TypicalSurfer - 12:52 pm on Jun 22, 2013 (gmt 0)
There are plenty of SE's out there, I built an index of 100M pages running on 5 midrange servers, crawling, serving, etc but the project fell apart due to lack of funding. Results served were indeed better than what google is providing.
To get a good prototype pre-production engine in place realistically is in the 200K range, figure 100k or so on machines, bandwidth etc., programming and labor another 100k.
How to monetize search is the challenge, today everyone is doing a "me too" bit with highest bid PPC, no incentive for CUSTOMERS, without ad buyers you have no model. I think a flat rate ad model would work as legacy PPC is just a margin eater, a tax of sorts that scales with your margin, it is not friendly to the bottom line and is chaotic in the sense that your ROI can change by the second.
If a marketer could buy 5 cent clicks at XYZ engine, that marketer would tell people to use XYZ engine and you really need that word of mouth.
Offering a better result set than google? That's easy peasey, they deliver junk and hide the goodies but you really need to offer something compelling to get the buzz moving. Compelling to me is a product that enhances my bottom line, restores sanity to my marketing budget.
BTW, you don't need a 100 billion page index out of the box, a quality 100 million page index will answer 95 percent of queries, you fill in the blanks from there, it's not that hard.