PPCChat | Google’s Latest Keyword Matching Changes | 9th February 2021

Posted by & filed under PPCChat.

This week’s PPCChat discussion was focused on the recent announcement from Google Ads regarding BMM folding into Phrase Match. Host Julie F Bacchini sought PPCers reaction towards this announcement, how are they planning to work in this new system, will they consider using BMM after this announcement, and more.  

Q1: What was your initial reaction to the news that BMM will be folded into Phrase Match, starting this month?

I was surprised they didn’t go the other way, eliminating phrase match. After all, that’s essentially what they’re doing – phrase match has been bastardized from what it originally was. @beyondthepaid

Definitely not surprised, but I would have preferred Phrase Match be rolled into BMM and not the other way around since that still technically works as it should. @_GilHong

The eyeroll heard around the world. I mean, all the platforms want us to just ‘let the machine think for you’ and this is the google version. @JuliaVyse

Kind of annoyed but not really surprised, had a little eye roll at them cutting off BMM and turning phrase into broad to make us ~feel~ like we have a little more control. @jennifer_lash

I feel like it was the announcement that made the rest of the marketing world aware of what I’ve been experiencing with matching for the past several months. @robert_brady

My profanity laced thoughts were basically that I already had ENOUGH to do based on Google changes and announcements from the last 30 days. My next thought was that they are definitely using COVID-instability as a cover. @ynotweb

Not surprised but definitely not expecting it so early in the year. @anna_arrow

I thought about two things: 1) That my keyword strategy got much simpler to QA, and easier for clients to understand. Much easier to say “no broad” than “well, this is broad but don’t worry, we know what we’re doing.” @ferkungamaboobo

2) How many accounts I’ve seen with BMM/MBM/MBMBAM errors because people didn’t understand the right way to do it or a small error was made in some regex. In general, this is a win for non-experts. @ferkungamaboobo

I think it’s trash but just a continuation of the trash we’ve come to expect from Google. “Take what we give you and be happy. @CJSlattery

I mean, if I was head of google voice search I would push to keep phrase over bmm. @JuliaVyse

This gif pretty much sums up how I feel. Really not surprised at all but I do think it would have made sense for Google to roll phrase into BMM instead of the other way around. @alexnicoll93

(More) lack of control. Google pushing Hagakurish automation thr ML and maybe the begining of the end of matchings entirely and even keywords as we know them at some stage. @edgar_sanchez

Five letters AYFKM. I’ve always perceived BMM as a bit of an experts play, and a hack that worked well to get over low volume KW in b2b. I thought that was a shot straight at us.@armondhammer

Modified Broad was a feature to give us more control – so they are taking it away. Google is “going back to the basics”… but this time the basics come with less control… But “trust us” – we’ll respect your intent, but hide the ability know if we do. @yaelconsulting

A face palm moment but also chastising myself for still not being better prepared for changes from Google that i don’t like. @TheMarketingAnu

Call me a conspiracy theorist, but this is some BS. just another step for Google to push us out of keyword usage in the near future, so they have more control of what we do. @JonKagan

Q2: Has your thinking or opinion about this announcement changed after some time to digest it?

Not changed, I just tend to prefer more options rather than fewer. Do new stuff to deal with voice search and conversational matches. but don’t take away options I had prior. @JuliaVyse

Nope. just got me more cranky. @JonKagan

Yes- this could actually make building accounts out a lot less stressful. @TheMarketingAnu

I also think there are a series of events that are moving us down a series of privacy initiated steps #ppcchat I’m trying to map it out, this is what I have for now, there is no Facebook on it, just because FB is a mess. @jimbanks

Trying to stay optimistic, and am also thinking about ways to report out on this update like pausing the BMMs instead of changing the match types and uploading the updated Phrase Match versions so that I can report out a MoM change. @PPCKenChang

It’s made me think more about the old standard “15% of all queries are new” statement that was given a long time ago…. I wonder if that rate has changed and contributed to some of these changes. @_GilHong

I was not surprised but felt discouraged – this wave of loosen targeting shouldn’t be surprising anyone anymore. @andreacruz92

Not really. And the shrinking of the search term report in September (plus COVID) gives them the power to do whatever they want with little accountability. @robert_brady

I’m confident that the impact will be minimal across client accounts. I actually see keywords at the lower end of “things to fix” and would rather focus on site speed, ad copy, extensions and the like. @jimbanks

“At least they gave us forewarning”, something they didn’t always do in the past. @ynotweb

At first I was annoyed that it was happening and that it would require some account restructuring but now I’m more like “oh well” It’s going to make management slightly easier in accounts using BMM and PHM, and it will cut back on some unwanted BMM queries. @MarketingByMark

2 I just miss phrase match the way it was and my feelings about this change are mostly saying goodby to an old friend lol @MarketingByMark

My feelings haven’t really changed since the news broke. This falls into the time I build in for every client for the inevitable “random acts of Google/Facebook” that I have to deal with at some point each year and it is not in an ongoing scope of management work. @NeptuneMoon

Not really. Unfortunately, I have come to expect this over the last couple of years. Google taking away more and more levers and controls its annoying but I’m almost numb to it now. @selley2134

Feelings remain the same. Expect these sort of changes quarterly as Google continues to try to get topline revenue growth by squeezing ever more dollars out of advertisers at the expense of said advertisers. @CJSlattery

It’s been interesting to see a somewhat more measured response out of the PPC community than just “GOOGLE BAD” I agree with @PPCKenChang that the implications are greater on Bing and Amazon and everyone else that adopted Exact/Phrase/Broad/Mod Broad. @ferkungamaboobo

Every single change Google makes to their platform is to boost topline revenue growth. This means taking shit, putting it between two slices of bread, and forcing us to buy it as shit sandwiches. @CJSlattery

My thoughts haven’t changed really at all. I’m fortunate enough to be in a position where I don’t have to explain the change to clients, so I can just adapt and move on with my work. @alexnicoll93

Not really, I’m feeling mostly the same about it and the keyword changes that have already been happening…just have to take it and make it work. @jennifer_lash

Not even surprised really. If you work with reps you can see they’re pushing for automation as much as possible. I’m more worried about the future and bigger changes they may roll out with. @ValenciaSEM

I’m still not happy, because I’d actually prefer it to be BMM that was retained over phrase. It’s going to be hard to explain to newbies because phrase isn’t phrase. Exact isn’t exact. Words mean nothing. (1/2) @armondhammer

However, given the mess that matching has already become, I’m actually slightly positive this might be OK. I’ve moved a lot from massive KW lists to much shorter ones. This is a further push in that direction (2/2) @armondhammer

The trend to take away our control is inevitable… How they do it may catch us off guard, but our only option has always been, and still it, to deal with it. So reaction was: “WHAT”, “NO”, “OK”, “FINE”, “NEXT”. @yaelconsulting

Q3: What, if anything, do you plan to do in your existing accounts and campaigns as a result of this change? Does this differ from what you plan to do for new campaigns and/or accounts?

We’re not changing much initially, but will need to restructure entire accounts that were based on exact/BMM. @beyondthepaid

I plan to replace all our BMM ad groups with phrase match ones. What won’t change: only making use of “pure” broad match keywords for research purposes with heavy negative keyword lists applied. @PPCHartman

I may test more phrase matches with heavy negatives. We don’t get enough volume from exact only, so I need to fill in the bmm void somehow. An voice search makes sense for food and public sector. @JuliaVyse

I don’t foresee there being a ton of value in pre-adding a bunch of BMM keywords before the deadline, but any new builds now will be with Phrase and Exact. @_GilHong

BMM vs Phrase match MoM reporting for the first 1 or 2 months after the update @PPCKenChang

We’ll switch from Exact and BMM to Exact and Phrase and continue to look at moving more budget to “managed” or “smart” campaigns, they fit better in the new privacy world, which Google is giving us in Chrome in 2022 @jimbanks

I had already stopped using BMM for the most part after the search query purge (a lot of clients are on strict low budgets). Probably use more and more Exact match but it doesn’t change my plan much. @selley2134

I don’t see myself changing much initially besides doubling down on the pre-emptive close variant negative strategies I started when we lost search terms With less keyword options I may actually test this broad match strategy. @MarketingByMark

I am actually leaning hard back into exact match only. I’ve been doing this since the September change TBH. I’m one man trying to hold back the tide. @CJSlattery

Same, just use more phrase than we did (ModBroad basically replaced phrase because the order of words was convenient for the most part. Trusting Google with when they determine order of words matter vs. not is hard. Especially when we get less query data. @360vardi

Am going to have to work on pausing BMM terms and putting equivalent phrase terms in (where they are not already in place) for existing accounts. For new stuff, will do a combo of phrase & exact. But, will have to see how broad phrase goes in the new system too. @NeptuneMoon

I mostly use BMM for keyword mining, so I expect most of my keyword mining to be picked up by phrase match from now on. Moving forward, I won’t build out any BMM keywords and just trust phrase match to take care of my keyword mining for me. @alexnicoll93

I don’t think much right away, but eventually will need to restructure things that were based on BMM and move to new builds using the new types only. @jennifer_lash

We will likely trim the fat, so to speak, more aggressively first and then do a phased approach to revamp. New campaigns, of course, will be built in a way to easily switch over. @ynotweb

Shouldn’t be too disruptive as we just add phrase and exact only terms. On the bright side, we won’t have to deal with those annoying ‘+’ anymore in our spreadsheets. @ValenciaSEM

This paid being behind the times, as I still mandated the team to use phrase and BMM. Now we’re going to have to toe the water in regular broad…like it is 2005 again. @JonKagan

Q4: Will you consider using Broad Match differently after this announcement? (Because we all know Google really wants us to do this…)

When it’s relevant. I’m always willing to test, but the nonsense that broad brings in just causes negative keyword list headaches galore. @JuliaVyse

Don’t see the point! Our move will just be from using BMM and Exact – to using Phrase and Exact. And now DSA to mine keywords we haven’t thought of. that’ll be my recommendation anyways. @TheMarketingAnu

I’d rather not but seems like I might be forced. It will ultimately come down to the client’s goals/budgets and what we can do without it. @selley2134

I would honestly still not use it. With Phrase becoming more like Broad I might loosen up some of my Phrase keywords for mining. @_GilHong

I loathe broad match. At least my past experience with it has been terrible. Especially because I think we are going to get less and less information and be forced to choose access vs. control (or the illusion of more control). @NeptuneMoon

The answer that we all hate to hear, “It Depends” will run some small test campaigns to see how it works out to see if there is anything different in the change month. @PPCKenChang

I’d be more willing to use broad if they weren’t hiding 50%+ of the click volume. @CJSlattery

I think I will definitely reconsider using broad match, especially when clients need more volume. I’ve been doing more testing with DSAs and would totally consider running broad match combined with in-market audiences as a test. @MarketingByMark

It will take some time to mentally digest that Phrase = MB + Phrase, per the definition… while fighting the urge to go Broad. But as with everything else, we’ll test and follow the data. @yaelconsulting

Ultimately, I’ve only ever looked at BMM to give us additional volume for clients when they still have money to spend and to mop up anomalies. I might be more considered in how I approach Broad Match, but it has to tie back to a business goal we are trying to hit. @jimbanks

That being said I’m definitely not EXCITED that I’m considering broad match as an option. It’s a tough pill to swallow. @MarketingByMark

Probably not. For what I do I employ very few true broad match keywords with a lot of negatives. My strategy will continue to be exact match and phrase match heavy with some DSA for additional keyword mining. @alexnicoll93

More exact, where I can. I know have the joy of mining external sources to try and add negatives. Pure broad is horrible for B2B, because of crossed up intent, consumers and people trying to figure out their homework. Give me an audience that makes sense. @armondhammer

Absolutely NOT. Way too much evidence from “suggested keywords” to indicate Google AI is wayyyyy too far off to trust it. Plus, the ubiquitous “and missing search query data!” gripe. @ynotweb

Broad testing has worked somewhat with heavy negatives and targeted audiences. So audience target (not observation) plus a broad term with about 4 negative keyword lists is the only successful test I’ve run so far. lots of setup work. @JuliaVyse

This is where Google messed up. If they want us to use Broad match more they can’t take away search queries. I’m not going to mine for something I can’t see and optimize. @ValenciaSEM

I think it’d have to depend on the account, generally I want to say I’d prefer not to (esp with the visibility issues) but I can’t count it out all the way for lower search volume clients – maybe with heavy audience targeting, negatives, and dare I say smart bidding @jennifer_lash

Q5: How does the less visibility into search query data play into this for you? Either your reaction and/or your plans to work in the new system?

1: Does anyone know if the legacy N-gram script is only fed by the search terms we can see on Google? I assume the answer is yes: searchengineland.com/brainlabs-scri… @andreacruz92

This is a huge factor! If I could see where the money is going I’m willing to test and optimize. When I can’t see what I’m spending budget on I am going to try & keep as much control as possible. @selley2134

See Answer A4 @ynotweb

Nothing Google has done has made me more disappointed in Google than the less visibility into query/placement data (whether in search, smart shopping, smart display). It is the one lever that you can always turn to and say “we are/ or not showing to relevant users” @360vardi

Loss of query data definitely makes me cautious of testing literally anything in search. IDK how I’m supposed to trust something I can’t see. @MarketingByMark

Well, it’s the big problem. I can do n-grams, I can do light mining every day, but if not enough searches actually show up, the relevance between my program and our audience is opaque. That’s why google wants to keep that data to itself. @JuliaVyse

The less visibility into search terms just means Broad will be showing our ads against a lot more terms we aren’t relevant for that we can’t see meaning we can’t exclude. So this just means DEATH to Broad! @TheMarketingAnu

The query data loss is my biggest concern right now with Google Ads, period. I am a big “trust, but verify” type of person in business and the lack of ability to, in fact, verify is VERY problematic for me. If the ML & automation & AI are so good, SHOW US. @NeptuneMoon

We are not getting less data, we use the valuetrack parameter for keyword (which is {keyword} btw) and pass it as an extra parameter and pick it up in Lucky Orange and BigQuery @jimbanks

We’ll need to get more proactive with negative keywords by leveraging @MSFTAdvertising data and our experience for the past decade. I’m planning to roll with the punches to phrase match and rely on the data and results to lead. Still can’t see need for pure broad. @yaelconsulting

Less search term visibility is certainly a loss for good PPCers, but I don’t really think the problem is exacerbated by the transfer of BMM to phrase. I think how we manage BMM with the change is simply how we must manage phrase moving forward. @alexnicoll93

I think it really highlights the need to think before you launch campaigns. Do the keyword research to find what isn’t the right intent for your ads Do the research to find what might be the right match This is a jab at my processes but IMO it was easy to be lazy. @ferkungamaboobo

I think this is the biggest issue. Making our targeting broader while taking away our ability to correct for it. Our ROI is in direct competition to Google’s ROI. @ValenciaSEM

Query data plays a HUGE part. they want us to use broad but won’t show us the matches so how are we supposed to optimize? using ML that we also don’t know exactly how it works…giving control & taking it away is way worse than never having control in the first place. @jennifer_lash

They are trying to put up an ugly man behind the curtain move on me @JonKagan

Q6: If you were going to modify how match types worked, assuming that changes to this point could NOT be undone, how would you do it?

I would make visible keyboard vs voice searches, then try to match emphatic terms. I didn’t mind vintage exact + misspellings, I didn’t mind bmm @JuliaVyse

This sums it up @JonKagan

I would make Exact actually be exact again. No “close variant” nonsense for Exact terms. Let Phrase be this new Phrase+ and have Broad stay fully broad. @NeptuneMoon

“Changes to this point could NOT be undone” I would create True Exact Match and it would work like the old exact match without close variants. @ValenciaSEM

Keep whatever changes they want, just give me query data. The only way matches are reliable if you can have visibility and can actually do something about it. @360vardi

Given that so many users journey involve other factors, not just keywords I think rolling something back because it worked before would not help. Keywords will eventually be worth less than context. And context can make the implication of the keyword different. @jimbanks

Q7: If you could influence how keyword matching worked without restriction or constraint, what would you do?

I would continue with the conversational learning, and identify intent words in a longer phrase. Like take a long tail, diagram it like a sentence and say to the algo, these three bits are the MOST important, or that one qualifier changes the meaning. @JuliaVyse

Full exact match back and bring search query reports to what they used to be. @andreacruz92

Go back to the old method, and make variant matching as an option, rather than a “hey advertiser, want to get sad for a bit?” @JonKagan

This is not exactly the intent of the question, but if Google wanted to help, they would have AI and recommendation to help us tell THEM what is NOT a match for our purposes. @ynotweb

1. Search query data 2. Option to turn close variants on/off 3. Modifiers for Exact & Phrase to prevent certain words from being dropped/modified in close variants (ie when “software” is swapped for “system”, you could put a + in front of software. @MarketingByMark

If clients wouldn’t abuse it (either intentionally or accidentally), I quite like the blind pay-per-conversion mindset of Local Service Ads. Who cares what you show for, what gets clicked, when you can just pay for qualified leads/sales? @ferkungamaboobo

A philosophical answer. Since focusing on looking ahead as to what is coming, rather than behind on what has passed has reduced my stress level a lot. Google now vs. 2002 it is a different multi-headed beast that operates under one roof, but it is disparate @jimbanks

Having 4 match types behave as intended, would be PERFECLY fine. We would then have the ability to eat exactly what we ordered. Problem is Google treats search like a Hotdog, shoving in “inedible” queries… figuring hiding “how the hotdog is made” will help. @yaelconsulting

Q8: Anything else you would say to @GoogleAds – if they were listening – on keyword matching?

We have to adapt, so keep doing what you need as more people search out loud. Bring back search terms – we’re all NDAd anyway – and show voice vs keyboard searches. @JuliaVyse

Literally LOLed at the thought of @GoogleAds listening. I would say A. Please start listening to feedback from your advertisers, we’re your customers. B. We can’t trust your automated products without data so either give us data or stop automating everything. @MarketingByMark

Give us data and we will test & choose the best option. Hide data and I’m going to search for any possible workaround not to use your system. @selley2134

Would like to know what their “timetable” is for the phase out of BMM. For example, do I have till the end of April 2021 or the end of July 2021? @PPCKenChang

Hahaha Again, see A7. Google needs to be using an entire account set of campaigns to determine fit and intent for KW. I, for one, am sick of dismissing suggestions for KW in one campaign when it fits another campaign better. @ynotweb

PPC is not a vacuum. Data they comes from paid search initiatives is very useful outside of your system (like landing pages) and for general messaging. SHOW QUERY DATA. SHOW QUERY DATA. SHOW QUERY DATA. SHOW QUERY DATA. SHOW QUERY DATA. SHOW QUERY DATA. @NeptuneMoon

What everyone else is saying – if you’re going to automate things, cool, but give us the data to do what we need to do within that automation rather than trying to hoard control. @jennifer_lash

We’re not idiots. We can help. Even the best of us might not be able to feed the machine all the data you want. (closed leads vs crap leads in b2b is hard) – The machine needs some supervision, and we’re an army that wants to make it happen. @armondhammer

Stop messing with stuff that doesn’t need messing with. It’s like you didn’t learn your lesson from Enhanced Campaigns. Also, we want the old UI back. Sincerely, Everyone @JonKagan

They are too arrogant to listen to what we have to say. They have an agenda to work on, they very occasionally walk something back, but they have a catalog of unpopular tweaks that we collectively don’t like and if we want to move elsewhere I am sure they are cool. @jimbanks

My issue isn’t really with the keyword matching, it’s with the search term visibility. I can adapt to the match types changing if I have the visibility to make them perform well! @alexnicoll93

As much as they think their AI knows to tell whether a query is appropriate to an account, EVERY account even if they sell the same thing, has different goals and needs more nuanced view that unfortunately, their system can’t tell (yet) @360vardi

You’re a new search engine or platform away from becoming myspace so you better start working with us (jk). @ValenciaSEM

@GoogleAds built itself on the ability for Advertisers to know what’s working. “Half the money I spend on advertising is wasted; the trouble is I don’t know which half” OBSOETE…. advertisers rejoiced, paid, and G built an empire from it. You’re going backward. @yaelconsulting

You can add machine learning and AI without diminishing our ability to view data. In fact, you need us to help you feed the machine the right data or you going to have a lot of unhappy customers. @jord_stark

PPCChat Participants

Related Links

Stop wasted ad spend with Karooya

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.

Start free trial Schedule a demo

Leave a Reply

Your email address will not be published.