of mobile operators which would not want to share their revenues, and the fact that the market has become saturated so the deployment of new phones has dramatically slowed down, and you obtain WAP as it is known now.

8) The success of SMS

While WAP stagnated, a low-tech player developed and boomed : SMS. This asynchronous, two-way, short message (160 characters) delivery system is generating high revenues for operators (in France, about 10% of the revenue of mobile operator come from SMS, and this share is growing).

Why does a very low-tech, poor media technology (SMS can be used to download custom ringtones and screen logos) succeeds where WAP has failed ?

Simply because all GSM handsets can receive SMS messages, that the charges are high but per-message (and not per-time), because it is lucrative (mainly for the mobile operators, but the SMS Premium charging system will allow content provider to receive a share of the revenue), and because there are killer apps that advertised the technology to the masses : chat, games, and custome ringtones & logo download for now (it looks stupid, but a lots of revenue is generated with these three apps). Revenue is a great incentive for content providers, so the content quality is slowly but steadily increasing, et voila, you’ve got a prosperous industry.

9) Conclusion

To conclude, we should not look at pure technical reasons to explain the demise of WAP. The problem is mainly rooted in the dynamics of the mobile phone market.

So will be the solution, because even if WAP looks dead, it is not. As I’ve said at the beginning of this mail (the previous one, in fact), there is a majority of WAP 1.2 or WAP 2.0 enabled phones coming on the market now. By creating the SMS Premium revenue sharing scheme, the mobile operators seem to have eventually understood that revenue sharing is the best way to create a market (or at least, they remembered the Minitel lesson).

Without revenue sharing, it was virtually impossible to make money on the WAP market. With higher bandwidth and volume charging rather than time charging, we can expect the advertisement market to appear. With advertisement and revenue sharing, we can expect the WAP market to become lucrative, and hence become a support for the technology instead of a dead-end.

What is the situation today ? Well, first, I’ve left the mobile phone application industry just a few weeks after writing this (I began at my new job in April 2002), and I’ve not really looked back since, so I’m far less informed than before. What’s sure is that WAP, WML and WBXML are now implemented in a vast majority of mobile phone. In France, there is at least one operator (Bouygues Telecom) who provides iMode services with a revenue sharing scheme. On nearly all phone, WAP or iMode services can be accessed with a single keystroke. A majority of phones now embed a Java Virtual Machine, mainly used for games that you can download for a fee. Plus, as everybody knows, SMS Plus services took the lion’s share in mobile phone revenue generation (to the point that SMS+ can be used to buy things or send money to the tsunami victims). So I guess I was not totally wrong in my analysis, which gives me that warm and fuzzy feeling.

What I did not see back then, however, is how adding a crappy camera inside new mobile phones would boost the phone renewal rate and prepare the field for the deployment of 3G. I’m quite amazed by the quantity of ads about 3G video mobile phones you could see in the subway last Christmas. Well, for now, rates are quite prohibitive, and network coverage is mainly limited to the main French towns, but I can really go and buy a mobile phone, make some pictures and movies with it and send them to another 3G enabled friend. I can even have a real-time video conversation with him ! Well, except I don’t have any friend with a 3G phone for now…

Anyway, remember that all this picture and movie making is not comparable to the usage of the potential high-speed Internet connection, which was (purposely ?) not advertised. For now, officially, a 3G phone is a way to communicate through voice (duh !), pictures and video. This strongly contrasts with the vision we had in 2000 when we expected the 3G phones of 2006 to be nothing else than a portable internet access point. Why is it so ? Well, I think that the point I made in my post back in 2002 still stand : the revenue models from the Web cannot be applied to the Mobile Web ; revenue sharing should be a rule, not an exception. It should be easy for anyone to enter the market and get a revenue provided its service is good. The technical issues come second, after this basic yet crucial economical issue.