It’s been a long while since anyone fully believed in the old, innocent “Don’t be evil” image of Google. The last 28 or so times I checked its quarterly financials, it appears to be running a business. And how!
Part of Google’s mission in “running a business” has been optimizing the PPC auction – and indeed, the entire search results experience – to maximize revenue. Doing that has often meant considering the user experience as well, to be sure. Google gets that far better than competitors ever did. But it’s also very good at squeezing this thing to extract maximum profit. Because that’s so obvious, no one is hesitant to call a spade a spade anymore: certain features of the AdWords platform are aptly deemed “revenue grabs.”
Lately, though, Google’s march to profitability is starting to look more like a square dance. Swing your partner aft and fore, make her cough up ten cents more! If she checks performance stats, she bids it down and that is that!
And don’t forget the do-si-do. Never forget the do-si-do.
What am I saying?
Google’s “partner,” of course, is the advertiser. In all of the engineering that’s gone into satisfying the user (searcher) experience while maximizing revenue, it’s the advertiser experience that has so often been taken for granted, at least as compared with the user experience.
Take Google’s complicated keyword Quality Score formula, which – given new and helpful levels of disclosure – turns out to be more complicated than one may have guessed. If my theory is correct, there is a far-reaching user experience theme being pursued as part of this mix. The relentless push for ad quality moves beyond modeling discrete, one-off user experiences. It also strives toward a model that seeks to impress on users “It’s OK – you’re safe here” (over here where we keep the ads, that is). If Google had engineered things in a more simplistic way, chances are the user would already have developed banner blindness. So: good for Google, and yes, in that regard, good for advertisers.
But financially speaking, Google has recently been faced with a problem of relatively simple origin. Google being Google, it figured it could engineer its way out of it. It’s not clear it can.
The problem stems from three basic causes. First, at some point click prices will reach their limits and advertiser budgets for certain keyword spaces will be maxed – period. Prices have to level off at some point.
Second, business confidence can create much less willingness to take chances – or put another way, a greater interest in revisiting assumptions about what are “fair” supplier prices, including what seems “normal” or “crazy” for a certain type of click. (Yes, advertisers can learn what is normal by watching performance stats, but in a dynamic environment where prices are falling, advertisers may develop a sense of “reverse confidence” – that if they drop their bid to return to more favorable ROI, weak hands in the auction may soon discover that they want to drop theirs, too.) We ran into a global business confidence crisis three-and-a-half years ago, and it’s been in the doldrums ever since.
Third, it’s precisely because more advertisers are measuring performance that Google will have more trouble fooling people with platform gimmicks. More and more advertisers doing a better job at measuring is largely good for Google, because it puts the price of a click on a solid foundation. If one advertiser decides clicks are too expensive, the market is smart enough to step in and buy the same click for roughly the same price. But has Google forgotten that more measurement leaves fewer stupid advertisers in the auction whose pockets you pick when you anticipate a weak quarter or two?
Starting around 2010, it appears that Google began unrealistically squeezing revenue in some areas, sometimes apparently in a hurry to reach short-term revenue objectives. Many advertisers absorbed and initially accepted Google’s claims for appropriate bid levels on channels like remarketing, for example. Then, it discovered the new level of the auction was 75 to 80 percent lower than it’d been led to believe, and went right back to trusting its own instincts and data on appropriate bid levels.
In addition, Google seems to have become vaguely alarmed by the general decline in click prices caused by the one-two-three punch of maxing budgets, declining business confidence, and more careful measurement. Picture a large graph on the screen, showing a typical average click price for a keyword of interest, falling half a percent a month for most of the past 18 months. Then picture fear and loathing at the Googleplex.
To head off the revenue hit that might result from the gradual but relentless freefall in overall CPCs, the natural Google response seems to have been to go to the product teams and ask them to “Engineer something!” to improve cash flow.
One such recent move is the widely-denounced move to make it much harder to “rotate” ads to test them. A certain percentage of advertisers will be conned into automatically reverting to the Google-friendly “optimize for clicks” setting. We’ll leave that aside for now.
Little tricks like “first page bid” annotations and the new “close variants matching even on phrase and exact match” are quite likely to have the predictable short-term effect of any engineered “feature”: they’ll create artificially high prices for some advertisers and put more money in Google’s coffers.
“First page bid” has been an ominous notation to most of us for some time now. This notation shows up next to a keyword when you lower your bid below a certain point where Google predicts you won’t always show up on the first page of search results because your Quality Score multiplied by your current bid is below that level. You’re supposed to go “Whoops!” and bump your bid a bit higher. Congratulations: Google’s framing exercise has tricked you into bidding inaccurately. It’s taken years, but advertisers are now ignoring this warning and trusting their own metrics. The main takeaway is that the notation often appears to be inaccurate and misleading. Your click volume might not decrease as much as you expect if you drop below the artificially-labeled threshold.
The new violation of phrase and exact match conventions is another revenue grab. However minor it may be, it’s important to consider it to understand the dynamic of click pricing.
Currently, a healthy percentage of advertisers who do not understand matching options, and who (in Google’s humble opinion) are targeting too narrowly to achieve the reach they should be going after, should be accepting “close variants” like plurals, verb stems, and misspellings in phrase and exact match, not just broad match. Such advertisers will (Google hopes) remain opted in to the less precise versions of exact and phrase match.
But those advertisers will then be seeing their ROI numbers get a little worse, just like they do if you use broad match without understanding it. Google’s effective CPM certainly rises, at least until these advertisers’ overall satisfaction level with AdWords drops, or until they simply assess their keyword CPA numbers. If Google’s eCPM is rising in a push-pull relationship with advertisers who are simply measuring what they take to be the exact ROI numbers on each keyword, then bids on those keywords will come down. Advertisers like this will be using a blunt instrument to mitigate the impact of Google-friendly feature tweaks. More precise advertisers who use the search query report, negative matching, all match types, and who will opt out of this new matching feature will do the best.
Long term, then, Google won’t see much of an improvement in the average CPCs on the same queries. It’ll be right back where it started, because increasingly, advertisers do a lot more than just measure visits, CTRs, and CPCs.
The problem Google faces is that, at the end of the day, it’s running an auction. It’s an auction with rules that advertisers are free to follow if they like, including bidding lower. It’s not a perfect market – it’s rigged in some ways – but it’s enough of a market that if click prices want to fall, they’ll fall…as surely and steadily as gold has fallen right on the heels of predictions that it’s going to $5,000 an ounce. Markets are powerful, and they often run on momentum and expectations. Any advertiser who has enjoyed consistent volume and rising profit in the past year while 90 percent of their bid change directions have been down and only 10 percent of them are up is now deep in the thick of timeless market psychology. Google is fighting a strong trend with short-term measures to trick some advertisers into paying more.
Why bother? Even if a solid 25 percent of advertisers remain stupid enough to leave some of the unfavorable features running, the rest will adjust, and the typical price for a click will go right back to where it was. Smart advertisers might get even more bargains, because the confounding nature of the auction will so frustrate the non-optimizers that they will simply assume AdWords “doesn’t work for them,” and in the classic “all or nothing” mentality of less accomplished PPC advertisers, will exit the auction entirely, bumping prices lower, not higher.
When you’re just starting out as a business owner it’s easy to become wrapped up in the seemingly endless number of metrics ... read more
Visual search on the web has been around for some time. In 2008, TinEye became the first image search engine to use ... read more
We’ve written an awful lot about Google’s open source accelerated mobile pages project (better know as Google AMP) over that last 12 ... read more