[HALtheWise]

My favorite framework here is that a successful startup/business needs three things: a problem that people desire a solution for, a technology that solves that problem, and a business model that connects those two things into a profitable enterprise. Most (all?) successful companies start from a place where one of those three is not obvious and requires insight and faith from the founders, and the other two obviously fall into place as a result.

For example, Google's problem was non-controversial "Yahoo, but more accurate" and business model was "just sell ads", but the technology was ambitious and required insight. In contrast, Facebook had obvious tech "just make a website" that a team of dropouts were able to execute on, but the market desire for a new social network wasn't obvious. Uber is the clearest example of innovation in the business model space.

One implication of this is that the best way to pick a startup idea is to search for one of these components, then check that the other two are present. TFA argues for focusing on (searching for) the problem, but the other two approaches are valid as well, especially for highly technical founders or in rapidly-evolving technology domains.

[nvm0n2]

Google's problem was controversial actually! Back then the general consensus in the industry was that web search didn't matter, users didn't care much and the big thing was "portals" where one company provides all the information you want in a curated way (invariably that meant weather and news headlines). It was the classic problem big companies have with recognizing the long tail value of UGC: search was that little corner weirdo nerds went to when they were desperate to find some crappy GeoCities site about aliens. Normies just want to log on and consume curated MSNBC/online shopping.

Beyond lack of user interest, there was a consensus it was a fundamentally a bad business too. Advertising didn't monetize well back then. The ads came from generic DoubleClick banner ad inventory because ad networks and search weren't integrated, so users didn't click on them and rates were too low to justify investing. Especially because a search engine does a lot of expensive computation to show one page/ad impression but there's no way to earn enough from that one impression to pay for the servers (especially in the dotcom bubble when server class hardware was super expensive and scarce).

That's why when Google tried to sell itself to Yahoo for the princely sum of one million dollars Yahoo said no. It's also why DEC didn't invest in AltaVista's tech and why Google got a head start of six years before their IPO required financial disclosure and everyone suddenly realized all at once that wow, ok, search does matter actually. But by then it was too late to catch up.

From the perspective of entrepreneurs back then, Google was a terrible startup: a solution in search of a problem that lacked a workable business model. Fixing it required three insights:

1. Technical: PageRank served using cheap Linux PCs whilst designing for failure.

2. Problem: That users did actually care about result quality and it only seemed they didn't because everyone was at the same level of suck.

3. Business model: That you couldn't just build a search engine, you had to build an ad network too, and it had to sell text instead of images as otherwise you couldn't get relevance high enough to monetize.

So I think it doesn't fit well into the model espoused by this article. In particular there was no specific group of people declaring they had a problem. After all, there are plenty of search engines already, people aren't saying they want anything better because as far as they're concerned the market must have already converged on the best that can be done, and anyway the UGC long tail was small so why would you want to search it (no adsense=no targeting=poor monetization=no business model)

[chrishare]

I like that model. Seemingly, some companies like Palantir are special on more than one of your dimensions - but those are likely super rare in practice.