There’s a common narrative in the search industry that Google AdWords match types adhere to a hierarchy of expected performance, with exact match driving the best performance due to higher relevancy, phrase match coming in second, with broad match sitting in the corner, eating glue with a bucket on its head.
Following the narrative further, a good marketer should be making use of all the match types available but should be controlling impressions using bids and negatives so that conversions tend to come from exact match keywords. This should mean that your cheapest traffic source is also your best converting.
It may come as no surprise that this is seldom the case. After all, different markets are sure to behave differently and there are many factors to be factored in. We’ll get to that.
Here’s a set of keyword data for 12 of our clients over the last 6 months, broken out by match type:
Keep in mind that this data set encompasses all keyword sets being marketing in search: brand, generics, competitors, your mum, etc. Note also that we stack our match types, so that converting search queries are built out into all three match types with negatives in place so that if a query can fire an impression, the highest-tier match type will fire.
What’s interesting is the narrative is only confirmed in Quality Score with exact taking top spot, phrase match coming in second and broad match bringing up the rear. However, phrase match has the lowest CPC but also the lowest CTR. Broad match shows the highest CPC but also the highest conversion rate. Wait, what!?
Brand can skew results by quite a bit, so let’s back that out to level the field a bit:
No surprises here; CTR decreases across the board, CPC goes up, quality score dips considerably.
Again, quality score maps as to how you’d expect, with exact match winning in terms of relevancy, phrase coming in silver and broad taking the bronze. However, conversion rate is vastly different with these filters in place; broad match again has the highest CPC but converts at nearly 1 in 3. What sorcery is this?
Well, it turns out that discovery keywords (that is, everything but exact) are way more useful and important than the narrative gives them credit for.
Often exact match keywords turn out to be Low Search Volume, which is Google’s way of saying, “we can’t run a profitable auction out of this one advertiser bidding on this one very specific search term, so we just won’t run one”. When this happens it forces marketers to bid on less complex keywords & generics to acquire traffic for their clients. This places more pressure on the phrase and broad match regions of their account to convert because the search volume for generics is generally higher than exact.
It follows, then, broad match could drive a higher conversion rate than exact by tapping into streams of search traffic that you either don’t have built out or are unable to bid into.
It’s also likely that phrase match delivers lower CPC than exact due to competition. Everyone knows what their killer keywords are and will optimise the absolute hell out of them, pushing traffic into exact match (because that’s what the narrative says you should do), but this leaves space in phrase match to pick up cheaper click traffic that the major players aren’t actively hunting after. This is further illustrated by the fact that phrase match exhibits the lowest conversion rate of the lot, even including brand.
The key takeaway here is that you can only trust the data you have, and pay little mind to ‘conventional wisdom’. The fairy-tale of exact match may be true as far as quality score is concerned, but there’s so much more going on that you can’t stop here.
So the canny marketer is invited to ask themselves what they want to get out of their traffic?
If I want to drive cost-effective conversions, then I probably want to push spend into exact match for high-relevancy traffic with strong CTR and respectable conversion rate, accepting that I may have to bid hard to get into this market. I also probably want to build out my broad match modified keywords to hoover up the converting traffic I haven’t built into my account and use SQRs and negative keywords to optimise efficiency moving forward.
If I want to play for airtime dominance, maybe I should pull back on exacts and push spend into phrase match for low CPC traffic, accepting that conversion rate will decrease as a result but traffic may ultimately go up.
If I want to be a jerk and bid on competitor terms I should go home and rethink my life might want to consider how my product compares with the competitor I’m going to war on as it could have huge implications for bidding strategy.
Here’s one advertiser’s competitor campaign that’s outperforming their own brand in search:
This one’s a pretty mature account, hence the majority of click traffic going into exact match due to tightly controlled negatives and 2 years of search query expansions. Still, broad match manages to drive a 50% conversion rate at a lower CPC than exact match. Note also that phrase match CPC, CTR and conversion rate are by far the lowest of all types.
In this case, if the goal were to drive cost effective conversions we would be bidding differently, however, this campaign is designed to go to war on a specific set of competitors in the same market as our client such that conversion rate & volume is not the ultimate measure of success.
Maybe it’s time to rewrite the narrative. Once the champion of PPC conversions, exact match probably needs to share its accolades with much-maligned broad match, while gentle, overlooked phrase match with it’s cheap traffic and a mouthful of glue quietly waits for the day it can show them.
Show them all.
Phil O’Connor (Searchinae Peculiaris).
Mammal: Digital Performance Manager for indago digital in Sydney, Australia; has been at the forefront of technological & tactical innovation of the search industry since 2007.
Natural habitat: Reddit. Can be found occasionally in whisky bars. Hibernates in winter.
Diet: Steady intake of marketing industry content. Crippling dependency on ham & cheese croissants.
Latest posts by Phil O'Connor (see all)