I had a great time participating in SMX West’s Great Structure Debate with 3Q Digital’s Caitlin Halpert and CMI Media’s Justin Freid, moderated by Ginny Marvin. My general take on structure is that it really isn’t one-size-fits all. (If you’ve heard me present before, you’ve probably heard me use the same one-size-fits-all complaint regarding other subjects, too: remarketing, for one, among others.)
In basically every aspect of paid media, I really, truly believe that most clients warrant a custom set-up that lends itself to the client’s goals, company structure and – if available – past performance. On the other side of the same coin, I know we are all tired of the token “It Depends” response. (Kudos to Matt Umbro for taking a stand.) Because we’re all tired of it and because frankly it isn’t a helpful response, I wanted to make sure that my presentation gave actual if, then examples for when and why I use different structural strategies for different clients.
In case you didn’t get a chance to check out our session, allow me to provide some context. I try to take the following into account when I’m building a structure:
- Company Goals
- Company Structure
- Geography (And Language)
- Products or Services
- Past Performance
The first four are probably fairly self-explanatory but the performance piece might not be quite so obvious. When I look at performance, any sort of oddity could trigger a restructure – some restructures smaller than others. Also, let’s be honest – we have to make sure that we weigh the potential gains of the restructure against any temporary performance dips that we might see. If we did large scale restructures willy-nilly all the time, we’d probably hurt more than we’d help. That said, these are the performance gaps that I’m usually looking at to determine if a restructure might be warranted:
- General keyword performance: Are there any high volume keywords that have big pain points? Possibly a higher-than-average CPA or a low quality score? Sometimes these are worth a minor restructure to create smaller, more tightly themed ad groups. Sometimes even single keyword ad groups (SKAGS)!
- Geo Performance: Are there any geographic outliers? We’ve had success in separating out high-traffic, high-competition cities from broader area campaigns in order to better control campaign settings that impact CPCs and CPA.
- Match Type Performance: Are there any stark differences in the way that different match types are performing? Sometimes if broad match has a significantly higher CPA, it can be worthwhile to pull it into its own campaign. This is especially true if budgets are capped so that you can ensure that exact and phrase match have full run of the budget that they are capable of spending with broad layered in on top.
- Search Query Mapping: Are queries mapping to the most optimal keyword possible? Based upon how your queries are mapping, you may want to pull certain keywords into new ad groups so that you can use a negative to direct query traffic to the best keyword. If you don’t actively check in on search query mapping, it’s something I highly suggest digging into. Here’s a post that details the value of search query mapping and another post that details how to use Excel’s subtotal function, which is my preferred method for reviewing mapping issues.
In summary, I don’t really believe in SKAGs as a full blown management structure but I do have a few SKAGs as warranted by historical performance data.
I also highly suggest that advertisers check their performance data regularly to determine if minor restructures might improve performance. Furthermore, although full account restructures aren’t taken lightly and shouldn’t be performed frequently, accounts can outgrow their original structure (and unfortunately some accounts are poorly structured from the get-go).
When considering a large-scale restructure, I suggest testing a few campaigns and monitoring performance before launching a full-scale restructure. This allows for data collection to ensure that the new structure will outperform the old structure. As much as it stinks, some accounts just have enough history that they defy best practices with their old, horrible structures compared to new, organized structures. Testing a few campaigns at once helps to mitigate risk.
Check out the slides for actual examples of these principles in practice!