On my one-year anniversary of working as an AdWords practitioner instead of a Googler, Brad Geddes interviewed me about Quality Score for his WebmasterRadio.fm show, Marketing Nirvana. He asked for my thoughts on AdWords now that I’m a user rather than someone building it, and I thought that was a great question and one worth covering in this month’s column.
I’ve always had plenty to say about the most common mistakes advertisers make, but now that I’m auditing accounts more in-depth and managing a few myself, there are some new mistakes I’ve seen that I think are worth sharing so that you can avoid them in your own accounts.
Handing the reins of bid management to Google or another third party has risks — but perhaps not for the reason I so often hear, which is that Google could use the information to their own advantage. I believe Google wants to use the information to give advertisers the results they want, and I am a big advocate of automated bid management because doing lots of repetitive math and keeping track of lots of variables is something computers are simply better at than your average account manager.
Automating bids is especially useful when you’re working with large accounts; any time you have more than a couple hundred keywords, managing all these manually is just too slow and tedious. But relinquishing all control to the algorithms can have unintended consequences, and it’s useful to understand the risks so you can decide what’s right for your account.
An advertiser I was working with noticed that they lost pretty much all conversions on the one keyword that perfectly describes their business. This happened several months ago, and because they were spending so much time managing long-tail keywords, the drop-off on their money keyword initially went unnoticed.
Only much later (during my audit) did they learn that this keyword had seen a week of poor conversion performance following a landing page redesign. The algorithm noticed it, too, and dropped the bid 60% — effectively taking them off the first page of results. They noticed the conversion rate decrease and quickly fixed the landing pages, but they never looked at resetting the bids for their keywords, so it continued to linger on page 2 of the results. Had the bid algorithm given the keyword another chance, it would have been restored to the first page of results and they wouldn’t have lost a ton of sales.
The takeaway here is that it makes sense to manage head keywords differently than long tail keywords. Even if you put them on automated bidding, you must monitor these keywords closely so that any change in performance will trigger a warning and you can investigate.
AdWords could make this much easier if they allowed advertisers to add charts of keyword-level detail to the interface. Until that time, we have an AdWords Script that builds a nifty little AdWords dashboard in Google Sheets.
The second issue with bid management is that your carefully calculated bids may not be the ones Google actually uses. Because of the vagaries of how Google selects the keyword from your account that matches the query, it’s possible that lowering the bid on a poorly performing keyword simply shifts its clicks to another keyword that still has a higher bid. This is exactly why we built the Traffic Sculptor – it helps find when Google gives an impression to a different keyword than what you intended. This way, you can be sure Google is serving the ad for your selected keyword, using its bid, ad text and carefully selected landing page.
A third pitfall I’ve noticed with managed bids is that they may not always play well with budget constrained campaigns. On several occasions, I’ve seen that a campaign that uses bid automation depletes its budget too early in the day, even when ad serving is set to standard (meaning that ads are supposed to be shown throughout the day). Obviously, when the budget runs out by 10:00 a.m. and no conversions happen, it creates a potential vicious cycle — the bid management tool will eventually reduce bids too much and the campaign may entirely miss serving ads at the times of day when they would have performed best.
Another common problem I’ve noticed is with account structure and keyword match types. I still come across accounts that combine display and search in a single campaign and accounts that have hundreds of keywords in an ad group — both not best practices in AdWords.
But one account structure issue that’s a bit more subtle (and that I hadn’t given much thought until this year) is that if you’re deploying a lot of broad match keywords in an account, it’s challenging to see trends or answer questions about why results change. That can make it hard to know if your strategy is panning out. As much as I like broad match keywords for their ability to help discover what users are really typing into the search box, they can make life for advertisers pretty hard….
We’ve all heard that CPCs change based on seasonality or because Google introduces things like Enhanced Campaigns. But knowing how this impacts your account is trickier than just taking a look at the average CPC. The reason is that most advertisers use some degree of broad match keywords in their accounts, and stats for broad match keywords are averages that include data from each of the hundreds of queries that matched the broad keyword. We’ve all heard the adage that averages lie, and broad match stats are averages that obscure what’s really happening.
For example, when you change the bid for a broad match keyword, you’re changing the query mix because it becomes eligible for a different set of queries. The different queries have different CTRs, conversion rates, etc., so it quickly becomes a nightmare to figure out why a keyword’s performance has changed. Was it a competitor, a change Google made, an optimization you did, or simply the fact that your query mix changed? If your results are dropping and you can’t pinpoint the reason, it’s hard to correct the course. So unless you’re analyzing exact match keywords, you’ll have to dig very deep into the search terms report to try and come up with an answer.
To further complicate things, even exact match keywords may see sudden changes in performance because of the aforementioned issue where Google may shift the clicks to another, similar keyword with a better rank. In this case, the best way to understand what’s happening is to run a query report rather than a keyword report.
These are a few of the things I’ve learned in 2013. Hopefully, you’ll avoid making some of the same mistakes in 2014 and beyond!