Downloaded the app and tried it out, having just picked up a used car last week.<p>As a 5/7 “car guy” I think of car buying in two phases: winnow the range of all possible cars into a few models (possibly one) and a few year ranges per model (also possibly one), then shop for cars in that set.<p>This app is exclusively useful in the second of those steps, though creating a decision matrix and scoring could be done similarly for the first step.<p>In terms of utility in the second step, I have weights in mind for different questions (“no significant bodywork” has vastly higher rating than “no stains on floor mats”), and many are subjective (no used car is completely free of wear marks, but different used cars present very differently and I probably want to record a 1-5 or similar rating for some categories).<p>Scraping AutoTempest (or Edmunds) and other sites to show me the function of asking prices vs miles and years would be worth something. I started creating that for my own shopping to try to augment my gut feel.<p>Certified pre-owned is a high-value checkmark for individual cars, typically being worth a few thousand dollars.<p>Where the car is is another large factor in terms of being able to consider it, from a time and money standpoint. I was shopping nationwide for two models, each with one series (a range of years where models were similarly designed/optioned), and keeping track of which cars I’d already ruled out was a drag. (For this, what I really want is a browser plug-in to change the CSS to grey out cars that I’ve already blocked [previous crash damage, missing option that I want, etc.].)<p>With multiple cars and the scraped pricing curves, you could do some visualizations (price on Y, miles on X as an example) of the cars vs the curves. (Or cars only, vs curves as premium feature.)<p>Absent data to show otherwise, I’d open the free version to at least 4 cars (probably 5) and provide a soft-delete (“I’m not ever buying <i>this</i> car.”) where those cars don’t count against the 5, in order to get more data on how people use the app, how they sort/filter, and to get more users. (If I’m only comparing two cars, I can keep that in my head.)<p>Overall, as-is I couldn’t/wouldn’t use it, mostly because of the fixed questions and lack of weights as above.<p>Other ideas: incorporate OBD2 scanning (perhaps with a rev share or affiliate marketing angle for the device). People will pay $25 for a device and $4.99 for your app long before they’ll pay $4.99 for your app alone.<p>Or to find hidden bodywork, affiliate links to paint thickness meter could bring in revenue.<p>Provide an “import details” button in pro version where you type in a VIN and scrape the web or find other sources for options, color, links, etc.<p>Provide a barcode scanner to scan the VIN sticker on the door. This could be an easy pro upgrade upsell.<p>I just went through this and spent 10 months of half-effort buying our car and there were definitely parts that could have been improved. Sites like AutoTempest (my primary) and cargurus* were helpful but still incomplete.<p>When I got to the point of traveling to inspect a car, I was much more in the mindset of “buy or pass” at that moment (the 37% [1/e] dating approach to finding the optimal mate via serial dating) than I was of score a checklist and compare overall. This app could pivot towards that by helping people declare how many cars they’re willing to go see, then frame each car in relation to this optimal stopping math problem.<p>Good luck!<p>* Small disclaimer: I know the cargurus founder and several people working in tech there. I think I’m unbiased in saying I used it consistently (and second to AutoTempest), but it’s impossible to tell with perfect certainty.