Result is – you dot get the target type of advertisers which fits into your website. For example, I have not found any advertiser in the category of "Employment". But my blog is all about job & employment!

The Google affiliate network support email address Evan promised (also in this thread) turned out to be the same one I found via Google search so I don't know whether the reply was to the first email I sent or the two subsequent emails or a response to all three (squeaky wheel gets the grease).

On your point #2: It's not hard to find Google distribution partners still up to no good. In surprisingly short order, I have identified many tens of thousands of typosquatting domains — each identical to or confusingly similar to a trademark or famous name, and each showing syndicated Google ads. I've also found plenty of examples of spyware/adware-loaded forced visit pages that show Google ads, and I find more examples each week. I agree that Google has taken action against some dubious practices, e.g. some kinds of PPC arbitrage. But there's plenty more to be done. In fact, I think you agree — see your point #5, where you also focus on improper distribution.

On your point #4: I agree that it's hard to find an example of Google enforcing the no-copying API policy. But that doesn't necessarily mean the policy is having no effect. In PPC Platform Competition and Google's "May Not Copy" Restriction, I cite Microsoft's 17-step procedure for copying Google ads to adCenter. Were it not for Google's API restriction, Microsoft could simplify that procedure dramatically, or a third party could do the same.

More generally, the fact that Google has not had occasion to enforce this policy doesn't mean the policy doesn't matter. Suppose Massachusetts passed a law imposing a $10,000 fine on anyone who wears an orange shirt. The next year, perhaps zero people would wear orange shirts, and zero fines would be collected. But from the zero violations observed, we ought not infer that no one wanted to ignore the restriction or that the restriction is important. Rather, we can only infer that the threatened sanction — here, loss of Google AdWords API access, and resulting difficulty in managing AdWords campaigns — is sufficiently serious to deter violations. Carry a big enough stick, and you might not have to use it.

Finally, if the AdWords API restriction is so unimportant, Google might as well remove it — "dead letter" that doesn't make a difference, in your view. Yet we don't see Google in any rush to remove this clause; to the contrary, the quoted restriction has been there for years and shows no sign at all of going away. That tells me that Google wants the restriction and thinks it matters.

Ben Edelman

Great post, Jeff; never have I followed so many links so gladly #:^)

I think that your thesis – that G has a disincentive to rock the affiliate boat – is wrong, though, because it's out-of-date. Everything you said *was* true, but only until mid/late-2006.

Your points paraphrased below, with my comments:

Point #1 – affiliates' #1 chanel is paid search, and that's good for Google.

CZ – aff mktg (& therefore VCLK) was a tracking stock of sorts for search engine mktg from 2000-2006, but that trend has plateaud/reversed starting mid-2006 because advertisers increasingly had sufficiently large inhouse ppc campaigns that they no longer needed or wanted the affiliate to fill the gaps in paid search w/their own ads. So whereas affiliates filled the search inventory *for* advertisers in 2004 & 2005, now the advertisers have filled search directly and from here on out increasingly want to keep affiliates out of search so they can't add inflationary pressure to already high CPCs. I don't have any data on the topic, but I'll bet that the chart you linked to was even more skewed towards search in prior years.

Point #2 – Why would Google want to upset a system whereby their distribution partners do all the dirty work for them?

CZ – again, I would've agreed with you 2 years ago, but Google's affiliate and CSE smackdowns of the past 24 months show that Google rightly realized that affiliates and comparison shopping engines were essentially placeholders for the end-merchants who have since arrived.

Point #3 – G's not enforcing rules for MSA's and affiliates that they've had for years, so that proves Google doesn't want/need to shake things up.

CZ – Remember Friday, April 13, 2007? Lots of MSA and Google-to-Yahoo search arbitrageurs do, 'cause that was the day Google manually destroyed their arbitrage operations, as evidenced by the now well-known story of Geosign's demise. Google did this because they realized with mathematical certainty that fewer ads actually made them more money. Having worked at one of the top two global SEM's from 2003-07, I can attest to that first-hand.

Point #4 – G's T.O.S. doesn't allow for functionality that allows advertiser clients to copy a G campaign to other SE's.

CZ – I was told by a friend at Google that that was an insertion by Legal that no one else was involved with. To wit, Google has *never* enforced that policy, nor have they enforced the policy prohibiting bid mgmt vendors from having a single G/Y!/MSN mgmt interface. I think this is a moot point.

Point #5 – G's convinced people click fraud's not an issue, and there's nothing in the fraud world that can hurt them.

CZ – I actually agree 100% with Google that click fraud is not a problem, mainly because I've seen the ROI characteristics of close to $1B in PPC spend tracked at the keyword-level, to very specific ROI goals. There *is* a fraud problem forcing Google towards affiliate-based traffic buying, though, and that problem is… distribution fraud. Distribution fraud is basically non-search distribution partners finding their way into G's search network, and we're nearing the point where advertisers' collective analytics deployments will allow them to know that distribution fraud exists and is costing them huge sums of money. To combat that fraud and make their market more efficient, Google will need to go beyond SmartPricing and into CPA-based selling of their non-search search distribution partners' traffic.

Not many people are [yet] aware of the following signs of slowing organic growth at Google, but they will eventually:

AdGooroo data showing a slowdown in growth rates for # of new AdWords advertisers;

OneStat data showing that query phrase length trends are reversing for the first time ever. Whereas 2003-2006 saw increases in 3+ word queries, we're now seeing a reversal towards 1-2 word queries = the long tail growth story is over, and search is now becoming direct navigation = trademark terms for which merchants don't need affiliates' help. [see data from OneStat released 6/16/08]

Increased time for click-to-conversion across many AdWords verticals. The macroeconomic situation is starting to affect many areas of search, and IMO Google will need to wring more out of its traffic via aggressive CPA-based selling of traffic to forestall the macro's effects on them.

Broad Match is getting broader, Automatic Match is in deep beta (>5% of all campaigns), and Google can't get anymore European marketshare than it already has; these too are all signs that Google knows it *needs* to keep shaking things up to match investor expectations.