I think number of pages is a good metric for pricing plans. Maybe the numbers could use some tweaking (e.g. something between 200 and 50k), but this seems like a good differentiator among people and how much money they have to spend.

It also seems like a good way to divide people by their scale. You want a metric that means something to a customer, and one that they will bump into eventually.

What you could do to add value to the upper plans, is also offer hourly scans/on-demand scans/scans invokable by an API for the big customers.

I almost did that. Ultimately I decided to take an "it just works" philosophy to crawl frequency (unless you need realtime).

Would you segment the number of pages differently? I think some sites with a lot of user generated content might actually have a ton of pages. But it is hard to get real numbers.

Maybe Patio11 will chime in with some ideas for your pricing plans? Everything I know I got from reading his comments/blogs.

(I think he searches for mentions of his username, so the bat signal is officially turned on :-)

They're undepriced, but I think the first problem is going to be going from 0 users to 10 users (bang down doors) rather than extracting the maximum proper value from those first ten users. The optimal pricing strategy will still lead to him having 0 customers until he figures out who needs this and how to reach them repeatedly.

You're right. Getting these first ten is SO hard.

After I have ten customers, I will have much more information to base my prices and differentiators on.

Hmm I still like the crawl frequency as a price differentiator personally.

Other ideas might be a service level guarantee for the high priced plans. And included email support for setup.