This is the subhead for the blog post

Over the last few months, Google has released a series of enhancements to AdWords that have enabled advertisers to see the full value of their AdWords spend. In particular, there are two improvements to the AdWords user interface of note:

  • “Many per click conversions” – you can now see if a user has purchased more than once from your site during the AdWords cookie window;
  • “View through conversions” – when someone sees a banner ad on the Google Content Network, but doesn’t click, and then later converts on your Web site.

These are useful metrics, don’t get me wrong. The more I can see the true number of conversions that resulted from a click, the more I can bid on that keyword and thus capture more market-share. And of course that’s great for Google too, as they give advertisers more reason to spend more with Google.

Interestingly, however, when it comes to metrics and tools that would help advertisers save money, Google has been slow to implement needed improvements. Here are a few examples:

  • The current keyword research tool has a drop down box that enables you automatically add negative keywords to your campaign or ad group – the new version of the keyword tool no longer offers negative keyword addition;
  • Despite the fact that Google Analytics easily integrates with Google AdWords, there’s been no effort to provide “multi-channel attribution” – that is, to only give partial credit to Google when a sale has resulted from other Web sites, like YSM, a comparison shopping engine, or a banner campaign.
  • The new (and useful) “see search terms” functionality in the AdWords UI auto-defaults new negative keywords to “exact match” and auto-defaults new keywords to “broad match.”
  • The geographic performance report in AdWords reporting only allows you to look at data on a “daily” basis, which makes it virtually impossible to draw conclusions if you don’t have an Excel macro or pivot table to roll data into a summary level.
  • Any AdWords report that includes conversion data will not give hour of day information, making day-parting impossible without log files or other external data.

In other words, Google gives you lots of reasons to increase bids, but does an average to poor job of showing you reasons to cut back on CPCs. The problem I have with this situation, is that it ends up giving advertisers a false sense of success. Let’s look at a hypothetical example.

Let’s say that you spend $100 and Google shows that you have five “multi-conversions per click” and three “view through conversions” for a total of eight conversions. So that’s a $12.50 CPA. But now let’s add a few more metrics currently not available from Google. First, let’s say that an analysis of multi-channel attribution shows that four of the conversions also visited YSM and two of the conversions visited If we give 50% attribution to Google and the other converting Web site, six of Google’s conversions should really only be counted as three conversions, bringing the number of conversions from 8 to 5 and the CPA from $12.50 to $20.

Next, if we had readily digestible geographic data, we could have discovered that 20% of our cost came from states where we have never received a conversion in the last six months, and if we had accurate hour-of-day conversion data, we could have excluded 10% of clicks. Combined that’s a 28% savings on spend, so our five conversions should have only cost us $72.

The end result: the advertiser should be paying $72, and counting five conversions, but instead pays $100 and thinks that he has received eight conversions. It’s a win-win for Google (more spend and more conversion credit).

I don’t want Google to get rid of their tools and reports that help us see more data, I just want them to work equally hard to deliver information that may not always improve their bottom line. Giving people the good news without the bad news is OK in, say, North Korea, but I think its fair to expect a little more from Google!