Larry Page argues that companies innovate in marketing, not technology. He shows how growth and communication costs kill innovation, and critiques the tech industry's obsession with standards like XML and .NET. A scanner example reveals how product development often misses what users actually need.
Chapters
So I think in general this doesn't actually happen. People are not innovating in business. They're innovating, they're generally innovating in business and not technology. This is because often doing the technological innovation requires broad understanding of both the business and the technology and the marketing and all the other things you need to do. There aren't very many good examples. So when we started Google, people said, oh, you build a better search engine. Why will that be good enough? And we said, well, we think people will care and they'll use our site and so on. And a lot of people laughed at us and a few people were nice enough to fund us. But there weren't that many examples of companies that are really founded on innovation and not just on pure business techniques or something like that. Um, also companies generally don't innovate that much. And I guess as Google's gotten bigger, we're almost 400 people now, it's, you sort of notice this as you get more and more people working on one thing, it's harder and harder for them to be innovative just because the communications cost and the inertia and all those kinds of things. And most companies end up being marketing or sales driven and not engineering driven. So let me just give another example quickly. There's a thing I call the technology trend bandwagon. And technology companies specifically get all really excited about various standards like XML or.NET or CORBA or any of the things I have listed here. And in fact, they're kind of like plumbing. If you had, for example, do you really care if you have copper plumbing in your house or plastic plumbing? Well, maybe one's a little bit quieter, a little bit cheaper, or so on. Maybe the water tastes better with copper. But there's pretty minor differences. And much of the things that the technology industry really gets excited about don't actually affect users very much. And in fact, you can ask, does this really affect programmers very much? Does anyone really care? People generally get excited about things that don't matter. I'll give you one other quick, quick example. If you go to Fry's and you buy a flatbed scanner, You know, every month you go to Fry's, it'll have like twice the resolution and they're up to about 2,400 dots per inch. And if you've never played with one of these, it can actually image every dot of ink on your paper, you know, and you can see that the dots of ink are round and things like that. And you can see every grain of, you know, grain and fiber in the paper. That's not tremendously useful. And it turns out as a result, the scanners are really, really slow. And in fact, if you— I challenge you to go to Fry's and find any scanner that's that says how long it takes to scan something. And the reason is if you— if it's set on the box, less people would buy it, right? Because it'll say like 3 minutes to scan a page. And in fact, what people probably care about is to scan, you know, with reasonable quality quickly. And actually, no one's built a product like that. There's probably, you know, a good market for a product like that marketed correctly. So I think a lot of things end up being, you know, you end up focused on one very specific thing and not looking at what users want.
