As a good manager, you’re constantly testing new ad variations, keywords, bid multipliers, etc. throughout the year and making optimizations based on your findings. Kudos to you! At the end of every year, I like to take a look back and see what observations I can make that weren’t entirely apparent in each test throughout the year and find areas to focus on in the coming year. This week’s post, as made apparent by the title, is Keyword Match Types. Here’s my process.
My general goal looks a bit like this: Exact < Phrase < BMM < Broad. In almost every stat column, I’m looking for Exact to have the highest cost, conversion total, conversion rate, average position, etc., followed next by Phrase, then BMM, then Broad. There are a couple reasons for this. First, I want to know exactly where the majority of my spend and conversions are coming from. It’s easier to find where up or down trends are initiating when your keywords are exact. Second, it’s easier to foresee bid change impacts when your keyword is only matching to one search query. Raising a bid on an exact match term is more often than not very predictable. On a broad match term however, raising a bid could do just about anything. It might respond the same as an exact match term, or you could have just made yourself eligible to match to a more competitive query, effectively pulling average position down and raising cost, CPCs, and CPAs. Too many variables there. No thank you.
Getting the Data
I put a label on each of my keywords for exact, phrase, broad, and broad match modifier. I use labels because the Match Type column doesn’t treat broad match modifier as a match type separate from broad, and I consider them to be very different. Once your keywords are labeled, hop on over to the Dimensions tab and select Labels – Keyword as your segment. If you need to, you can filter for only the four match type labels to clean up your view. Then change the date range to all of 2013. Here’s the breakdown for an account we inherited at the beginning of last month.
All CPAs are in an acceptable range, but BMM keywords are getting a little high. Might need to take a look at bids and search queries to cut some fat. It could be that one of the words I have as a root keyword is actually hurting performance.
Average position is exactly how I like it to look. Exact has the highest at 2.7, with Phrase next at 3.1, then BMM at 3.2, and regular Broad coming in last with 3.4. Assuming that your keyword match types are tied closely with relevancy, your breakdown should look something like this one.
Cost and Conversion Sharing
Here’s where I start to have problems. Broad match accounted for 51% of overall spend and 53% of conversions while Exact brought in only 8% and 7% respectively. This isn’t even close to the Exact > Phrase > BMM > Broad principle above. Not to mention that Broad actually has the lowest CPA for the year. My takeaway here is that we’re missing tons of converting keyword variations within the account. Search query analysis should produce lots of new ad groups and keywords for us to target using all match types and hopefully shift the cost and conversions focus over to those more targeted types.
CTR and Conversion Rate
Phrase and BMM have the highest conversion rates, while the Exact conversion rate is a bit low for my taste. CTR follows about what you’d expect, highest for Exact and lower for those in lower positions. My thoughts here are that we might have a few too many Exact match head terms in high positions, or at least not enough long-tail Exact match to make a difference. Somewhere between ad copy and landing page we’re losing interest. Our existing Exact terms might need a slight bid adjustment and some qualifying copy.
In a couple months, I’ll run a similar report looking to see if I’ve made any progress on my goals. Hopefully at that point, we’ll have a much lower dependence on Broad match keywords and will have Exact as the main driver of the account.
Next time we’ll cover how your past ad copy tests could help shape your future tests. Wahoo!