This week’s PPCChat session was hosted by Julie F Bacchini who sought experts’ views on types of keyword matching, The biggest challenge with the current state of keyword matching, Future of keyword in search and more.
Q1: What type(s) of keyword matching are you currently using in your Google Ads – broad, broad match modified (BMM), phrase or exact match?
All of the above, except very little broad match. @Mel66
Never broad match. I hate those so much. We will use exact, ModBroad and sometimes phrase if the modbroad has issues in the order of words. @360vardi
All of the above. Each has its role to play. @stevegibsonppc
Mainly using BMM and Exact Match. I’ll throw in some Phrase match here and there, but only when word order matters. @_GilHong
BMM, phrase and exact. Basically never use straight broad. @CJSlattery
My too-typical setup is duping all keywords BMM/Phrase/Exact unless I can see a lot of intent mixing with different word order. Broad is good in small doses if after hours of research I’m still flying blind on “good” terms for the client. @ferkungamaboobo
Haven’t had to use broad in awhile. Mostly exact, phrase and MBM for me. @robert_brady
BMM, exact and phrase. And I hate every minute of doing search queries. @nikkijumpfly
Everything but straight broad. @IamNextSTEPH
Like others have said, a mix of BMM, exact and phrase. Very little straight broad match. @alexpeerenboom
Still using a mix of modified broad and exact, but no longer segmenting by ad group (in new/restructured campaigns) because match types aren’t what they used to be. @akaEmmaLouise
All but straight up broad. Although tbh I don’t know why I bother with exact anymore. @amaliaefowler
Primarily BMM for data mining and to find converting keywords. Then exact match to separate out campaigns and to push spend for the ‘key’ terms and converting keywords. However swaying more so towards phrase match for data mining lately, esp for high cpc industries. @AmoreDigital
We mainly use Exact & BMM, with Phrase almost exclusively being used for negative keywords. @nataliebarreda
I use all match types except Broad for the most part. I have been testing more broad keywords but as a separate campaign to control budget and not let it run away with my other campaigns money. @KyleShurtz
Q2: Has your strategy for keywords changed over the past year with the changes to how keyword match (or don’t match, depending on your world view!)?
Yes, a lot more time is being scheduled for negative keyword work each week. Google opened the flood gates for random terms on accounts. @lchasse
This chat topic was inspired by a Twitter conversation after I tweeted about how pulling converting exact match terms into their own campaign no longer really works like it used to… Lots more SQR time and negative application. @NeptuneMoon
A lot more work involved in SQRs for Exact match, which never used to need to be done. There is very little difference between the match types now. @Mel66
I used to use broad in spurts – for a day or 2 – to capture long tail and learn. But now phrase match tends to do the same thing. @tunadonut
I’ve been trying to pay more attention to how Google is matching to intent (however loosely) and subsequently adjusting how I decide when a potential ad group theme is worth segmenting or not. @_GilHong
Looking for a lot more cross negatives in ad groups to make sure the query matches the customized ad language. I do not want someone looking for Vegan to see an ad that says vegetarian. @360vardi
I also like to make sure I don’t have seemingly duplicate KWs in exact. If Google now re-arranges the order what is the point in having all these instances in exact. @360vardi
This has been one of my main experiences as well. Not just from the SQR, but proactive research into what queries could be triggered as well. @alexpeerenboom
We don’t build out all the plural/word order variants anymore. Still segment with themed ad groups. Ie Flower Vase versus Florist Vase ad groups, with opposite as negative, since Goog/Micro don’t pay attention to keywords present in the campaign. @nikkijumpfly
I really haven’t had to shift at all on how I build campaigns due to this. I’m obviously in the small minority so I’m so confused as to why… The biggest change over the years has been dropping geo modifiers from targeted keywords in initial builds. @ferkungamaboobo
Definitely more time spend on search term reports to add negatives. Also not really building out long-tail ad groups as much. Just creates more and more potential places for a query to end up. Fragments data all over. @robert_brady
Yes! We said bye to SKAGS. @amaliaefowler
Google’s changes haven’t made a huge impact on me TBH. I’m vigilant with my SQRs anyway, plus I’m lucky I don’t have any clients with brand names that create issues. I try to have a solid, up-to-date understanding of match type intricacies to field client? @FindingAmanda
Seeing a lot more cross-pollination between ad groups that never did before. Having to add a lot more negatives where they weren’t necessary. Expecting this to get worse. @jmloquist
Yes! Uses to hardly ever use phrase match. However have been using it more and more as broad tends to mean you spend half your life adding negatives. Phrase is great for high cpc industries or clients that are running a watertight budget. @AmoreDigital
Absolutely! Exact match has worked very well for me. I used to rely on having more BMM and since the updates to Exact match, BMM became more expensive. @loogleppc
Strategy has for sure changed. I honestly think Google is heading towards no match types in the near (5 year) future. I have started doing more longtail and also playing with match type campaigns. @KyleShurtz
My strategy for KWs hasn’t exactly changed, but how I think about the use of a KW has changed. I’ve tried to embrace the change & understand how to work within the changing ecosystem, so I try to focus more on the intent captured & less on building out every KW @nataliebarreda
Q3: What is your biggest challenge with the current state of keyword matching in Google Ads?
Definitely more time spend on search term reports to add negatives. Also not really building out long-tail ad groups as much. Just creates more and more potential places for a query to end up. Fragments data all over. @robert_brady
Google SUCKS at understanding nuance. It assumes two separate four-word phrases mean the same thing when, to the user, the one word that is different changes the meaning entirely. Diluting match types has made this much worse. @tunadonut
It’s creating havoc for some of our B2B campaigns. We carefully sculpt keywords to reduce B2C traffic but then get matched to those queries anyway via close variants. It’s a real problem. @Mel66
Chasing negatives, instead of spending time optimizing, ad writing, etc. I’d be one thing if the results were there and I could let it ride. But with results dropping for clients (ROAS or CPA), that’s not possible. @nikkijumpfly
Finding all potential Close Variants to exclude before having to pay for them is a pain. I like to search product names in other sites like Amazon and Pinterest to see if there are same-named products out there. @_GilHong
Probably Google changing it’s mind on what a “close variant is” @jord_stark
I think the biggest challenge now is unreliability of keyword-level data. Much harder to say “this works and this doesn’t” because it could be luck of the draw which kw was matched to which user/audience. Harder to confidently exclude based on performance alone. @akaEmmaLouise
It used to be predictable, now it’s unpredictable. @amaliaefowler
It starts with an expectation setting, as most things do. You almost have to allow for and prepare clients for a certain percentage of crappy/irrelevant terms every month even with SQR vigilance. @NeptuneMoon
I think for me, it’s doing the matching of search term to customer journey stage – that’s got less to do with anything technical and more that I’m trying to make connections to the real world now. @ferkungamaboobo
Simply put, keeping the garbage out. Seems like so many more junk queries slip through even with the most careful search query reviews. @timothyjjensen
I don’t overthink it anymore. What I use is working very well. I do, however, continue to test other match types periodically. I have a killer, extensive Neg KW list and know that has helped a great deal. @loogleppc
It’s all broad match now. Just 4 different versions of broad match. I’ve accepted this is the way it’s going to be. @stevegibsonppc
The cross-pollination is annoying, but not a deal-breaker for me. I’m more concerned about what this means for the future of the KW. All these changes are v obviously leading up to a huge KW match type change that I’m afraid will be a true black box, like UAC. @nataliebarreda
Q4: Have you found any successful ways to mitigate these challenges?
Managing multiple negative keyword lists at Campaign and Ad Group levels. @tunadonut
A lot of communication, SQR report mining, and conversations with clients. @amaliaefowler
I make sure reviewing their quality, reviewing search terms, and researching trends is something I do regularly (on my calendar!). @loogleppc
Mainly re-prioritizing what work on the account is going to have the biggest changes… As much as my OCD hates junk query matching…. @_GilHong
More regular SQ checking. More proactive negatives on the build. Complaining to all of you. @nikkijumpfly
Intent is king (sorry content). If a keyword might have a different intent then what my advertiser wants, I put it in its own campaign to monitor costs./performance. @FindingAmanda
Putting on my “marketing hat” and trying to get more creative with copy that addresses an *audience* versus a specific keyword. It’s terrible and tiring but I think it’s where we’re heading as an industry tbh. @akaEmmaLouise
Tons of SQRs and audience exclusions when clear patterns have been found – Using tons of Google Analytics insights for this. @andreacruz92
That’s not to say that keyword management and optimization isn’t still important, I’m just leaning more into the messaging side as we gradually lose control and transparency on the keyword/bidding side.@akaEmmaLouise
I monitor the search terms report and make sure to add negative keywords. I use DSA’s for keyword expansion and once I have found keywords that work well I remove them from the DSA and add them to a standard ad group. @joshuadubs
I don’t have what I’d call a great solution for the squishy keyword matching. I’m mostly just muscling it right now, honestly. Checking SQRs every couple of days and adding negatives. Trying to figure out ways to be more predictive, but it is hard.@NeptuneMoon
I think the biggest win I have seen is longtail terms that convert as exact. I am also pro broad match or even BMM as it’s own campaign paired with RSA’s with Target CPA. I am like Google’s poster child LOL but I am telling you it is slaying it. @KyleShurtz
I set up campaigns and ad groups in some cases using only specific match types (exact, phrase, BMM) to see what we get in each one. The extra work has just come from managing negatives a little more closely. @lchasse
SQRs are a much more regular occurrence, embracing the intent-focused mentality, leveraging the N-Gram script & complaining to my Google reps non-stop. @nataliebarreda
Q5: What are some examples of crazy matches that you’ve seen? Please provide the query and the term it matched to.
This was a few years back but still my favorite. I worked at an online university “Western Governors University” and our brand campaign triggered “Vienna sausages for sale”. @KyleShurtz
Not keyword matching, but I’ve started to notice the Kleenex affect a LOT more in Google Shopping, matching brand competitor names to off-brand shopping ads. @_GilHong
[t mobile] matches to anything and everything about mobile. @nataliebarreda
See twitter.com/timothyjjensen… @timothyjjensen
Company name is Plethora. B2B custom manufacturing. TC Electronics just released a new guitar effects pedal called the Plethora. We’ve paused the branded campaign for a bit. @tunadonut
Google thinks “lice” and “dandruff” are interchangeable…@NeptuneMoon
Oh, something I can’t say…shouldn’t say! All I know is that I now have a Neg KW list labeled “XXX Terminology”! @loogleppc
So goes without saying…those SQRs are quite the task and my negative lists are long. @alexpeerenboom
Here’s some that are just fun. Don’t remember exactly what triggered them. ‘malpractice lawyer for lawyers that lie to murder victims’ ‘stripper I can punch’ ‘how to get cerebral palsy fl’ ‘amputated bitch’ ‘rubber mold of my girlfriends vagina’ @nikkijumpfly
Q6: Do you spend more time now with query reports and negative lists? Do you have any tools, methods or strategies that help make it manageable?
I have rules built to look for search terms spending over $x and just have it built into my weekly process for accounts. Makes it easier and you identify the top offenders so you are not spending hours doing it each week. @lchasse
I definitely spend more time in query reports and adding negatives. I don’t have a great solution at any kind of scale. @NeptuneMoon
Yes. Coffee and @mastodonmusic. @tunadonut
Yes, a lot more time. And more use of Negative Keyword lists than ever before. Wasn’t a big user of them, but now with some accounts (particularly those with Shopping campaigns) I’m hitting negative limits on campaigns. @nikkijumpfly
We use optmyzer, N-Gram script, Google recommendations (sometimes those are useful) @360vardi
I actually review my search queries each morning (the first thing I do). It used to take me a couple of hours at first and now it only takes about 30 minutes. The more proactive you can be, the less time it will take you to manage it. @loogleppc
I do spend more time in those areas, and has helped, but also led to a question that I haven’t been able to find the answer to yet. Perhaps can help now…@alexpeerenboom
And under the predefined reports for Search Terms, I’m only getting for 514 impressions and 150 clicks. So where is that missing data? Is it possible to see all the queries that generated 616 impressions and 29 clicks? @alexpeerenboom
Q7: If you could wave a wand and change one thing about how keywords match, what would it be and why?
Make Exact Match Exact Again! @jmloquist
I would change match types back to their OG matching: exact = only that query, phrase = words in order with maybe other words, BMM = words in any order + other words, broad = Wild West @Mel66
First, I wouldn’t go back completely to true “exact” match. I don’t have a problem with very close variants like slight misspellings, plurals, etc. in most cases. Would like to tighten up matching rules a LOT and include a method to report bad matches. @timothyjjensen
The + in BMM has basically become pointless at this point. Would return BMM to work the way it did previously instead of basically being actually broad. @timothyjjensen
Rollback to where: 1. Exact match was literally only an exact match. 2. Phrase was actually useful, as it only added words before or after your desired phrase & didn’t reinterpret it Seriously, BMM already covered what they are making others do. @NeptuneMoon
Ok, yes – this as well – not completely exact like 20yrs ago – but definitely not, different orders of the keywords. @mindswanppc
Exact match would be exact match. Phrase match would be phrase match. @stevegibsonppc
If Google could do what Bing does and “break associations” on an account level bases that would be bomb (baum!). @FindingAmanda
Exact match should be what it says it is Exact. Phrase match adding terms around your phrase was also useful and made sense. Broad Match was the wild west if you wanted to test seeing if there were terms you were missing and this was always fine. @lchasse
Have a bad match feedback mechanism within Google Ads. Let us help teach the machines why a match isn’t good. @NeptuneMoon
When Exact match used to be Exact! I have to admit, the new variation IS working well for me (it’s my go-to match type now), but….I’d still like a true Exact to be out there! @loogleppc
Already been said but give us a slider on how liberal we want to be with each match type. If I am bidding exact I want exact. @KyleShurtz
I don’t like the synonyms as close variants. I also don’t like the adding words to exact match. Just give me the words that are in the KWs. Google’s algorithm may be good, but it does not do a great job with nuance. @360vardi
I wish it was easier to explain to clients and non-practitioners that it’s all a bit fuzzier than everyone thinks. You can pretty quickly get folks to understand the 3 match types, but then you get “well wait but why?” questions that are totally fair. @ferkungamaboobo
Q8: What do you think the future of keywords in search is?
I hate to say it, but would not be surprised if there comes a time when keywords aren’t even an option. @timothyjjensen
At some point, we won’t get to pick them anymore and Google won’t show us the search queries (which will be taking us back to 2007 in the world of paid search). It will be plug and play, and take what you can get…@nikkijumpfly
IMO, Google would love to get rid of keywords. They’re already doing the hard sell to less sophisticated advertisers. And, of course, there are shopping ads… I reckon they want text ads to be like the same. @stevegibsonppc
My prediction for the future of search – fewer keywords, more feed based matching, and audiences. @FindingAmanda
It is hard to imagine keywords going 100% away, but I could see them being more categorical or harvested via a site crawl with little control over them…@NeptuneMoon
I see two potential futures for Search: 1) Audience-only or intent-only bidding (audiences built using keyword lists) 2) Keyword targeting as a modifier layered with intent/audience targeting but not the sole factor Not sure which is more likely. @akaEmmaLouise
The removal of Phrase match and more machine learning…somehow Google will fit that in there. @loogleppc
No KWs. Point Google to LPs and they will do the rest. Advertisers will have to trust they are working in the benefit of the advertiser. Meanwhile, we all know Google just wants to squeeze the most Dollars possible. @360vardi
I don’t think they will be fully going away in the near future. But I think match types will be dying sooner than later. Without keywords, it would have to be predictive, and no one is nearly good enough there. Yet. @armondhammer
Google is going to have “smart” keywords one of these days. The way things are going, probably not too far into the future. They are working to try and increase shareholder value and that means increasing advertising revenue (cost) for advertisers. @lchasse
PPCChat Participants
- Julie F Bacchini @NeptuneMoon
- Amalia Fowler @amaliaefowler
- Melissa Mackey @Mel66
- Daniel Vardi @360vardi
- Steve Gibson @stevegibsonppc
- Gil Hong @_GilHong
- Collin Slattery @CJSlattery
- Doug R Thomas Esq.@ferkungamaboobo
- Robert Brady @robert_brady
- Nikki Kuhlman @nikkijumpfly
- Steph Cocker @IamNextSTEPH
- Alex Peerenboom @alexpeerenboom
- Emma Franks @akaEmmaLouise
- Amore Digital @AmoreDigital
- Natalie @nataliebarreda
- Kyle Shurtz @KyleShurtz
- lchasse @lchasse
- Matt Lukens @tunadonut
- Amanda @FindingAmanda
- Jeff Loquist @jmloquist
- LooglePPC @loogleppc
- Jordan Stark @jord_stark
- Tim Jensen @timothyjjensen
- Andrea Cruz @andreacruz92
- Josh Warner @joshuadubs
- MindSwan @mindswanppc
- Steve Hammer @armondhammer
Related Links
- Broad Match vs Broad Match Modifier Match Types in Google Ads
- Google Ads Keyword Match Type Best Practices
- Seven Crucial Google Ads Mistakes To Avoid
Stop the wasted ad spend. Get more conversions from the same ad budget.
Our customers save over $16 Million per year on Google and Amazon Ads.