Cross-Campaign Traffic Segmentation

Cross-Campaign Traffic Segmentation

Thomas Byskov Madsen & Lasse Rasmussen - November 3, 2017

In standard search campaigns (with keywords), segmenting different traffic types goes without saying. The reasons for doing so include: ad messaging, bid management, reporting and providing the account specialist with a better general overview. In Shopping campaigns, the main reason for segmenting traffic is that different traffic types have a higher expected value than others, which should ideally mean different bids.

As mentioned, the most common segmentation is to split brand and non-brand/generic search terms into different campaigns. In Keyword campaigns, this segmentation is easily done by adding the brand(s) as part of the keywords in the brand campaign. However, this has to be done quite differently in Shopping campaigns, given that there are no keywords.

Using the Campaign Priority Setting

Rather than using positive keywords, a combination of negative keywords and the Shopping-specific setting: ‘Campaign Priority’ can be used. This can be set at three different levels:

 

3 levels of campaign priority

 

What is unique about this setting is that it is the sole factor for determining which one of multiple campaigns will display an ad (assuming that they are all eligible otherwise to display ads for a search term). This is the opposite to keyword campaigns, where the mix of Max Cost per Click and Quality Score are the deciding factors. This means that, all other things being equal, in a scenario involving two Shopping campaigns that are both entitled to display an ad for a specific search term, the campaign set to the highest Campaign Priority level will always be the one that displays the ad. Note however that ‘Campaign Priority’ has no effect in relation to how one company’s ads are shown against the competitors’. So how can we use this priority setting to segment certain traffic types?

Brand vs Non-Brand Segmentation

Let’s assume that we want to set up the common traffic segmentation of brand traffic vs. nonbrand/generic traffic. In the traffic flow chart below we have two different search scenarios representing the two traffic types – one where the search term contains the brand, and one with a generic search term.

 

Brand vs non-brand segmentation

 

As priority comes before bids, the high-priority campaign will attempt to trigger its ad for both search terms. For the generic search term, it is this high-priority campaign that ends up displaying an ad (ignoring external factors such as competitors). For the branded search term however, a conflict arises as the brand is negated, thus preventing the ad from triggering. This allows the low-priority campaign to come into play; here, no conflict arises, and the ad is triggered. With this method, it becomes possible to segment brand and non-brand traffic, and consequently to differentiate bids between these two types. Note that if product bids in the two campaigns are managed from a CPA or ROI-based perspective, the generic search terms will end up with a (relatively) lower bid than the branded search terms, as a result of the expected difference in conversion rates from these two types of searches.

Additional Cross-Campaign Segmentations

Besides the generic versus branded segmentation, which in the majority of cases will be the most relevant, it is also possible to segment traffic in other ways, for example:

 

 

Conditions for a Working Setup

One important condition in order for the above structure to work is shared (non-capped) budgets across the campaigns. It is very important to note that if your high priority campaign gets budget-capped, the low-priority campaign will have no restrictions in terms of triggering its ads for all search term types. Your search term data will then get mixed up and more importantly, the bid differentiations made between traffic types will discontinue.