I am guilty of sometimes being an over protective parent with my campaigns. I can lock down and restrict my keywords to the point where I may be missing out on clicks. I may be missing good clicks — even converting clicks. I am maximizing click through rate, like a good PPC manager wants to, but am I sacrificing missed qualified visitors? How much impressions waste am I willing to put up with to get qualified clicks? What is the intercept between the two? I call this the Keyword Exposure Threshold, or KET.
What is KET?
KET is the fine line of balance between getting qualified clicks and conversions, your risk tolerance and flexibility with your budget, and the opportunity cost of missing out on other converting clicks because you are too conservative.
Typical Audit Tactics
A normal campaign audit finds me looking at the keywords that were being used and evaluating the actual search query reports. This allows me to see what queries are actually generating impressions and clicks. I often come across poorly matched impressions that result in unqualified clicks and that costs the client valuable budget. For example, a basement waterproofing client that had “basement waterproofing” as a broad match receives 200 impressions and 50 clicks for the actual search query “basement remodeling contractor.”
In my typical playbook I pause all of the broad match instances, and extracting the queries that are working, and then implementing those queries into an ad group as new phrase, or exact matches. I then rewrite test ad copy to aligned keyword saturation with those matches. Most of the time, this audit results in improved click through rates, cleaner ad testing, and an increase in conversion rates. However, it does not work all of the time.
At what point does keyword underexposure creep in?
The reason behind determining KET is that sometimes potential customers for my clients do not search in nice, perfect ways; I find this particularly true with my B2B clients. Often times my client has a product or a service that is not in the keyword sweet spot. For example, the keyword may be “steel plate fastener” but needs to be found through a clarifying question or a related process query, such as “steel joining components.” So when I lock down the exposure on my keywords, I am also being too overprotective and missing possible converting clicks. Impression shares on your matched keyword will only explain so much. What of the market and the overall pie are we missing? This is a challenge.
Many would argue that using modified broad match types would help solve this. I would agree. I think it is a good idea to experiment with modified broad match in a test run to try to generate more exposure for a keyword without totally exposing it to poor impressions. My experience with this is that modified broad match is only a sliver better of protecting against keyword exposure. In going through search query reports of broad match modified keywords I have seen some really bad correlations Google has made between query intent and query, thus triggering my ad and a misguided click. While negative keywords help, let’s face it. By then it is too late, and there is no such thing as future negative keywording.
Feeling out KET
One way to combat keyword underexposure is using the keyword planning tool. Use this tool to look up monthly search volumes for differently matched keywords combinations. I find if you do this and combine it with your current impression share numbers you can make some assumption on how much of the market you are exposing your keyword to. Another way is working with your client and getting “boots on the ground” feedback from their sales teams on how clients are finding them or talking about their products and services.
A possible scenario I have experienced may work like this: I have my client’s budget. I gut the broad matched campaign and start with phrase matching, utilizing keyword saturated test ads as I referenced above. While volume drops off a cliff, I seem to get good click through and conversion rates. However, the client is clamoring that traffic is way down and leads are down. Note: I didn’t say cost. Often times in these scenarios my costs numbers hold steady because I am a bit more aggressive with my phrase matching because I want to be sure we get good exposure and I am confident in my selections.
So what is the next step? I may do more keyword research and add in more ad groups with phrase or exact matches. Or I may test out with modified broad match keywords, using any negative keywords I can generate from past broad match search query reports and client feedback. I may also use broad match keywords, but use multiple term queries to help further qualify when I want an impression to trigger. But when I do this, it is critical that I keep a very close eye on the actual search queries for poor ad triggers. Additionally, I keep a close eye on that budget, and pair it with how the cost of conversion is playing out. At some point into the campaign, I start to establish a percentage of the budget that is more risky, and I know I have to endure bad clicks to get converting clicks. If I then hit that ceiling, I have to take more action. (client meeting, more negative keywording, etc)
Over time I have learned that discovering the right KET is like anything in PPC. I have found that it follows this path:
1. Trial and error.
2. Knowledge of your client, and their goals.
3. And your client’s budget.
4. To truly kind the right KET you have to chase it over time, and the target never really stops moving. It often does not play out neatly or nicely.
5. It requires using a combination of multi-variant broad, modified broad, phrase, exact and negative keywords.
6. It also requires good ad copy testing. If you don’t write and test for effective ads you will have a harder time determining what keyword match types are working and how what your risk tolerance is.
7. Did I mention the part where it never really ends and you chase it constantly?
8. Managing and maximizing cost per conversion per match type. (Proper ad group structure is critical to this)
9. That it is a moving target
So that is how I determine and stress about Keyword Exposure Threshold. How do you handle it?