As valuable as the Dimensions tab has been, it doesn’t solve every PPC issue. While Google does a good job of recognizing the major segments applicable to all accounts (search, content, geos, time of day, etc.), you’ll need custom filters to group the pieces that matter most to you.
Filters will allow you to quickly roll up data for various segments at the campaign, ad group, ad, and keyword levels (filtering is a bit more limited when it comes to placements, so Excel manipulation is advised – we’ll cover this in a future post). You can select from a variety of different metrics and characteristics – leveraging CPA and ROI metrics to weed out poor performers at all levels or, more applicable in our case, text filters for grouping similar campaigns.
Google can’t inherently tell the difference between your brand terms and your non-brand terms. You need to tell them what makes up each segment, and the best way to do this is through labeling. As you’re starting to figure out, efficient campaign check-ups are largely dependent on how you structure things in your account, so having a consistent method for grouping and labeling campaigns (and ad groups) is essential. See below for an example of how you might label things.
This naming convention enables us to set up a series of filters we can use in the future. With this setup, we can quickly isolate search, content, brand, non-brand, topics, audiences, country, images, text, as well as alpha and beta terms (this terminology is central to PPC Associates’ Alpha/Beta build methodology for account structure – white paper available here).
Executing the filter will provide you with visual trend data (graph) as well as rolled-up data for the segment. As quick check-ins go, these two things should be enough to make sure everything is on track. I tend to default to some combination of Conversion volume/ROI metric (for the graph), which is enough to catch major shifts in either direction. The real challenge is that Google does not enable the user to apply any segments to the summary row for what you’ve filtered. You can apply the segments to the individual pieces but not the filter’s total.
While you can still see trends for individual pieces, and try to correlate them to the trend image the graph option provides, you probably won’t get enough detail to take action. In this case, it’s best to revert back to the Dimensions tab, adding whatever parameters you need and take this analysis to Excel. These in-depth checks should be executed regularly, but your filters should be run with enough frequency that you can immediately tell when something is going wrong.
Digging through large data sets is an opportunity to let your inner quant geek go wild. Manipulating data and finding hidden trends is akin to searching for buried treasure. It’s long and arduous, and there are lots of clues and false leads, but when you get there, the satisfaction is immense. As much fun as this is, these daily checks are meant to catch trends faster and act on them in the moment. Catching the culprit (or profit driver) early makes for more efficient work down the road. Some new variable is bound to be introduced and obfuscate the real cause of a trend. Set up a few filters and see what the impact is on your day to day. Chances are, you’ll save yourself a lot of time by doing a little more work up front.
– Sean Marshall, Director of Search Engine Marketing