If I understand you correctly, the service would work as follows: You enter a URL you want to monitor. My service would show you trends of similar websites to watch out for. It would base it on some sort of secret sauce that measures similarity and popularity of competitors. Correct?
That's an interesting idea. What types of current websites/businesses do you think could benefit from this right now? Could you give me examples? This could help me see if it's possible to monitor similar websites and develop trends.
Measuring a website's popularity is pretty straightforward -- you make some sort of formula based on how often that URL is discussed (from backtype), linked to (various sources), bookmarked, and ranked (alexa, quantcast, et al). It's a matter of mashing up that data with the similarity data (over time) to determine the trends of similar sites.
I didn't start the thread, but here's a tweak of this idea:
"You enter a URL you want to monitor."
OR - I copy/paste a line of javascript of your service on my site, like google analytics... your service then figures out which are the competitors of my site and how popular they are with my visitors (from their browsing history). The list of competitors should be updated with time. I think this could be quite useful if packaged well. Let me know if you'd like to collaborate on something like this or just further iterate the idea. My contact is in the profile.
That's what I was thinking of doing, and I'd love to collaborate. But, the URL sniffing hack is being batched up in the next generation of browsers, though. When it gets patched, that would essentially make the service useless... unless you can think of some other value offering. It's a pity, because I came up with a beautiful way of checking thousands per second with no locking up or freezing or any sort of negative feedback on the UI.
That's right - though you enter your own URL. Extra idea: it might also show you complements (products/services that fit in nicely with yours; these are usually pretty obvious in established industries - razors+shaving cream; milk+eggs; tyres +brakepads - but not so obvious in emerging industries. You might not have even thought of the category of a certain complement, let alone the name of it, nor specific brands that do it. It's valuable, because you might consider targeting people who buy the complements. It's a kind of market-research.
(Unhelpfully) I think every website could benefit from it. Probably emergent and fast-moving markets would benefit most (because they don't yet know their competitors/complements + it's all changing very fast anyway). A very specific start would be... y-combinator offerings (that have just launched - not the established onces); for fun, you could also do it for emerging languages (like Mirah). These aren't ideal biz markets, but they would gain attention here on HN, and might validate the basic idea (to see if it's possible.) For example, if you were able to find a competitor to a YC company that they didn't know about, it would be pretty compelling evidence (and will surely happen, since startups rarely look very closely for competitors; they are too busy doing. Which is a good thing. Knowing about competitors too early can be artificially discouraging. So this is probably most helpful for companies that already have enough traction that they wouldn't be discouraged. Competitors are good for giving you ideas; for demonstrating market existence (good for raising funds...); for suggesting markets you hadn't thought of. Complements are always good.
Perhaps also, companies that are always launching new products might appreciate it: eg Proctor and Gamble (though in their market, there are only a few competitors and they are well-known...). It might help in industries that are often disrupted - ie high-tech industries. Examples might be the established competitors of YC companies. If you could demonstrate legitimacy, big enterprise might pay a lot (perhaps equivalent to how much it would cost to hire someone to do the same work manually). You can be an arms dealer, and sell to both sides. :-) That's the nature of competition.
EDIThttp://www.mirah.org hasn't got results yet. JRuby's competitors are:
Groovy and http://rjb.rubyforge.org... There's a japanese (I think) news site. That might be helpful, if you can detect similarities across languages...
That's an interesting idea. What types of current websites/businesses do you think could benefit from this right now? Could you give me examples? This could help me see if it's possible to monitor similar websites and develop trends.
Measuring a website's popularity is pretty straightforward -- you make some sort of formula based on how often that URL is discussed (from backtype), linked to (various sources), bookmarked, and ranked (alexa, quantcast, et al). It's a matter of mashing up that data with the similarity data (over time) to determine the trends of similar sites.