Our development shop works with a number of startups. To keep billing simple (and in such a market, to reduce our potential liability) we make each startup sign up direct with the provider for hosting, monitoring etc.

That worked with the old pricing model, but the new one assumes either a) each company is large, or b) we’re going to resell.

For my own startup, we are still looking on how we can optimize our pricing. At this moment, we know that we are kinda cheap, but that has been a decision we made initially for increasing our reach and awareness.

We are currently in the process of properly segmenting our current user base and create different kinds of packages with specific features added to cover their needs.

What is still a difficult one of me is how you can properly AB test different pricing schemes without users being aware of this. Apparently for ServerDensity this hasn’t been a big issue, so I think we’ll do this anyhow ;-)

Customer loyalty and retention could be very important factors indeed.

Unfortunately in A/B tests Patrick describes he does not even mention retention, but that likely plays very important role in overall revenue.

It’s hard to test for retention, in particular because such tests take very long time (at least several months).

Are there any studies on how many people would self service a $299 plan though? Seems like quite a big number for a web signup type service.

I don’t think Amazon will go away from the linear ‘pay only for what you use’ pricing. I am no longer there so I have no inside knowledge but it is core to their model. If other companies compete directly with AWS, it may be hard to compete on pricing models that are significantly more expensive so you will have to differentiate in a way that adds customer value over what AWS offers.

But I definitely agree with the goal of simplified pricing and value-based pricing when possible.