Google – the leading digital advertising company on Earth – is poised to release a new wave of innovations to build on the array of tools and formats already available. There has been a steady stream of buzz about new ad formats, new tools like Wallet, more advanced behavioral targeting…and that’s just the tip of the iceberg.
Lest we get carried away talking about Android, YouTube, Plus, and other big bets the company has made, Google sometimes cites a figure related to its resource allocation: 70 percent of its resources remain dedicated to search. Many advertisers share that perception: for them, Google is 70 percent search.
Why, then, if search is still so big and also “old hat” to Google, is it still so easy to bleed money in the core AdWords functions we need to perform as a routine part of optimizing an account? Why are so many advanced features being added, while basic fixes to workflow that affect testing and targeting are not being fixed?
Here are three ways that the AdWords platform is (still) costing you money. Watch out for them, and while you’re at it, ask Google why it doesn’t fix it.
1. There is broad match bias everywhere. I’m willing to speak to the sometimes surprising virtues of broad match, but some accounts find themselves getting in recurring hot water as broad match dilutes the ROI in the account. The problem seems to be our initial overconfidence that a sizeable list of broad-matched terms will continue to match in about the same way a few months from now as they did when they were first added. Many times, they don’t. Wasteful matches reappear and persist until you come back and stamp out the small fires.
Ambitious account managers with above-average (but not superhuman) initiative may keep on top of most of these nagging issues, but “it comes to something” (to quote the random old woman in “Repo Man”) when even power users of the platform tire of the constant reversion to broad match in many places in the workflow. In a hurry, they may toss in a bunch of new phrases with the expectation they’ll “come back later” and fix it up. They had better come back. All it takes is one or two broad-matched phrases to start going to town with expanded matches to eat up proportionally more spend than you want or would expect!
(Regarding negative keywords or exclusions, be mindful that some account structures and the use of too much broad matching can open up a bottomless pit of optimization challenges. Negative matching isn’t necessarily an exciting strategy, just a necessary one. Rather than leading you to veins of rich opportunity beneath the surface, a poor account strategy can be more like digging into a landfill full of junk that needs to be tossed out again and reburied. Delightful!)
Broad match ubiquity will cut into the effectiveness of some of your outstanding work. For example, one of the most thrilling features of the platform is in the “view all searches” report under the keywords tab: you can highlight a long list of torso and long-tail phrases that successfully converted, yank them out as separate phrases, and bid separately on them. Unfortunately, the default is broad match, so if you added 20 to 30 phrases, you’ll now need to take the trouble to manually change them all to phrase match. Why doesn’t Google add a one-click option to make them “all phrase match,” or simply make phrase match the default?
Finally, it’s worth noting that broad match modifier, which is broad match without the pesky expanded broad matching, is more predictable than regular broad match. That being said, broad is still broad, and even minus the expanded matching bundled with regular broad match, you still open yourself up to a range of unpredictable queries. For example, even a cautious bid on +ontario +place will potentially show your ad against a search query like “place to buy cookware in Kitchener Ontario,” etc. All those unwanted matches can wear you out.
2. Even ad rotation. Conversion Optimizer and AdWords Campaign Experiments allow you to reduce the impressions given to the experiment group(s), so the testing process doesn’t cost you too much money and disrupt the business metrics you’re accustomed to. Unfortunately, you don’t have that type of control over ad rotation when you set it to “show ads more evenly.”
Extreme ad testing is often done by relative newcomers to the industry, and one of their favorite look-busy tactics might be to enthusiastically test a high number of new ads in ad groups. Taken to extremes, this dilutes quality scores and ROI while testing is going on, and ends up hurting clients. I know that Google wants to give us settings to automatically select the “best-performing” ad, but many advanced testers are pretty hands-on with it. What we need is a setting for power users that allows advertisers to split the traffic among ads in rotation, in any proportion you like. AdWords Campaign Experiments (ACE) does something like this, but it’s not the same thing, and ACE is definitely an acquired taste.
In the meantime, continue to use whatever kludgy workarounds you can to mitigate the financial impact of testing.
3. Automatic matching: has it regressed? To borrow an analogy from the recent housing bubble, you can securitize and bundle subprime any way you like, but some investors feel inherently safer when they can look inside and decide for themselves which assets they want, and which ones they don’t. Investors didn’t have to believe and trust Moody’s and S&P “AAA” ratings. And similarly, advertisers don’t have to trust anyone’s stamp of approval on a vast network of publishers. Even if there is a sophisticated methodology involved in the targeting and pricing of the media, there’s also a leap of faith involved.
For performance-based advertisers who have gradually built files of Managed Placements (which are targeted if you select a network targeting method now called “specific reach”) that perform like clockwork, bravo! But as a means of discovering good publishers and pages, the Automatic Matching flavor (now called “broad reach”) seems to have regressed. Junky, geographically inappropriate publishers, etc., simply shouldn’t be showing our ads and showing up in our reports…at all. In many cases, not only do they show up, they show up by the thousands.
It’s no better when Google “spikes” you with YouTube, Gmail, Maps, and other Google-owned property placements you weren’t expecting. Sure, they’re plausible, but the relationship to ROI often proves specious. Even with so-called broad reach, some power control features would be nice. And really, you should have to opt in to channels like YouTube. The way things stand, it can unexpectedly dominate your spend, generating 50 percent or even 75 percent of the clicks in a campaign.
A key pitfall to the broad reach scenario today is that optimization efforts are constrained by the long-tail-iness of the net you’re casting. You may see thousands of one- and two-click publishers in your reports. Hard to come down on any one publisher as being “bad” in that scenario, and hard to optimize and improve by eliminating any of them. The scientific methodology that goes into the matching attempts is consolation if the group as a whole has a dismal ROI. It’s a shame that the practice here seems to belie the potential; we don’t mind paying low prices for lower quality inventory, but we don’t want terrible inventory. We also don’t want to have to exclude hundreds of publishers one at a time, and then hundreds more over time in a game of endless whack-a-mole.
As much as the remaining community of MFA (made-for-AdSense) publishers might scream and moan about Google’s draconian ways, most advertisers aren’t interested in those sites, period. If anything, Google remains too lenient with marginal publishers. Should Google be doing business with them at all? As “liar loans” were to investors in securitized consumer debt, “rogue publishers” are to advertisers.
It’s counterproductive to speculate much more on the “why’s” of the foibles in the AdWords platform, or on the “when’s” of future feature releases or revolutionary changes to the platform. In the here and now, as always, you need to plug the leaks in your campaigns. I hope these reminders have been helpful.
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