Here's how I'd do it... I'd massage this a bit with some closed alpha testing, etc... but this is the gist:
Sell ads and kickback a percentage to the siteowners while working in charity donations to remind folks that you're not evil.
Create a script to categorize your install base. Eg: Gadgets, Tech, Food, Personal, etc. In version 2, plan to categorize in real-time contextually. A good ad network partnership (see below) will help you with this technology.
Change your TOS. Send an e-mail to everyone asking them to verify that your script categorized them properly, to check out the new TOS and the new ad feature. Clearly and honestly tell them why you're making this change - the project will die if you don't.
Test some "run of network" ads. Input a few different variations, measure clicks (volume and geo are important).
After a couple of weeks, approach a few different ad networks at the same time. An Adtech conference is coming up in NY in November... use their exhibitor list as a hit list.
Negotiate with the ad networks. Depending on the demographics and CTR's, I bet with a 5m install base, you'll have some good offers. Reliability is key here - ad networks start and stop all of the time. In contrast, remember that big networks come with more constraints.
Here's the part that's going to keep this from being evil: Allow your siteowners to sign up and insert their own affiliate / user codes. Give them a percentage of the ads. If someone does not sign up, give that same portion to a charity such as Kiva.
As an alternative to selling ads - start looking to sell the whole thing. There are a few funded startups I can think of that would keep your beer fridge full for a couple of years in exchange for what you have built.
As a final (tacked on edit) idea: Survey your siteowners. With such a large install base, if you ask what advanced features people would like, you may be able to get some people to upgrade to a paid subscription model. You might be able to use this in conjunction with the adrev model above. In your survey, don't give the siteowners any ideas - you don't want to sway them.
Setting up an ad network that charges advertisers, targets ads and then compensates site owners is not a fun way to monetize a widget. It's recreating AdSense from scratch. Think of the tax issues alone!
If you could build that ad system, coming up with 5 million impressions would be the easy part.
You would not need to build an ad network. There are a few hungry ad networks out there already that offer text ads and have systems similar to this already in place.
The only challenge really is in categorizing the sites where the widget appears (to target the ads so that the CPM ratio doesn't make it worthless). Past that, it's really just seeing who has a good base of customers buying text ads and letting them handle all of the dirty work.
Sell ads and kickback a percentage to the siteowners while working in charity donations to remind folks that you're not evil.
Create a script to categorize your install base. Eg: Gadgets, Tech, Food, Personal, etc. In version 2, plan to categorize in real-time contextually. A good ad network partnership (see below) will help you with this technology.
Change your TOS. Send an e-mail to everyone asking them to verify that your script categorized them properly, to check out the new TOS and the new ad feature. Clearly and honestly tell them why you're making this change - the project will die if you don't.
Test some "run of network" ads. Input a few different variations, measure clicks (volume and geo are important).
After a couple of weeks, approach a few different ad networks at the same time. An Adtech conference is coming up in NY in November... use their exhibitor list as a hit list.
Negotiate with the ad networks. Depending on the demographics and CTR's, I bet with a 5m install base, you'll have some good offers. Reliability is key here - ad networks start and stop all of the time. In contrast, remember that big networks come with more constraints.
Here's the part that's going to keep this from being evil: Allow your siteowners to sign up and insert their own affiliate / user codes. Give them a percentage of the ads. If someone does not sign up, give that same portion to a charity such as Kiva.
As an alternative to selling ads - start looking to sell the whole thing. There are a few funded startups I can think of that would keep your beer fridge full for a couple of years in exchange for what you have built.
As a final (tacked on edit) idea: Survey your siteowners. With such a large install base, if you ask what advanced features people would like, you may be able to get some people to upgrade to a paid subscription model. You might be able to use this in conjunction with the adrev model above. In your survey, don't give the siteowners any ideas - you don't want to sway them.