Why don't you launch with no ads, and see if you actually get the traffic you are predicting? Worry about the ads later.

It takes two months(!!!) to get money from Adsense, for example, anyway.

OK, I cobbled it together this afternoon so HN gets an early (and ugly, designless) preview.

http://www.mpghead.com/

Basically allows you to determine how much you'll spend different from one vehicle to another just for gas.

If you get 14mpg and you pay $3.50 a gallon, you're spending $0.25 a mile! It's almost cheaper to take a taxi!

Cool idea.

From your initial description I thought maybe you were comparing specific makes and models. This would be more difficult, because you'd have to obtain (and maintain) this data from somewhere, but it's probably doable.

In that case, you could help potential buyers decide between one vehicle purchase and another based on fuel costs (not that your app can't do that now--the users just have to get their data someplace else). Nice of me to expand your project's scope, eh? ;)

I have to be somewhat careful playing in the automotive space since I used to work for a company that was in that space.

Luckily, getting automotive data is fairly simple as there are a few companies that provide it and images (and dealer inventory too) for a fee (of course).

I've thought about having a list of "popular vehicles" like Prius, TDI Jetta/Beetle, Mini Cooper, Honda Civic, Yaris, Smart Car, etc in a pull down that will prepopulate the vehicle data.

Also, I've been ruminating about a better ridesharing/carpooling site/app and this would be a good driver to it.

Timely ideas.

I was wondering if you could license it to others (Edmunds, etc.) but I guess it depends on them. Maybe with your industry experience you know something about that.

Anyway, good luck!

i'd consider trying to solicit sponsorships for lengths of time. i.e. a banner that maintans a persistent banner ad for a month for the monthly sponsor or something similar.

this would make you independent of pageviews and clickthroughs. you might not make as much $, but you'll not be as open to traffic changes and higher tech folks blocking ads.

I think that this is a good route.

I'm also looking at taking the tech "pluggable" and be able to resell it as a component for some larger sites.

Pandora is a fantastic example of how this works.

You'd get better and more straight forward answers if you gave a little info about your site. I think it'd vary on the type of page you were running.

Fairly simple/straightforward calculator/trend viewer.

Really one page unless I forced the user to jump through post back hoops (blah).

You can have users refresh the page when they want to see the results of a calculation. I'm not sure how often that's done, though.

AJAX changes the CPM (cost-per-1000-impression) advertising model, so you'll need a different type of ad network. The more mainstream ad networks haven't really addressed the issue for smaller sites, despite the fact that AJAX's popularity is peaking.

There's something called CPI (cost-per-influence) advertising, which isn't based on pageviews. Only one network comes to mind (Coudal's "The Deck": http://coudal.com/deck/) but maybe some Googling may lead you to some others. The Deck will pay you with a predetermined sum that is not pageview based.

It seems like it might be an opportunity for one of us here. Definitely and interesting project idea.

Cool site, but not quite in the space I'm going to be in.

My wife used to be an ad manager for a huge site (ya'll would all recognize it) but it's been a few years and she's amazed by how everything has changed.

It really depends on what you want to measure on an Ajaxy webpage. If you can discern metrics you can sell some form of ads based on that. Also, consider making the ads iframes/includes and refresh that at a given time interval.

Start without ads, gain your traffic, ebay your adspace and then advertise these listings in the adspace your selling. Depending on your volume, niche and longevity it might work.

Just an idea.

EDIT: Sorry, just saw noodles point.

I'm confused. Are you trying to be funny?

Maybe you should head over to chompchompdead.com...

Adbrite.com would allow you to sell ads for a length of time (day, week, month) instead of being based upon impressions or clicks.

Just have a timer that changes the ad every 30 seconds (or whatever time frame you want). Coding wise, should be pretty trivial.

google explicitly prohibits doing that with adsense; other ad networks may be similar.

Use a CPM company instead like ValueClick where you get paid per impression, regardless of click.

This doesn't solve his problem. His problem is that he only has one page--CPM means less return in his case, because if someone stares at the same ad long enough, they're probably more likely to click it than if they only see if for a few seconds. CPM would be good if he had thousands of pages and thousands of page views. And, I'm pretty much certain that ValueClick has terms of service that prevent him "solving" the problem by reloading the ads on the page periodically (since that's a tactic that would be very easy to abuse, and I certainly wouldn't buy ads from a company that charged me for multiple views on a possibly unseen page--what's to say the guy using the app didn't go to lunch or go home for the night and leave the app loading new ads every minute...that's an awful lot of wasted impressions).