Loading ...
Imagine going into a casino and being met by a friendly hostess at the front door. As you approach, the hostess smiles and says, “Hello, thanks for coming in. How much money would you like to lose today? Just write me a check and I’ll maximize the amount of fun you have losing that money!” Not many people would like that sort of program, but that’s an apt analogy for the recent “dynamic ads for search” program Google has recently announced. While it’s currently only an optional beta, if this is the direction AdWords takes, it would be disastrous for advertisers, agencies, and, well, everyone but Google.
According to a recent AdWeek article, dynamic ads for search eliminate the notion of keywords or queries entirely. Instead, “With the new product, advertisers just point AdWords toward the pages that they want to promote. Then Google matches the ad with the best searches and generates an appropriate headline.” As a Google product manager tells AdWeek:
“(Baris) Gultekin predicted that this could eventually represent ‘the future’ of AdWords, because it’s a more natural way to think of ads—businesses usually start with a product or Web page that they want to promote, then with AdWords, they have had to think of a long list of relevant keywords. With Dynamic Search Ads, Google can take ‘the middle layer of keywords’ out of that process.”
Ah yes, that dreaded “middle layer” of keywords, which, by the way, happens to be the raison d’etre of AdWords’ incredible success in the first place. Why? Because keywords enable smart advertisers to precisely target users based on the intent they express in their search query. Smart advertisers quickly learn that the phrase “mortgage rate” has totally different intent than “mortgage rates,” and that “buy night stand” is totally different than “one night stand.”
Of course, there are many advertisers who I would not classify as “smart,” at least when it comes to AdWords. These advertisers typically buy too many keywords on broad match (resulting in lots of irrelevant clicks and additional revenue to Google) or too few keywords on exact match (resulting in missed opportunity for the advertiser and missed revenue for Google). My assumption has always been that keywords had a net positive impact for Google’s bottom line (i.e., that most advertisers err on the side of being too aggressive than too conservative). If my theory is correct, why would Google ever want to abandon keywords at all?
I believe the answer comes down to reducing transparency. With keyword-level and query-level data, smart advertisers can deftly pick and choose where they want to show ads and create highly efficient campaigns. Take away the ability to control spend at a keyword/query level, and these smart advertisers are now knocked down to the level of less smart advertisers, the ones who are already happy to buy broad-match keywords and assume that Google will match them on relevant queries. To put it another way, “dynamic search” would have no impact on dumb advertisers who are already running inefficient campaigns, but it would make it harder for smart advertisers to continue to run at their current levels of efficiency. Translation: less transparency for advertisers, and more control of revenue for Google.
Google notes in the AdWeek article that advertisers can still run keyword-focused campaigns, even if they are in the dynamic search beta. And they also claim that those who have opted in to dynamic search are seeing a 5-10% increase in conversions. But as Google’s product manager notes in the article, this could be “the future” of AdWords, so keywords might someday be a distant memory.
If AdWords turns into a system where advertisers show up at the door with a check, and a friendly hostess takes their money and then magically delivers the “most relevant” ad text, landing pages, and bids, it is truly no different than a casino placing your bets, or a fox guarding a henhouse. I don’t like it one bit.
- David Rodnitzky, CEO, PPC Associates
- Questions? Comments? Email bloggers at ppcassociates dot com.
Good post, although I mostly agree with you, there may be cases in which this feature could work, as long as conversion tracking is working properly. For websites with thousands of products this could be a way of discovering hidden gems in the product line, these could then be optimized manually for even better ROI.
Thanks for the comment, Carlos. I do agree with you that for large ecommerce accounts with structured data, an automated solution – at least in theory – could work well. I believe companies like adchemy and datapop are working on third party solutions to this.
As long as this is just an option, so be it. I’d be very concerned if this really was “the future” of AdWords.
Thanks or the heads up about dynamic ads. Funny, but everytime I test an Adwords automated product, my spending goes up. As an example, my test of automatic bidding was not a disaster, but I ended up hitting my budget limit every day without enough of an increase in conversions to justify the increase in spending.
Randy, you are not alone in your experience. I think Google comes out with some great innovations, but whenever they try to wrest control of ad campaigns away from advertisers, it only seems to work out in their favor!
Great post – thanks. My fall back position with Google’s revolving door of “innovations”: “Distrust, Then Verify!”
Excellent analysis. I can’t think of anything less inviting than issuing a ‘blank check’ for Google. It looks like another case where they’re building more ‘gotchas’ in their system, rather than adding value.
Agreed Barb and Darren. There is often a debate as to whether Google’s motto is “do no evil” or “do KNOW evil”. Mostly I think they have good intentions, but sometimes you have to wonder!
I’m just a transplanted Blogation groupie, but as such I’m aware you’ve been predicting the end of the keyword for some time. I for one always thought it would come part and parcel with optimization based on deep pattern-matching (probably implemented using Hadoop). The story from Google would then be, “we can get a higher conversion rate, but you’ll have to trust us, because to get it we are using behavioral data we can’t in good conscience share with you.”
p.s. David, in case there was any doubt, that Hadoop reference was just for you.
Steve, I would have been more impressed if you had referenced “Map Reduce” instead of Hadoop: http://en.wikipedia.org/wiki/MapReduce
[...] Google dynamic ads – Will they go live? (David Rodnitzky hopes they don’t.) [...]