In this episode of the Show, I had the opportunity to interview Navah Hopkins, Services Innovation Strategist at WordStream, a leading PPC management software provider, on how to approach and optimize keywords in paid search today.
How do you approach keyword theory when it comes to paid search?
Navah Hopkins (NH): Keyword theory is the staple of how I look at all of paid search which is: What are the keyword concepts that are going to give you the best access to your desired market at the most cost-effective and scale-oriented price?
Now part of that is the actual keyword concepts you go after. So, for example, going after attorney versus lawyer.
But the other part of it is also what match types you select. And as SEJ has done a great job of reporting, there have been a number of changes in how the match types function.
And so I consider that in keyword theory, not just looking at the keyword concepts but at what match type – broad, modified broad, phrase, exact – you have those keyword concepts on.
Because how you match those keywords (when you tell Google to stop trying to match your keyword to a query) can have a really big impact on whether the queries that you’re showing up for are relevant to you.
- Are you exposed to auction prices that are not quite profitable?
- Are you being put into auctions because you’re on a really restrictive match type that is cost-prohibitive to you and your business?
So there’s a number of things that go into keyword theory. I would say that if you had to distill it down to just one general idea, though, it’s: “How are your people searching and how can you get access to them?”
Do you think third-party tools do a better job in putting together keyword lists?
NH: I like using every tool possible. Because even garbage can be used for scrap into beautiful things.
And this is not to say that the [Google] Keyword Planner is garbage, but what it is to say is that the keywords that are going to be suggested are all going to be on what is called “broad match.”
So I sometimes love the ideas that come out of the Keyword Planner, they can be exactly right. But because they’re added on broad match, you don’t actually guarantee that you’re getting that specific idea.
You’re getting a synonym. You’re getting an implied word. Only one word has to be there.
My favorite way of doing keyword research, ironically enough, is actually doing dynamic search ads (DSAs) where I am targeting bits of the site…
And when you run that DSA campaign, you’re actually given back a list of actual things people searched in relation to your site or service.
I also love doing simple two-word or three-word long strings of keyword concepts that I know will teach me how people are searching, rather than bidding on everything at once.
So it’s not that those tools aren’t useful. It’s just that it’s very tempting to have a whole bunch of keywords dumped in prematurely and then have a real mess of a time and a very time-intensive period of cleaning it up.
They can be good. You just have to be really mindful that they’re being suggested on broad match, so you’d have to actually change the match type.
What are close variants and why do they matter?
NH: Close variants are variations on keywords that Google Ads or Bing Ads will use to make a logical leap. Some examples include:
- Singular/plural
- Misspellings
- “ing”, “ed”, “er”, etc.
- Abbreviations
- One word being split into two words
- Two words being combined into one
Say, for example, you’re looking at teeth whitening, you could come up for tooth whitener for the same keyword.
Now, the strategy that you choose to leverage for close variants is having a negative for every single idea.
Active keywords have close variants, negatives don’t. So if you’re going to try to keep your variants and actually bid accordingly, you need to have every other variant as a negative in the ad group or in the campaign.
Why that can be a problem, there’s actually a limit of how many negatives you can have in a given campaign. It’s currently 10,000.
The other strategy is that rather than bidding on every single idea, you pick the variant that is the most cost-effective for you and is the most predisposed to drive transactional traffic.
Now, what do I mean by transactional? I define search result pages or SRPs as either being research-oriented or transactional.
- A research-oriented SERP will favor mostly organic.
- A transactional-oriented SERP will favor mostly paid.
And so if you see that a particular way of searching by and large, and it can even be a misspelling of the keyword, is how you’re getting the majority of your traffic, bid on that and pause the others.
Odds are you are still going to get those other keyword concepts. You’re still going to be eligible to SRP and you can check for yourself in the search terms report.
If you pull in the column of “matched by” and the target keyword, you can see what keywords resulted in what close variants…
It’s worth noting that by the new rules of close variants, which is mildly scary for some of us legacy PPCers, “attorney” and “lawyer” are now the same thing on every match type.
So the panic that comes from close variants and the doubt around how to act, is that keyword theory – for a really long time – had the same set of rules of engagement.
The rules on match types and how they functioned didn’t change. So it was very comfortable, very easy to work in a specific way.
Close variants and the implied words that now come with the different match types are shaking it up a bit. And so the structures that used to work really nicely maybe no longer are as effective.
Do you feel like it’s shaking up in a way that you can work around still?
NH: It’s absolutely still doable. You just have to be okay taking a step back and recognizing that the structure you used 3-5 years ago no longer works.
My favorite case of this is SKAGs or single keyword ad groups. Singe keyword ad groups used to be the quintessential way to not only get a really good quality score but also to, essentially, handpick what auction price you’re going to invest in for a given bit of creative.
The problem is, SKAGs require you in today’s world to have so many negatives that the odds of you perfectly protecting your SKAGs structure so that there isn’t keyword duplication, there isn’t budget misallocation, and budget shortages, gets compromised.
There’s this new idea of STAGs or single theme ad groups. And these are really a way to bypass the close variant issues where, for example, say you were to have an ad group of “lawyers near me,” “best lawyers near me,” “lawyers taking clients,” and you have each of those keywords as an ad group.
In today’s world, those would be one ad group. And you would have creative talking about the lawyer near you…
What’s your take on keyword valuation and determining whether your keywords are working for you or not?
NH: A couple of things. First and foremost, we have to ask ourselves: “Is our conversion tracking working?”
If our conversion tracking is working, conversions are almost always going to be the gold standard on whether or not a keyword is working.
The reason I say “almost always,” sometimes conversions are tricky and there’s a multi-step funnel that you can’t perfectly protect against.
My favorite example of this is you have keywords converting well, but then your sales team can’t actually close the deal on those leads.
You’re converting – marketing is doing its job, but there’s a disconnect in sales.
So put that aside. If you do not have conversions, the supplementary metrics I like to look to, number one actually, of all things, is impression share.
The reason why I like using impression share to assess if a keyword makes sense or not is that I’m a pragmatist and I’m also lazy.
I don’t want to have to do more work than I need to, and I want to make sure that what I go after, I am pre-disposed to win.
You have a two-week learning period baked into any new thing that you do when it comes to paid search. Where that cycle gets longer or shorter depends on how much you are willing to invest.
I’m a big believer in investing aggressively in the beginning, but you then have enough data to make a decision that you could optimize quickly and efficiently later.
I might budget and bid maybe 10-20% more, in the first month or two of a campaign. And then roll that back as I’m working, and the reason for that is two-fold.
The first reason is Google and Microsoft love data. They genuinely need it in order to operate and optimize and allocate budget and serve your ads and work with your keywords in an efficient way. And if you’re depriving the system of data, you’re making it actually harder for it to know that you are a reasonable advertiser and get those placements.
The second reason is if I invest early, I am going to get a plethora of queries, auction prices and potential prospects that I can then audit and make decisions on whether I’m focusing in the right way.
It’s almost always better to have a larger budget at the onset of a relationship then get the buy-in that you’re going to pull back, than to start off super small where you’ll have no data because then you can’t assess anything.
What kind of meaningful decision will you make off of a hundred impressions and two clicks? You’re not.
So it’s better to invest aggressively in the first month or two, but then know that you’re going to pull back and audit those queries.
More Resources:
To listen to this Search Engine Show Podcast with Navah Hopkins:
Visit our podcast archive to listen to other Show podcasts!
Image Credits
Featured Image: Paulo Bobita
