An excellent matrix that is broadly applicable beyond SaaS but should come with various caveats and realistic expectation setting. I imagine that many SaaS providers would position themselves with just a fraction of the success factors you describe and have some pretty high hurdles to overcome to move forward - ie; legacy infrastructure, integration issues, budget constraints, cultural challenges, development expertise and change management to name a few. The ultimate success factor described here smacks of the 'one-to-one personalisation' utopian model expounded by the likes of Broadvision - http://en.wikipedia.org/wiki/Broadvision - in Web 1.0 which made a large contribution to inflating the dotcom bubble. True, time has moved on but there is still a big gap between the 'marketing ideal' and 'practical implementation' of these types of capabilities, particularly for a typical SaaS service provider. Would be really great to see some tangible examples in these pages to support matrixes and diagrams.

Personally I think it is a mistake to think that the Amazon personalisation model is broadly applicable to other businesses. Yes, it works for Amazon because of the nature of their business but, to be honest, the type of 'smoke and mirrors' approach you are proposing above is most likely to be counter-productive. Users can quickly identify if they are being steered towards purchases and will not take kindly to it if they feel it's not in their best interests or meets their needs. I can't see that the majority of SaaS service providers have sufficient 'value add' to make personalisation meaningful or beneficial and would be far better focused on getting the basics right first. Your whole model is around moving to web-centricity to attract web-centic buyers. I would say that 'web-centric' buyers are far more directed in their online decision making and understand that websites are a window on an organisation's operation and can easily see through sub standard and ill-conceived personalisation techniques.