Instant (Bad) Karma’s Gonna Get Ya: AdWords Account Quality At Startup

This one’s for all of you consultants, agencies, and in-house SEM’s who find themselves “inheriting” a sloppily-built account started by someone else.

It’s also a cautionary tale for anyone thinking about “casually” starting out an AdWords account. Hopefully I’ll save you some cash today, but in particular I hope to save you time.

I’d love to say this is a niche topic, but it’s actually a very common situation, since the default approach to trying AdWords is for a business owner (yep, it’s common practice for even a CEO of a decent-sized company) to go in there and plop in a few keywords and ads to see what things cost. The result, usually, is yecch.

That didn’t used to be a problem back in the day. And speaking of the day, this whole pattern of inheriting “yecch” accounts has a nice long history now, doesn’t it? I remember one of our first accounts (we still manage it) was one of those ones that got built by a brainy CEO short on time. “Bill” built a campaign called “Bill’s test,” and while he found the experiment fascinating, it was yecch. He didn’t have the patience to really do a good job, so there it sat until they found a campaign manager.

That lucky web development and yeah-we-can-do-that-too shop, we’ll call Clueless Interactive. The good folks at Clueless doubled down on the yecchiness of Bill’s test campaign, cranking bids to the moon on general keyword terms and adding a few more for good measure.

And so Clueless was out, and presumably went back to building overpriced websites that would later need to be completely revamped by qualified information architects and user interface experts.

Upon inheriting the account, we of course cleaned it up and things proceeded healthily from there in a testing and then monitoring environment.

Dig out, or start fresh?

Try to run through the same playbook today, though, and the “of course cleanup was straightforward” part of the story is far from the same.

That’s because today, people building poor accounts in the early going are actually building up poor account history. It is widely accepted now that Quality Score includes an account-wide component. If you’ve inherited a “dog” account, it could be a long, long time (like maybe forever) before minimum bids on a poor account get down to sustainable levels. It’s as if Google starts out assuming you’re an idiot, but then you have to go and prove it by ringing up piles of untargeted clicks. Now they know you suck.

So if you inherit this puppy, the long struggle to right the ship may wear you down, and the client’s budget may be exhausted if you aren’t first.

But why should you have to suffer indefinitely?

Google reps may tell you in specific cases that even taking into account the time it takes to build up good signals for strong account-wide quality, it might still be better to dump the old account in favor of a fresh, new one.

I’ll repeat that takeaway to be clear. You may save yourself weeks or months of pain by huddling with an AdWords rep who can frankly advise you on whether you should keep an existing account (if its Quality Score history is mediocre to good) or whether you should delete it entirely and start a fresh one (if that overall Quality Score is truly stinky).

(If Poor Quality is demonstrably related to poor landing page and website quality, especially if those are caused by any part of your business being de facto banned by Google, of course the advice above and below do not apply.)

The right way to do it

Reminder: you need to adopt a staged buildout strategy for new accounts in AdWords these days. Start with broad and experimental types of targeting and you’ll be fighting the Quality Score gods. Build up a strong account history with a tightly targeted (smaller but more beautiful) campaign, and you should get the green light to ramp up sooner.

Either way, though, unless it’s a very conventional e-commerce account, it takes time to get Great Quality Scores. Established accounts that are already doing well work differently – initial quality on new keywords can come in “Great.” But on new accounts, it’s harder.

This might hurt a bit

In less mature markets, we continue to see the Quality Score bar set less strictly. But what is particularly pesky, especially in the US where keyword auctions are already clogged with willing participants, is how sensitive and smart the predictive keyword relevancy algorithm is. Certainly, for example, there often seems to be a difference in how users respond to queries typed in the singular (for example, stone mason vs. stonemasons). But should this really be something we have to worry about from a “quality” perspective if we’re targeting reasonably diligently? When a singular term is min-bid at 50 cents and the plural’s at 15, it makes a material difference to our efforts, and not in a good way. Not only do we have to worry about all this stuff, if we want to “have to worry about it less,” we have to pay a hefty premium.

It’s somewhere between a root canal and a painful cleaning at the dentist: you sort of know why you’re there but do they have to make it hurt so much? And do they really think those colorful balls floating in your face, in elementary-school-friendly primary colors, are going to distract you from the pain you feel?

Yep, that is one sensitive and smart algorithm. Is it possible to be too sensitive and smart? Maybe so. Take my word for it, that was me between grades 4 and 6.

Ramp-up costs are now slightly-higher-to-much-higher

Ordinary companies in an ordinary marketing planning process naturally want to get some budgetary sense, not just of theoretical media costs or target CPA’s, but of the total true cost of the initiative, including any investments in sunk costs, technology, startup/testing costs, etc.

Considering the “initial learn and burn” phase, the smartest companies have always been eager to get a realistic sense of what they might burn through in testing, getting a campaign to the point where it is reasonably close to profitability, or close to comparable to other media they’re more familiar with.

As consultants, we’ve always appreciated a frank discussion about this. While we’d love to hit the jackpot of a self-funding campaign from Day One, that’s an approach typically taken by small companies or affiliates willing to look for small slivers of light in a vast darkness of negative ROI. (In fact, that “timid strategy” of “bid everything at 5 cents because you can’t possibly lose if a few of those clicks convert” is part of the reason why Google min-bids new accounts higher – so you have no incentive to come into the auction with a million keywords and an “I only want free money” attitude.)

Testing budgets are tough to estimate, but a company with a new account always had to take a realistic (iterative, intelligent) go at it for 1-3 months. So of course this startup cost could range anywhere from around $1,000 to $150,000+, minus any revenues made along the way.

In a Quality Score world, you have to bid high on some keywords for weeks before they finally “establish quality,” which is Google’s way of saying “build up enough data to reliably signal to the AdWords Quality Score algorithm what your ‘real’ quality is on a given keyword.” And the account as a whole is also establishing quality; to what extent the account-wide overlay affects any given keyword at any juncture is, of course, a black box.

You don’t often hear us in the industry discussing the real dollar cost of this new AdWords ramp-up regime to new advertisers, but they sure must be discussing it on Wall Street and on Google’s accounting department.

Again, to generalize is difficult, but in today’s environment, to the 1-3 month “testing budget” for new accounts, you can probably tack on $5,000 at the low end, and $50,000 at the high end – just to run your account normally, at increased bids to ensure greenlighting (establishing account-wide quality) sometime this decade, or before climate change boils us all stupid, whichever comes first.

Let’s explore that math. Except if they get very lucky with an extremely conventional ecommerce account, many small firms will now see their “get up and on Google” cost increased maybe sixfold. Am I outraged by this? Not at all. Google’s running a successful publishing business that commands high effective CPM rates; it’s not a charity, public utility, or free classifieds site. The current “tough ramp-up” scenario represents an opportunity for committed companies because it sets up a barrier to entry to lazy and uninformed people. But I do wonder if Google had thought all this through when it was developing the “easy entry” messaging that lured so many small businesses in with a self-serve, “ads running in minutes” myth.

For the big guys – they’re not seeing a 500% increase in ramp-up costs. It’s maybe more like 25% or 30%. Again, it’s a no-brainer that companies should do it, especially where competitors allow themselves to be stalled out and frustrated by what will prove, in the long term, to be a trifling sum.

All of the above is a sober assessment of a wallet-bruising reality that proves just how grown up online targeting is becoming as a medium. But that telling of the story is based on typical experience with normal, reasonably well-built accounts. Yes, uh oh. It can get even worse than that.

For anomalous, sloppily built, poorly targeted campaigns, the startup burn can be higher than the above. It might be a little bit or a lot higher. It could even be indefinite. Since no one will endure negative ROI pain indefinitely, why not just avoid the root canal experience and go for that “painful cleaning” as a preventative measure?

What you can do

Here are some smart things new advertisers can do.

You can do it yourself, if you’re somewhat aware of how AdWords works, if you focus on “tight targeting.” However, there are so many moving parts in the interface now, I am not confident that DIY works very well unless you’re hiring a truly qualified professional (in house).

You can outsource to professionals. The professional’s role is multifaceted in this process. They should be able to reduce your “time to green light” and from there, continue to test, build, and improve campaign performance. Beware, however, of the Clueless Interactives of the world, who aren’t schooled in this realm. Some agencies are still no further developed in their awareness of this space than their knowledge of vague media buying principles allows. Meaning, they’ll take the harried client CEO’s hastily built campaign, crank up bids even more, and release new, clumsily targeted campaigns into the wild, putting you way over budget and deeper and deeper into Poor Quality Score territory.

You can request guidance from Google directly. This is not a panacea. Google reps are very professional and helpful. But they don’t know your business intimately and will have some goals in opposition to yours. Service levels, inevitably, will vary by budget but also may fluctuate depending on whether your business model fills perceived holes in Google’s ad inventory or whether it is congruent with their favorite ‘types’ of advertisements.

It isn’t 2002 anymore. You can’t just open an AdWords account, go in there, and play around. Sorry.

Andrew Goodman is the founder and principal of Page Zero Media and author of Winning Results with Google AdWords. The Paid Search column appears Mondays at Search Engine Land.