Search Query Reports are incredibly useful for the SEM marketer – as seen in my previous post on Search Query Report Mining

One other use of the SQR is to figure out what terms are driving up your cost per conversions.

If a particular theme seems to have lower conversion rate and hence higher acquisition cost, we could consider excluding the term.

One easy way to detect these themes is to use the Negative SQR Tool I’ve built here.


The principle is simple:

  • Bubble size is the cost
  • Y axis and color is how expensive it is to get a conversion (Cost Per Conversion)
  • X axis denotes conversion rate

Simply put – themes that are expensive will be deeper red bubbles that float to the top and those that are due to low conversion rates will drift left.
the larger the bubble, the more important it is that we “pop” it by adding it as a negative keyword and excluding it from the campaign.

This dataset is for a cafe buying SEM into location keywords. We can see some very intuitive patterns appear:

  1. MRT (Train station) is a bad buy – even through the cafe sits 5 mins away from Taiseng MRT, users searching for MRT are most likely are “grab and go” hence a likely candidate for negative keyword
  2. Location, location location.
  3. The further away the keyword location bought, the less effective it is for converting. Taiseng and Payalebar are slightly more distant location compared to Taiseng. As a direct marketing term it doesn’t work as well – these could be shifted to a branding account and bidded on differently.

  4. Bidding plays While taiseng is more expensive, it is not entirely worthless due to the high conversion rate. By lower bids on these terms we might find it eventually profitable to run

With a simple chart we could have real actionable insights that we could optimize against.