Incidentally, so that I don't earn a reputation as Resident Google Critic: I love Google, I think they've created more value for more people than any company in history, and but for Google I would still be a cog in the machine at a Japanese megacorp. I just don't think they are infallible, and in particular I have learned that their PR is as likely to be accurate as any other megacorp's is.
For what it's worth, there are several people in Google who will use your article to argue that we could improve our customer service/user support in different ways. I pinged one of them on Twitter and I'll pass this article around within Google.
Here's my personal opinion: I think the idea of paid inclusion had a deep influence on Google--we really didn't like the idea of paying to be in Google's index, because it meant that Google's incentives would be misaligned (not indexing pages would lead to revenue). I think that some people might have had the same worries about paid user/customer support (messing things up such that people had to call us would lead to revenue). That might sound crazy, but back in 2000 I saw quite a few companies that almost seemed to make deliberately bad software so that they could overcharge on support contracts. [Of course, there's the notion that Google tries to be scalable as well.] In my personal opinion, if anyone had the notion "we don't want to be paid for support, because that means our incentives are misaligned" in the early days, that notion is outdated and should be re-examined. Just my personal take.
I know the teams that support users and advertisers work really hard to scale and that it's a tricky problem.
> I know the teams that support users and advertisers work really hard to scale
If the examples given in the original article are true, esp. the fact that when one copies an ad to a different category, without changing anything, the copied ad has to be manually approved again, then it doesn't look like scaling is a concern.
But I think part of the problem is that ad-copy approval is a "ACB": ass-covering business. Kind of like the TSA.
If something gets through, you need to be able to show that you did anything you could to prevent it, and it's really not your fault.
ACBs don't care for efficiency: what they want is documentation.
It's also possible to just undercharge for support (and maybe even advertise that fact to limit expectations), but use the money as a filter. Ideally, you could even charge exactly cost, and then you wouldn't have misaligned incentives (and I would factor into the cost the information you'd gain from working with customers, so I'd say that to exactly align incentives, you'd want to charge somewhat less than the "support department" costs).
>I think that some people might have had the same worries about paid user/customer support (messing things up such that people had to call us would lead to revenue).
That's pretty easy to fix: provide an acceptable level of support without charging for it. Then it can't be an incentive. In fact, in that case it can only be a motivator to have good software because support would just be a loser for google.
