Message us

All great projects start with a conversation.

Thank you. We received your message and will get back to you soon!

Browse our work

or Send another message?

Google Encrypts 28% of Ads Queries – What Does It Mean For You?

Gareth Cartman

In a move that should surprise absolutely nobody, Google has decided to encrypt around 28% of your Google Ads queries. This means that if you are spending on Google Ads, you won’t know what queries are receiving clicks for around 1 in 4 keywords, and those keywords will mostly be the low-volume or “long-tail” keywords.

In other words, Google is saying “give us your money, let us spend it for you”.

Now, this is undoubtedly awful, and let’s not mince our words, Google is awful too. In a just world, this wouldn’t even be legal. You probably agree, but it’s important to assess what we can do from here – and what the implications are for the future of paid search given the very clear direction of travel.

Let’s look at the impact first to understand what’s really going on.

Not knowing what you’re getting clicks for is a terrible thing. 

Of course it is, there’s absolutely no benefit to not knowing what your users are clicking on.

Wil Reynolds posted an excellent example last night of what can happen:

Even on the best-run Adwords accounts, you need to keep a close eye on queries and what’s bringing in traffic, and what’s driving conversions. A long-tail keyword that converts like crazy is worth its weight in gold – this may now be encrypted so you don’t know what it is.

Worse, you may not know which keywords are draining money from your account, either. Like the bank that spent $60,000 on clicks from people who were searching for a rapper.

Google may argue that if people click on things, then they’re interested, but the data has always shown that this is not true. People click on all kinds of rubbish, and end up wasting you money. It remains essential that you retain the ability to weed out the rubbish clicks.

Add to this the awful scam that is Adwords Express, and it’s quite clear that the loser in this is the customer. There is no way that this makes it easier for Google Ads customers. But Google have the monopoly, so it is what it is.

OK, so is this the way things are going?

Yes, and it has been for some time. We’ve already seen Google suggesting new ads using its proprietary Artificial Intelligence, and while it’s still not as good as humans, it’s improving.

You can imagine a scenario whereby Google no longer tells you ANY of the search queries you’re paying for, and demands your blind faith that it’s doing the right thing, so long as you obtain a specific target Cost Per Acquisition for instance.

Take this a step further, and you could imagine a scenario in which you give Google £100,000 a month, and Google hides all resulting search queries and perhaps even starts sending your traffic to an AI-created webpage of its own to generate leads.

This is not so far-fetched. The message is – trust us with your money, AI will learn how to spend it, so sit back and trust Google.

But we don’t trust Google, so what do we do?

Many of you may remember when Google encrypted organic search keywords. That was the day I went grey, by the way.

But I’ve always said that if you want to understand how a species survives even the worst catastrophe, you should look at SEOs. Even after a nuclear apocalypse, the SEOs will still be there, glued to their desks, optimising for “post-nuclear apocalypse life hacks”. 

We’re good at this kind of thing.

One of the first things we did was look at landing pages instead of queries. Yes, it’s not as good, but it started to give us an idea of where the traffic was going, and what level of engagement those pages were receiving.

Secondly, we started looking more at site search. This gives you a much better idea of user intent when people land on the website, and probably reveals better keywords because the user is refining down from the original keyword. 

Thirdly, it put the emphasis on better content and a better user experience. I remember the Britney Spears days where all you had to do was put Britney Spears in the meta title and you would receive millions of hits. The early days of the Internet were just as vain as the modern era, only in a different way. It wasn’t a great user experience, and nor was stuffing the page with relevant keywords either. But we still did it.

Fourthly, we looked at Adwords keywords a lot more than we ever had done. There has always been a tight correlation between what you pay for and what you don’t – after all, it’s just one webpage. And from there, we looked at more data sources such as Trends, Search Console (still some keyword data in there) and third-party tools. We made do with the gap, and got better at serving people content that would convert.

Google wants to remove the paid management aspect of Google Ads. It would like to think it can run your Ads account for you.

In the future, this may well be the case. Despite the experience small businesses have had with the appalling Adwords Express, I don’t doubt the AI will eventually replace us. 

In the short to medium-term, however, it puts an even greater emphasis on paid management in order to prevent Google from wasting/stealing your money. If you’re reliant on a Google Rep to optimise your campaigns, I would suggest either moving away from your Google Rep or calling them multiple times a day so that you’re managing them really closely, like a team member you’re performance managing.

They have KPIs that are not your KPIs.

Furthermore, it puts an emphasis on a wider vision for Google Ads. It does not live in isolation, and you need to understand more than just how much it costs for each click or each conversion – you need to use your actual website data to assess the impact of queries known and unknown. High bounce rates, low pages per session, low time per session, poor conversion rates – all of this is probably feeding into the algorithm, but it educates you more as to what will improve your campaigns when you don’t have the query data.

Like what we do in SEO, basically.