If you’re like most retail marketers, you’re accustomed to crafting exhaustive keyword lists for paid search campaigns. But as this webinar demonstrates, it’s time to re-evaluate your keyword strategy.
Below is the transcript from our webinar, Zero-Keyword Search Campaigns: Reality or Hype? Led by Sidecar’s Mike Perekupka, this webinar examines how Google’s own updates, including changes to exact match, as well as innovations with natural language processing, are creating a world where marketers no longer need to sculpt expansive paid keyword lists.
Learn if you are over-segmenting your keywords by creating bloated lists, and how to change your keyword strategy for the future of paid search in retail. Mike presented this webinar on May 19, 2020 as part of the TechTalks series for RetailX.
This transcript has been edited lightly for brevity and clarity.
Keyword Lists Are Bloated
Today what I’m here to talk about is the concept of zero-keyword search campaigns and whether they are reality or hype, or maybe somewhere in between. The number of keywords required and the best practices in this channel have ebbed and flowed over time. There was a time where people had to become [keyword] experts. The more keywords, the better. There was this ramp up of how many keywords people were using in their campaigns.
Over time, what has happened is that people have come down this curve. There’s been new innovation in the channel that has caused people to re-evaluate that and see, do I really need to do all of this? That’s why we’re going to spend this time talking about today.
I want to relate it to the time period that we’re in right now, because right now what’s going on is we’re being forced as a society, as marketers, to re-evaluate everything, not just within this space. We’re being challenged to re-evaluate everything in our lives, from the simple things like buying groceries, to how you spend your money, your time, and how you interact with your friends and family.
Digital marketing and performance marketing are not immune to it. There are a lot of big-ticket items that marketers are re-evaluating, like budgets. There was a very moral crisis going on in the beginning of this, about sensitivity and should I even be advertising right now during this time? What products should I sell?
I think one of the most interesting things is competition. Whether we want to admit it or not, some businesses are going out of business. They’re forced to close. Some businesses are pivoting their models and getting into other verticals. Just keeping an eye on what your competition looks like now, versus what it might look like in the future, is really interesting.
Is This Bloat Needed Anymore?
What I want to do is pose a couple of questions throughout this webinar. The first one: When’s the last time you actually evaluated your tactical keyword strategy? Not the number of keywords. Not your budget. I’m talking about your tactical keyword strategy and how many keywords match types that you’re bidding on.
Paid search has changed, and you may be behind the times. You may be not adapting as well to the market if you haven’t taken the time to do it. And if you’re already reevaluating everything else, this could be a perfect time to do it.
Paid search text ads have been around for a long time. And a lot of things have changed — not just the way they look.
There was an interesting study that pointed out that 25% of consumers today cannot identify an ad versus an organic listing. Not only has the look changed, but the technologies that exist to manage the channels have changed. Google itself and the technology and the algorithms have changed.
How Retail Marketers Ascended the Keyword Bell Curve
There have been these advancements in paid search that have caused people to climb that bell curve to add more keywords and make their strategy more robust.
A couple of callouts would be mobile, especially after 2007. More mobile searches were happening. Google kept up with that and made you not need to create mobile campaigns. They would optimize your ads for you. That caused people to really expand keyword strategies.
New ad formats are always being tested. More imagery is coming in. Gallery ads came out and then they were rolled back. Now there are some picture extensions and that always cause people to add more keywords.
A major change happened a few years back where the real estate itself was affected. Google removed text ads from the right rail, and that caused people to add more keywords, but it also changed the competition. You’re now competing for less space. So people added more keywords and fine-tuned their strategies. That ended up with climbing up that bell curve and adding more keywords to their accounts.
Whether We’ve Reached the Inflection Point
I think a lot of people reached the top of that bell curve and kind of wiped their hands clean and said, “Okay, good. I’m here. Everything is done and now let me just let this run.” But more recently, there have been some advancements in technology that have caused people to come down that curve.
One is an advancement from Google itself, with Dynamic Search Ads. This ad format is a true zero-keyword approach and has gained popularity where Google will actually pick the keywords for you. It will write ad copy for you. It will even generate a landing page for you. A lot of people have tested that. I’ll talk about that as well.
But one major change that is more recent is match types. The functional way paid search [operates] as a channel, has changed. I want to dive into that because it had a very large impact and I’m not sure how many people spent the time to really understand this.
This happened to correlate very well with us developing an automated solution. The biggest thing that I want to point out here is that paid search exact match has fundamentally changed. It’s not exact anymore. It is matching to more queries. Google says it’s matching to anything that has the same intent, paraphrase words, and words that are implied are now matching to what are called close variants of your exact.
Google said they’re doing this to do you a favor, just for transparency. They say that every day 15% of searches are net new. So they’re saving you 15% of your time by not having to dive through keyword reports and having to add these to your campaign.
This was a massive change and we studied it a lot and I’ve learned a ton myself with our team and debated this impact. My next question for this audience is a simple question, but it’s loaded with content here. Have you spent enough time understanding how these changes have impacted your campaigns? Have you gone back and looked at how that change has impacted your performance and strategy?
How Exact Match Is Impacting Paid Search Performance for Retailers
What I want to do is walk through an example of why I think this is such a big deal. I’ll use a field or a vertical that we’re very familiar with. If you sell apparel, before these changes, the way you built keywords was to start with women’s shirts and build out every variation of keyword relating to women’s shirts.
You put an apostrophe in it. You made it singular. You made it plural. You changed the word orders. You thought of synonyms and misspellings and you ended up making a lot of variations of the exact same keyword.
Then you have to think about all the variations of that. And then you have to think about, “Alright, now I sell women’s pants and women’s jackets, and now men’s clothes and kids clothes.”
Is this necessary anymore? Do you have to do that? Are you wasting time? So what I wanted to do is run a controlled experiment. I wanted to show some data on this webinar today. What I did was set up a one-week test of a women’s apparel retailer.
I put the exact match keyword of women’s shirt and set it out into the wild, with new match types, with new exact match close varients. This is some of the data. This is a snippet of 50. On the right side of the screen you can see it matched to 91 total searches. Where in the past that would have matched to one.
Consider the very first one — shirts for women. That is not an exact match as we know it. You can see there are synonyms — ladies, female shirts. There’s t-shirts. There are synonyms of both. There are misspellings. There are plurals. It is matching to all these. Where in the past if you went and you built out all of these keywords, you would have had to build them to match to these. That was just one example of one keyword.
Why You May Be Limiting Your Paid Search Performance
By creating these exhaustive keyword lists, are you hurting your performance? Are you actually detracting from what you can do in paid search by doing this?
I want to elaborate on what I mean by that. Just continuing that example through, and then giving you some suggestions and things to think about. Before the change, it was very simple. Woman’s shirt matched to woman’s shirt. You got data around that keyword. You could make a smart decision about it.
If you wanted to match to other search queries, or you wanted to match to other keywords that were out there, you had to build them out and put them on exact match. So if you end up getting a woman’s shirt ad group, what that might look like is this extensive list of keywords where each one of these is built out.
But here’s the problem. Each one of these keywords then needs to be monitored. Each one of these keywords then needs to have a bid where you need to make an intelligent decision. You start debating with yourself. Should I bid it higher? Should I bid it lower? And then some of these keywords have $2 of traffic. Some of them have 52 cents of traffic. And how are you ever going to make a smart, intelligent decision about those keywords?
So I’m on this webinar claiming that by doing that and by staying with the old approach or the traditional approach of building out keywords, you are detracting from what you can do in this channel. You are over segmenting your keyword list by creating these bloated lists. You are actually hurting yourself.
If you don’t take advantage of this new functionality, you’re not doing yourself a service. This third concept here, I think can probably hit home for a lot of people. You’re also wasting time. You don’t need to do it.
Steps to Improve Your Retail Keyword Strategy
So I want to offer some suggestions. I want to talk about some quick tips for what you can do. If you think, “Oh wow, I am in that old world, the more traditional world. I am over-segmented.” I’m going to offer five tips for how to get yourself into this more modern scenario.
The goal should be to go from what I was talking about on the left, where you have this over-segmented list with all of the segmented data, and roll it up to this scenario on the right where you’re matching to more queries.
You’ve got a lot of data, a thousand clicks. You get $700 in spend. You can make a decision about that keyword every single day. You can use that data to do more meaningful things with your time. And you’re going to be more strategic. A/B testing is the name of the game in paid search. And if you don’t have enough data to do it, it’s another form of wasting time.
You can be strategic and recommend a holistic strategy across paid search and shopping, or maybe your other channels. That is the benefit of doing this. If you can get out of this traditional way of building keywords and move into this more modern approach, and adopt some technology to help you with this, there are a lot of areas of beneficial improvements that you can apply in your own campaigns.
Evaluate the Impact of Changing Match Types With an N-Gram Analysis
Go back and just take 45 minutes, take 30 minutes, take an hour, to evaluate the impact of the changes that have happened and see what they’re doing to your campaign. Are you over segmented?
One, download your search query report. Pick a timeframe. I usually recommend 90 days. On the right, create what I’m calling an n-gram analysis table. I’ll talk about what n-gram means. Set it up using the formula that I gave you in column L. Columns M through Q are where you’re looking for how those words perform.
What this table does is roll up performance. You can type words into column K and see how search queries perform with the word “best” or with a brand in it or ones that say “near me.” You might be able to get a look at data at that level.
Pick some words, pick some phrases that you are interested in, and see how many searches they are matching to. What does the performance look like? I guarantee you’ll find that to be interesting. Then what usually happens is you want to see how every word performs, every phrase performs. This is where my next tip is.
Leverage Natural Language Processing and Automation
Consider using natural language processing (NLP) and automation, whether that be your own resources or a partner. There are a lot of benefits to adding this layer of technology onto your paid search account. I think that a lot of people can stand to benefit from that.
NLP is studying how humans interact with the technology. How they interact with search bars. I want to call out three quick benefits here. The n-gram analysis that I just talked to was about how you can do it manually.
Imagine doing that at scale and looking at every word and every phrase across the entire search query reports and seeing how they perform. That is one of the major benefits that you can get out of this, and you can make such important data-driven decisions about that, that can really drive your keyword strategy, even port over to your Google Shopping strategy.
The second benefit is an automation benefit. Via automation, you can take so much of your day back and put it in your hands. You can have technology matching keywords and automatically negating and adding new ones based on some thresholds that you might be interested in. There’s no need to have a human being doing that every day. That can be done via automation.
And then the last bit is match types. There’s a concept within natural language processing called PMI, pointwise mutual information. That takes the guesswork out of match types. I think most people say, “Oh, I’m going to put this on broad match modified because it feels right. And I’m going to see how that performs and if it does well, I’ll move it to exact over time.” PMI, without getting too weedy, is a measure of how often or how frequently words come in together.
That can give you a mathematical calculation of whether you should put this on broad. You should put this on phrase or exact match. You can really take the guesswork out of that process.
Use Dynamic Search Ads to Support Your Channel Growth Strategy
Dynamic search ads can be a fit for anyone. But before I talk about my suggestion, I need to level set just for 30 seconds. The pro of DSA is that it is a time saver. When doing new keyword research, it is very hard to predict how every human will search. Google is pretty good at mining for new keywords. Like it said, 15% of searches are new every single day.
However, a major con here is that you must give up a fair share of control over your account to use these. You don’t get to pick keywords. And if you don’t have automation, that can make DSAs a little daunting.
My suggestion would be go forward, if you don’t have one, and create a catch-all campaign. Maybe if it’s a trademark catch-all campaign or an overall catch-all campaign, set to a very aggressive goal. You may get some incremental revenue and then get the benefit from learning from the matching that Google is doing on their end.
You might be able to take that and apply that to your strategy. DSAs is a tool that we have studied and there are some benefits. But you just have to be very careful about this true zero-keyword approach, because there is some control that you do have to alleviate
To go back to our initial question — is it reality or is it hype, the idea of zero keywords? I think it’s somewhere in the middle. I think that the reality is people have too many keywords, but going to zero is very, very scary for a lot of people.
Check Your Keyword-to-Ad-Group Ratio
Here’s an easy one. If all you have time to do is a five-minute check, just go to your Google Ads campaigns, and just check your keyword-to-ad group ratio. Count how many keywords you have, how many ad groups you have, or just download an Excel report and do a division.
If that number has gotten too high, if it’s gotten away from you, if it’s over 20, even at over 15, I would recommend it’s time to take some action. It’s going to hurt your relevancy. It’s going to hurt your quality score. So check that. You want to take some steps to go back and to reduce that.
Regularly Test With a Subset of Your Budget
My last strategic guidance that I want to offer is, when testing, pick a subset of your budget, maybe it’s 5%. Or 30%. Maybe it’s some percent of your budget. Maybe it’s a part of your catalog. If you are going to advertise those products on text ads today, how would you do it? Knowing match types have changed, knowing all of these technology advances have changed. You’re going to start fresh. I challenge you to think modern about them. Set a campaign live for a good amount of time. Monitor it and let it run.
You’ll find there are some benefits. There are some time savings you can have. There’s some technology and some automation that can really help take your campaigns out of the stagnant old world where they’re sitting there and getting bloated, and move them into this new agile, fresh, modern way of doing it.
Think about it being holistic with your other channels. Think about a more omnichannel approach. Can you think modern about this channel that’s been around for 20 years?