Why I’m not sweating Google’s recent changes

If you met someone at a party and, without any prompting, they told you their motto was, “Don’t kill hitchhikers,” what would you think?

Me? I’d be wondering how many hitchhikers they’d killed. 


When I first heard Google’s motto was “Don’t be evil,” I thought, “Shit, I bet these guys are evil.”

And so it turned out to be. Not “evil” evil, just your “run-of-the-mill monopoly” evil.

And, when they dropped that motto 2 years ago, I burst out laughing. Imagine telling the world that “Don’t be evil” was too restrictive.

That takes huge balls.

So, let me pay my respect to Google’s balls… and let’s talk about their latest attempts at being evil…

The PPC world has lost its mind

My fellow PPCers are up in arms.

Why? Because Google has done something PPC managers feel is beyond the pale. Something they see as a game changer.

Well, I don’t know what game they thought they were playing – because, in my world, it’s the same game Google’s been playing for the last decade.

What game is that?

Google’s favourite radio station: WDYGTF

For as long as I can remember (and I’ve been managing PPC accounts for 14 years), Google’s has had a problem with ad inventory.

Let me explain with an analogy…

In Europe, radio stations are allowed to have up to 12 minutes advertising per hour.

They can’t go above that. And, if any of those 12 minutes go unsold, they can’t roll it over to the next hour. That ad revenue has gone forever.

And that’s a problem.

It’s a problem because advertisers want the best slots. They want rush hour and other peak times. They don’t want to pay for their ads to run in the middle of the night.

That’s why smart radio ad salesmen package ad slots. So, if you want to advertise at prime time, you need to buy ads for other times as well.

And Google has the same problem.

Just like radio stations, it has high-value inventory and low-value inventory. They have no problems selling ad slots for the high-value searches… but struggle to get a good price for the low-value stuff.

And, if someone does a search and there are few advertisers, there’s less chance they’ll click on an ad. And, even if they do click on an ad, because there’s less competition, the cost per click – i.e. Google’s revenue – will be lower.

So Google uses packaging. 

And those packages are called keywords.


The right way to think about keywords

Back in the day, Google Ads had four different match types: exact, phrase, modified broad and broad.

Now it just has one: broad match.

But there are four versions of it. There’s exact broad match, phrase broad match, modified broad, and “are you feeling lucky, punk?”

That means every one of your keywords is a basket of search terms. Some of those search terms convert well – and deserve to have higher bids. Some of them convert less well and should have lower bids – or be blocked altogether.

Now, in an ideal world, you’d know exactly what is in each basket and you’d be able to…

#1: Decide whether you want that search term to be in the basket.

#2: And, if you do, how much you’re willing to bid on it.

That would be good for you, the advertiser. But it would leave Google with a bunch of inventory that’s either unsold, or has to be deeply discounted.

Not good for Google. 

Fortunately for Google, it’s a monopoly – i.e. if you want to run Google Ads, you have to buy those ads from Google.

So they get to make the rules.

And, as far as they’re concerned, if you don’t like those rules, then – as we say in Scotland – WDYGTF.

(“Why don’t you get to fuck?”)

And that brings us (finally) to what the PPC world is so angry about…

Google’s latest douche move

Google has decided to reduce the amount of information in search term reports.

I mentioned earlier that keywords are baskets of search terms. Well, search term reports let advertisers see what’s in those baskets. 

Or, at least, some of it.

In the past, if a search term produced a conversion, we’d almost always get to see it.

If it produced a click, we’d see it less often, but still most of the time.

And, if it produced an ad impression, but no clicks, we would rarely see it.

And that was useful because it allowed advertisers to…

#1: Identify searches they don’t want to show for – and block those with negative keywords.

#2: See which search terms convert well within a keyword and add those search terms as keywords – and set their keyword bids accordingly. That way, you’re directing your spend away from low value searches and towards high value searches.

#3: Spot relevant searches where you have a low click rate and move them to a new ad group with a more relevant ad.

#4: Identify overlaps between ad groups and remove them.

Having less data means less ability to do those things.

So, inevitably, your accounts will be less well optimised. Which means wasted spend and wasted opportunity.

This sounds bad. And it is bad.


How much data are we losing?

One agency claimed they went from seeing search terms for 95% of their clicks, to seeing data for only 33% of their clicks.

Another agency says they’re still seeing data for 72% of clicks. 

I trust that report more. But 28% is still a lot of data to lose. 


Why Google’s doing this

Google says they’re doing this to protect users’ privacy. Here’s their explanation:

“Starting September 2020, the search terms report only includes terms that a significant number of users searched for, even if a term received a click. You may now see fewer terms in your report.”

Frankly, that sounds like bullshit to me. I think it’s a thinly-veiled cash grab – making advertisers advertise on search terms they don’t want. 

That would be consistent with Google’s past behaviour.

But, either way, at least it tells us what data they’re removing: they’re taking away the lowest-volume searches.

So what are these search terms?

Well, according to a 2017 report, 15% of Google searches are unique. That is, they’ve never been seen before.

Think about that, there are around two trillion Google searches done each year… and 300 billion of these have never been seen before.

That’s extraordinary. (Though many of them will simply be typos.)

And there’s going to be a large overlap between the 28% of data we’re losing and these 15% unique searches. 

So half of what we’re losing is data for searches we’ll probably never see again. So blocking them with negative keywords would have had limited value.


“But what about n-grams?”

There’ll be PPCers who, right now, will be asking, “What about n-grams?”

Now, if you don’t know what an n-gram is, it’s basically grouping search terms together to see how different words impact conversions.

An example would be to take all your search terms that included the word “free” to see how they performed. 

The reality is that we can still do n-grams. Not to the same level as before, but we can still do them with the data Google is still showing us.

So, let’s recap: we’re missing around 28% of our data.

But, 15% of that is probably unique searches we wouldn’t see again – so it has limited value as negative keywords.

What about the other 13%?

Well, let’s understand one thing, some of these terms you would have blocked… but many of them you would have left.

So it’s not 13%, it’s maybe 5% or 6% that the real problem.

And not only that, but…

Your account isn’t getting worse, it’s just not getting better

Now, remember, we’re already showing for that 5-6%. So, if Google hides it, then our accounts aren’t getting worse. 

It’s just harder for us to make them better.

So, what’s the impact going to be?

Well, in theory, your numbers should stay the same.

You’ve got the same search terms triggering the same ads and going to the same landing pages. So, in theory, you should have the same CPA/ROAS.

And, because your competitors are in the same boat, their bids should stay the same, too.

So the net effect is zero. Everyone stays where they were.

Except, of course, there’s opportunity cost. If having this data would have made your campaigns a few percent more profitable over time – then you miss out on those extra profits.

But that’s normal.

Click inflation outstrips real inflation 10:1

You see, when I started managing PPC accounts in 2006, the typical cost per click was around $0.40. Ten years later, it was around $2.10.

That’s an increase of 425%.

To put this in context, if click costs had increased in line with inflation, they would have increased by 19.1%. Or, to put in dollars, by just 8 cents.  

This means click inflation increased at around 18% a year, while real inflation increased only 1.76%/year. So the cost of clicks is growing 10x faster than inflation.

So dealing with rapidly increasing costs is just part of being a PPC manager. And 5-6% is nothing out of the ordinary.

I’ve got 99 problems and these clicks ain’t one

So, yes, it’s a drag that Google is hiding this data from us.

But if you think that’s bad, try working with a SaaS company that’s getting outbid by a competitor that only cares about growth and doesn’t care about profitability.

(And has millions of dollars of VC cash behind it.)

That’s a problem. 

This? This is an inconvenience.

If you were planning to stand still, you had the wrong plan

As we saw, the cost of a click goes up every year – far faster than inflation. 

So, in PPC, if you’re standing still, you’re going backwards.

If you’re standing still, you’re a sitting target for any competitor that’s…

* Optimising their accounts.

* Split-testing their ads…

* Split-testing their landing pages…

* Optimising their cart or signup pages…

* Increasing the value of their offer…

* Adding up-sells, cross-sells and re-sells…

* Increasing the lifetime value of their customers…

(My book, The PPC Multiplier Method is all about how to do these things strategically.)

So, presumably, you weren’t planning to stand still. You were planning to improve the performance of your PPC funnel.

You were planning to do all the things I just listed. Me, too.

Well, any one of those will have a far greater impact than losing this search report data.

And that’s why I’m not sweating these changes…

Steve Gibson

PS I know some PPCers will say, “Don’t you realise this is the start of the slippery slope towards having no control over our keywords?”

Listen, if you haven’t realised we’ve been on that slippery slope for the last 7 years, you’ve not been paying attention. 

PPS You know what’s funny? Google could easily increase its ad revenue by at least $1bn a year – without doing any of this bullshit.


By using an old idea from newspaper advertising.

That’s all I’m going to say. If Google wants the idea, they can pay for it.