I really do agree with your first idea: I would love to give my referals a login/pwd that would show the stats associated with their referal links: pages view/goals/$
There's many different taxes as you can imagine, but I'd say a little bit more than 50%.
With that you get plenty of benefits (school, health care, unemployment, you name it)... but we also have a huge government deficit.
At some point, we'll have to change stuff. And that will be painful ! (Meaning that do should really not plan your vacations in France at that time)
Going after the "typical large outsourcing firm" seems too much for us: we suck at sales.
Internet is more our cup of tea, and easier to market I think, and yes I have to admit the dream to set up a service, let it run, then sit back and collect the checks it very tempting :)
Any hint on the pricing model you would see for the online service ?
If I were you I'd avoid anything overly complicated.
A lot of potential customers are going to balk if they have to put any work into figuring out how much your service is going to cost (eg: if it was based on some formula involving the table count and the column count(s)).
Sometimes a clever and innovative pricing scheme works well, but a lot of times customers shy away from anything they're not already used to. Keep it simple to figure out what this is going to cost.
If I were you I'd split your offerings into three tiers, basic, medium, and enterprise, and then sell them on a subscription basis. When you subscribe to a tier, you get unlimited use of that tier's offerings (!).
Basic would be aimed at projects a small ISV would tackle; probably put in some limit on the model complexity, or omit some of the crown jewels.
Medium would be aimed at larger projects; relax the limits on model complexity and omit fewer of the crown jewels.
Enterprise would be aimed at "enterprise" stuff, would have essentially no limits and would include the crown jewels, if you have any.
Candidate prices: 99/yr, 999/yr, 9999+/yr, but you should do more research to home in on something.
I suggest yearly subscriptions b/c it's more money upfront, it puts less mental strain on a potential purchaser, and the larger #s make you look more serious.
About the "more serious": $99 is cheap, but a believable price for "serious software". The equivalent monthly rate is ~$8.50/month, which is hard to take seriously as being "serious software".
I want to reiterate that you'll probably have better luck monetizing this in some different way; no harm in trying the service route first, if that's what you want to do.
Even as a service, though, you guys need to pound on some doors and do some interviews to figure out which features are wanted by what potential customers, and why; this will help you refine your offerings, get a better handle on the right price(s), and figure out where best to draw the lines between tiers.
(!) If you do offer unlimited use within a tier make sure your terms of service prohibit resale of the service itself. You'll also need to plan on building a monitoring infrastructure to alert you to "suspicious use" (eg: high volume on an account, or a large # of distinct ips using a particular set of credentials).
