Broad Vs. Exact Match Types: The Hard Data
Search managers are often asked the question if they are over- or under-exposed on broad and exact matches. I have heard several heated discussions debating this issue. Since this is an issue that touches many a nerve, I decided to take a dispassionate data driven look at the numbers.
I selected keywords in my company’s (Efficient Frontier) databases that had both exact and broad analogs. Phrase match was grouped under broad. These keywords represent about $36 million in monthly spend. So in short, it is a large keyword set. Next, I calculated the total spend, clicks, cost per clicks (CPCs) and impression volume per advertiser and per vertical. Finally, I calculated the broad to exact ratio. For example:
Impression ratio = broad match impressions / exact match impressions
CPC ratio = broad match CPC/ exact match CPC
…and so on.
The ratios have been plotted as a boxplot above. Note that the leftmost line in the box plot represents the 25th percentile, the middle line the 50th percentile (median) and the right most line represents the 75th percentile. Outliers have been plotted as black dots.
How to interpret the ratios
A ratio that is less than one means broad matches are less dominant than exact matches for that metric. For instance, if the CPC ratio is less than one then it means that broad match CPCs are lower than exact match CPCs. On the flip side, a ratio greater than one means that the broad match dominates the exact match.
Just remember that all these keywords are in campaigns optimized for performance so this is, in a way, a biased dataset.
Median CPC ratios were always greater than one. On average, for the same keyword the broad match is more expensive than the exact match. I found the biggest skews in automotive and travel where they were 2.8 times and 1.5 times more expensive, respectively.
Median impression ratios were between 2 and 3 for all verticals. In other words, the broad match analogs get between 2 and 3 times as many eyeballs as the exact match.
The click ratio was slightly less than one. Therefore, broad and exact matches drove about the same number of clicks. Taken in context with the previous statement, the effective CTR of broad matches is about a third of exact matches.
Advertisers spent between 1.2 and 2 times on broad match as exact match (between 55% and 67%). In other words, broad match dominated the spend.
I found that smaller advertisers spent more on average on exact matches than broad match. This makes intuitive sense; advertisers with a small budget would sacrifice volume over quality more than a larger advertiser.
The findings can be interpreted in many ways. I have presented the data as it is for you to draw your own conclusions. However, here is my take on the findings:
The findings corroborate what most of us already know. Adding broad matches will get you more volume but will cost you more per click. This appears inevitable. The quality score on the broad match keyword should be lower than an exact match on and hence CPCs should be higher. This, in my mind, is an open area for additional research—any takers?
The CTR conclusion makes intuitive sense too. At least after this analysis, you should have a baseline of what CTRs you can expect from your broad matches as against exact matches.
The ratios do vary widely by vertical. So keep this in mind. In terms of CPCs, retail, finance and classifieds have the most equitable ratios while automotive, travel and business services have bigger skews.
Smaller advertisers depend more on exact matches than broad matches. Again, this shouldn’t be surprising. However, one must keep in mind that as a campaign expands, broad matches will have to be added and there will be a loss of efficiency, just like with any growing business. Bottom line: You can run a low CPC, high CTR campaign on exacts but your ads will only reach a small audience.
If you are running a campaign or are looking to expand an existing one, keep these factors in mind. This will help set expectations both for you and your organization and avoid unnecessary heartburn.
I have not talked about conversion rates in this post, though keep in mind they will affect this discussion too. That will be a topic for a future discussion.
Some opinions expressed in this article may be those of a guest author and not necessarily Search Engine Land. Staff authors are listed here.
(Some images used under license from Shutterstock.com.)
Analytics news and expert advice every Thursday.