I am at RIM’s Blackberry DevCon in Amsterdam (where it is so cold that the canals have frozen). Attendees have been given a free Blackberry Playbook, the neat 7” tablet running an operating system based on QNX, acquired by RIM in 2010.

The Playbook was launched in spring 2011, and sales have disappointed. Exact numbers are hard to find; the Guardian estimated that RIM ordered 2.5m devices, while Crackberry.com says 5m. How many sold? In the three reported quarters, RIM said 500,000, 200,000 and 150,000 were shipped. Prices have been falling, naturally, but it seems that there are plenty left.

Nevertheless, this is an attractive device. The operating system is smooth and the size is convenient. Why has it failed?

One factor is that the device is designed as a companion to a Blackberry smartphone. Email does not work unless you have a Blackberry, or can get by with a web browser client. RIM thereby reduced the market to existing Blackberry owners, a mistake which should be rectified when version 2.0 of the operating system is released – expected later this month.

The second problem is the the extent to which Apple owns the tablet market. When you buy an iPad you know you are buying into a strong ecosystem and that every app vendor has to support it. That is not the case with the Playbook, making it a riskier choice. RIM’s fix is to introduce support for Android apps, though there are a few caveats here. Perhaps the biggest is this: if you want to run Android apps, why not just get an Android tablet and avoid any compromises?

The Playbook is a delightful device. The big question – for RIM and other new entrants into the tablet market – is what will make it sell, other than pricing it below cost?

Amazon found an answer for its Kindle Fire: low price, Kindle brand making it an e-book reader as well as a tablet, and a business model based on its retail business. Amazon can sell the device at a loss and still make a profit.

It is not yet clear to me what RIM’s answer can be. The most obvious one is to make it truly compelling for the large market of Blackberry smartphone users, but not if that means crippling it for everyone else as with the 1.0 release.

Another factor is that the device has to be nearly perfect. On the conference device, it took me 10 minutes to send a tweet. The reason was that the supplied twitter app is really a link to the twitter web site. That in itself is not so bad, but I found the soft keyboard unwilling to pop up reliably when twitter’s tweet authoring window was open. Making a correction was particularly frustrating. A small thing; but one or two frustrations like this are enough to make a good experience into a bad one.

Version 2.0 of the operating system does promise numerous improvements though, and watch this space for a detailed review as soon as I can get my hands on it.