I’m not as strict on duplicate keywords as many of my fellow PPCers. While keyword overlaps don’t add any value in themselves, they are not nearly as problematic as they are often treated, for reasons explained in my post on exactly that topic.
That said, overlap does reduce clarity into the job being done by each keyword… and clarity is important for optimisation. So when an account grows, and its campaign/ad group structure begins to sprawl, you will want to check where your keywords are competing for the same clicks, and tidy up the ‘division of labour’ between your keywords where you can, so that you can treat each keyword as a properly discrete entity.
Here’s how to find out where ‘overlapping search territory’ is cropping up in your account.
N.B. This is distinct from finding duplicate keywords, which can be done at the touch of a button in Google Ads Editor…
Here we’re delving a couple of levels deeper and looking for search terms that sit within territory covered by more than one of your keywords, in order to find the zone of overlap:
With the whole account selected (or at least any campaigns you’re interested in…) open the keywords view and click search terms.
Take a decent date range – say at least three months, but more may be needed depending on data volume.
Add ‘keyword’ as a column in the table.
(a useful move in itself, as it’s always worth knowing which keyword produced which search term…). Make sure your table is showing ‘conversions’ too.
Export the table straight to Google Sheets (a nice new feature… https://support.google.com/google-ads/answer/9055275)
In your newly created Google sheet, lop off the first two summary rows that appear above the data, and – under the ‘Data’ dropdown – create a pivot table. No – wait – don’t go! – I promise it’s not that bad…
To set up the pivot table, under ‘rows’, first add search term, and then keyword.
Under ‘values’, add the metrics you will want to compare between the different instances (keywords) of each search term.
Impressions, Clicks, Cost, conversions ,and revenue if you have it, will all be interesting. The ‘functions’ of these metrics (CPC, CTR, conversion rate, cost per conv., ROI) are also valuable for this analysis, but don’t try to pull them in directly into the pivot table or they’ll get skewed by the aggregation.
Once the metrics are in, you can go back up to ‘search terms’ under ‘rows’, and sort by number of impressions, so that you see the significant overlaps first…
In most cases, you’ll see single search-term-keyword pairings as above. No overlap. Where you see an expanded row, as below,with one search term matched with two or more keywords, you’ve found what you’re looking for.
Sometimes, match type or context alone will dictate that one keyword is more appropriate than the other/s to be the rightful owner of the traffic for that search term. If so, use negative keywords to exclude the search term from the other keywords’ ad group or campaign.
If there’s no obvious preference, you can now scan along the metric columns to see whether the results guide you towards a winning keyword. Does traffic from the search term perform better in terms of CTR/ conversion rate / CPA with one keyword rather than another? Again if so, exclude the search term from the keywords that are not making such good use of it.
Search-term crossover is a normal (and usually tolerable) part of search campaigns – but it’s good to run this sense check and tidy up every now and again, just to make sure your activity doesn’t deviate too far from where it should be – and your stats don’t lose their clarity.
It’s like a haircut for your campaigns – and worth doing about as often.