How A “Good” AdWords Campaign Structure Can Hinder Advanced Insights & Strategies

While there is disagreement in the paid search community over the best way to structure an AdWords account, columnist Andy Taylor discusses some common variables to consider.

Chat with SearchBot

google-adwords-bigA4-1920

Optimal campaign structure is an interesting topic of discussion, complete with different takes across the paid search industry and Google’s own recommendations.

The ways in which bid modifiers can be set in Enhanced Campaigns should influence how advertisers structure their accounts. Further, understanding how Google calculates its estimates for conversions influenced by paid search (such as in-store visits) can also prove valuable in deciding on a campaign structure that works best for each account.

To expand on the pros and cons of different setups, let’s walk through one hypothetical example of the considerations an advertiser might take into account in deciding on an AdWords account structure.

Mobile Value By Collection Of Keywords

Let’s imagine a retailer with a large brick-and-mortar presence and established paid search program — Gray Blanket Company (GBC), the leading purveyor of gray blankets and only gray blankets.

GBC has a well-managed paid search program: great keyword list, advanced bidding system, solid team of analysts digging in and making optimizations. As such, it uses mobile bid modifiers to effectively bid all of its smartphone traffic to meet the same ad-spend-to-sales ratio as its desktop traffic.

Cumulatively across all non-brand keywords, let’s say smartphones bring a “value per click” relative to desktop right in line with Merkle|RKG’s Q1 Digital Marketing Report figure of 37%.

 rkg-q1-2015-paid-search-mobile-vs-desktop-rpc

However, in digging into performance, the crack team of analysts discovers that mobile performance relative to desktop varies more based on category than it does by keyword specificity.

(Note: There are lots of categories of gray blankets — short, tall, fleece, terry cloth, wool-lined, water-resistant, etc. — and GBC sells all of them.)

Gray Blanket Company’s Current Setup

Currently, all of the company’s campaigns are set up by keyword specificity, with SKU level terms across all categories in one campaign, generic terms in another campaign, etc.

Campaigns may end up being set up this way for any number of reasons (including simple analyst preference). In GBC’s case, it’s the result of first building out head terms across all of its categories by hand in one campaign, then later using information found in the product feed to make for faster keyword creation of longer tail terms in a separate campaign.

As GBC wants to set its bids as granularly as possible, it chooses to use ad group-level mobile bid modifiers (currently the most granular possibility) as opposed to campaign-level mobile modifiers. For ad groups that lack enough traffic to assign a mobile bid modifier based on just that ad group’s performance, GBC has been using campaign performance to calculate the modifiers.

Should GBC reorganize all of its campaigns to be based on the category instead of keyword specificity in order to assign better mobile modifiers to these ad groups? If the difference in smartphone value relative to desktop is great enough, then it should definitely be considered.

Geo Performance Varies By Category

Another trend the GBC team has identified is that different categories of blankets perform differently across various geographies. For example, the fleece and wool-lined blankets do great in the northeast where the winters are harsh, but aren’t as popular in warmer areas. Other, lighter categories of gray blanket don’t see the same trend.

In Google’s Enhanced Campaigns, geographic bid modifiers are set at the campaign level, unlike mobile modifiers which can be set at both the campaign and ad group levels. As such, GBC is seeing even more reason to reorganize their campaigns based on the category of the keywords as opposed to specificity.

Let’s say they go ahead and pull the trigger on organizing their ad groups and keywords into campaigns based on the category.

But Mobile Performance Varies By Geo?!

After a campaign reorganization, one of the analysts comes to the team with some new data that suggest that mobile performance varies based on attributes of the geography searched from. GBC finds that its mobile value relative to desktop varies hugely by the average household income of the area searched from.

Since Enhanced Campaigns don’t allow advertisers to set different mobile modifiers based on geography (as in, say, bid smartphones down 40% in New York but bid them up 10% in Virginia) within a campaign, the only option to account for this is to divide the category campaigns up into separate campaigns targeting different areas based on their geographic attributes. This is a real headache, but the difference in relative phone performance is so significant that GBC decides to go for it.

The team is optimizing geo and mobile bid modifiers like never before and feeling pretty good about it, until they notice…

Where Have All The In-Store Visits Gone?

GBC has a huge hypothetical brick-and-mortar presence that spans across the entire U.S. It knows that its paid search ads have had at least some impact on its offline sales, but it’d always had trouble with offline attribution, and holdout tests just weren’t giving conclusive results.

Then Google rolled out in-store visit estimates as part of its Estimated Total Conversions. The GBC team jumped on getting in-store visit tracking rolling for its account, and had been enjoying several months of being able to see estimates of just how many visits had ties to a paid search click.

But just after the team members divided up the campaigns to target geographies based on the average household income of different areas, they found that they were no longer seeing any in-store visits attributed to their campaigns. Had they broken something?

Talking to their AdWords dedicated account manager, it became clear that in dividing up their campaigns, they had split up their traffic such that the data for each campaign was no longer significant enough to receive in-store visit estimates, which are tracked and estimated by device for each campaign. Without a campaign reorganization or significant uptick in store visits driven by each campaign, they would no longer be able to see in-store visit estimates.

Do they give up some of the bidding optimization made possible by their current campaign structure in order to go back to seeing those estimates?

The Impact Of Campaign Structure Is Complicated

All of this is to say that an account’s campaign structure can impact a number of different variables in terms of bid optimization and data collection. While the structure may change as a result of different initiatives over the course of a program’s lifetime, it’s crucial to set up campaigns to succeed given the current rules of Enhanced Campaigns.

Driving the best results requires that real thought be put into how ad groups and campaigns are organized, and the best structure stands to be different for each program.

Even the dullest of product selections may see significantly different performance based on variables that can only be optimized for at the campaign level. As GBC’s hypothetical company motto goes, never judge a blanket by its grayness — what lies beneath the surface may be far more important.


Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.


About the author

Andy Taylor
Contributor
Andy Taylor is vice president of research at Tinuiti, focused on creating unique views into digital marketing performance based on $3 billion in annual ad spend under management. In addition to Search Engine Land, his work has been featured in major publications such as The Wall Street Journal, Bloomberg and The New York Times, among many others.

Get the must-read newsletter for search marketers.