How the mix of people who see your ads affects AdWords performance.

This is an extract from my book Diagnosing AdWords: A troubleshooting guide for solo AdWords professionals.

The book will help you get to the root of the problem by following a methodical process. Narrow down possible causes. Home in on the exact source. Eliminate hit-or-miss guesswork and be assured that you know what went wrong and how to fix it.

Your audience is all the people that Google could show your ads to. - People whose search matched your keywords and match your location, demographic and other filters.

  • Some of those people will search on mobile, some on desktop.
  • Some will be at home, others will be at work.
  • Some will search on a Monday, others on a Sunday, and every day between.
  • Some will search early in the morning, others during working hours and others in the evening.
  • They’ll be scattered throughout your targeted location.
  • Some will be searching as part of their research before buying. Others will be ready to buy now.
  • Some will find your ads on Google, others might find them on search partners or the display network.
  • etc...

This audience mix changes all the time. Those changes make a significant difference to almost every aspect of AdWords performance.

Sometimes change in traffic characteristics looks like a problem.

Panic! Your conversion rate dropped from 13% to 6% compared to last month. The months are similar enough to rule out seasonal fluctuation.

Here are the numbers.

Month 1: 800 clicks. 106 conversions = 13% conversion rate.
Month 2: 800 clicks. 50 conversions = 6% conversion rate.


On the surface this looks like a conversion rate problem. But, if we segment the clicks by device, a different picture emerges.

Month 1.

Desktop: 650 clicks. 98 conversions =15% conversion rate from desktop clicks.

Mobile: 150 clicks. 8 conversions = 5% conversion rate from mobile clicks. Total: 800 clicks. 106 conversions = 13% average conversion rate.

Month 2.

Desktop: 100 clicks. 15 conversions = 15% conversion rate from desktop clicks.

Mobile: 700 clicks. 35 conversions = 5% conversion rate from mobile clicks. Total: 800 clicks. 50 conversions = 6% average conversion rate.

We don’t have a conversion rate problem.

  • Desktop clicks are converting at the same rate (15%).
  • Mobile clicks are converting at the same rate (5%).

What we have is a change in the mix between mobile and desktop clicks. If we tried to solve this as a conversion rate problem we’d get nowhere. Instead the answer to the lower conversion rate lies in finding out:

  1. Why we got fewer desktop clicks and more mobile clicks.
  2. Why the mobile searchers convert at a much lower rate than the desktop ones.

The reason for the search changes the mix.

Your future customer will do some research before buying. How much research depends on the cost and complexity of the goods or services they buy from you.

A few keywords let us target people closer to the buying end of the process but most keywords don’t offer that insight. That means that most of the time we won't know exactly where the searcher is on the buying process.

The people who see your ads will be somewhere between starting their research and being ready to commit. That makes it difficult to write ads and landing pages.

A headline "Fast Free Quotes" or "Fast Free Estimates" might appeal to someone who is close to committing. If lots of your searches this week came from people close to buying you'd have a high clickthrough rate. It would be less attractive - and thus less likely to be clicked - to someone early in the research phase.

It's the same after the ad has been clicked. Your landing page will favour a searcher at one end of the process over a searcher at the other end. That will show in the conversion rate.

I consulted on a campaign that had a great landing page -> lead conversion rate. A junior employee ran the AdWords campaign. She'd worked hard to get it to produce a strong flow of sales enquiries.

The boss decided that it would be more efficient if people could do their own quote on the website. That way the sales team wouldn't have to phone the people who filled in the enquiry form.

The boss hired a web developer (at great cost) to build an online quoting system. It took a couple of months but when it was done a visitor could complete a series of forms on the site and get their own quotation. The quoting system tied into the business's accounting system and workflow planning. It had all the bells and whistles.

The employee contacted me for help about 3 months after the quoting system went live. Her page conversion rate had tanked. The flow of quotes dried to a trickle.

The boss wanted to know why AdWords wasn’t working. Our hero was the target of his unhappiness.

We looked at the search terms that brought people to the site. A few indicated that the searcher was in buying mode. For the rest it was impossible to tell. The searchers could have been in the early research stage or ready to buy.

An online quoting system is a turn off for someone who is researching. It’s too early in the process. It's like suggesting spending the night with someone you've just met.

The boss was pretty attached to his online quoting system because he’d spent so much time and money on it. We couldn’t get rid of it right away so I suggested splitting the traffic.

We sent half the people who clicked to a page with the old enquiry form and the other half to the online quoting system.

The people who saw the old enquiry form filled it in at about the same rate they did before the change.

AdWords wasn’t the problem. The problem was a mismatch between the searcher’s intent (research vs buy) and the ads and landing page.

When people search changes the mix.

I have a client who repairs electric driveway gates. His ads run 24/7 but there are two periods in every work day where his conversion rate skyrockets:

  • In the morning when people leave for work.
  • In the evening when people get home from work.

That’s when they hit the remote to open the driveway gate and smoke comes out of the motor.

His average conversion would be much lower if his ads were shown mostly in the middle of the day or late at night. The number of impressions and clicks might stay the same but people searching at those times don't usually have an urgent problem. They're less likely to hit the call now button.

Where people are when they search changes the mix.

I have a client who runs motorcycle tours from Cape Town in South Africa. We advertise in two locations: South Africa and the United Kingdom.

He gets 100x as many impressions from people in South Africa than he does the United Kingdom.

His South African ads face stiff local competition. His campaign averages around 5% CTR because there are lots of ads to choose from.

His UK campaign gets > 40% CTR because he's usually the only one advertising.

The words people use to search changes the mix and the results.

Your campaign probably has an assortment of keywords. They’ll range from wide to very focussed long tail keywords. You might have the same keyword with different match types.

Some keywords are cheaper, some have better CTR, others convert better.

The distribution of impressions, clicks and conversions across that medley of keywords changes. That causes a change in results.

What device people search on changes the mix and the results.

The mix of traffic by device - mobile phone vs desktop computers - is significant. I don’t think I’ve ever seen a campaign where desktop and mobile performance is the same across any metric. CTR is different, CPC is different, conversion rate is different.

When the ratio of desktop searches to mobile searches changes you should expect big downstream changes.

Where your ads show changes the mix.

You could show your ads on the display network and the search network.

  • The display network is made of sites that host Google adverts alongside their articles. Usually news sites or blogs. You'll see a bunch of ads in a block at the top or the side.
  • The search network consists of Google search (,, etc) and search partners.

Search partners are hundreds of non-Google websites (like AOL), as well as Google Maps, YouTube and other Google sites.

New campaigns default to show ads on both Google search and search partners. That gives your ads the biggest possible exposure.

Performance on Google search can be very different to performance on search partners. One of the campaigns I manage had the cost per conversion double on search partners. Disabling search partners and showing ads only on Google search freed enough budget to fund an extra 60 enquiries a month.

This is an extract from my book Diagnosing AdWords.

The book will help you get to the root of the problem by following a methodical process. Narrow down possible causes. Home in on the exact source. Eliminate hit-or-miss guesswork and be assured that you know what went wrong and how to fix it.

Other articles you might find useful:

How long does AdWords take to work?

You’ve chosen good keywords, written decent ad copy, set up a landing page that should convert. You’re getting impressions and the CTR looks ok but you’ve still not had any conversions. Is this normal or should you worry?

Is your AdWords running at 100%?

Are you worried that it might be underperforming? Perhaps you've neglected your AdWords account. Maybe it used to do better but now it’s costing more and delivering less. Maybe you’re behind on some updates or not confident that everything is as it should be.

How to get fast traction on a brand new AdWords campaign.

I worked out a process for getting fast results from new AdWords campaigns so I didn't have to worry as much. You're welcome to copy (and improve) my process.

I ignored quality score for a year.

I used to think quality score was vital, but in the year I’ve ignored QS and focussed on other metrics: - our ads have been clicked more often. - Our landing pages are better at turning visitors into leads. - Our clients are getting superior leads from their AdWords spend.

Doing the wrong things right.

I've been lucky enough to consult with several clients who run their own AdWords or have an in-house AdWords team. I usually get called in because they're not getting the results they anticipated. They’d hoped for a reliable flow of new business but they're paying Google more and more for fewer leads. One mistake crops up all the time...

Putting a little science into AdWords

In the past I relied on a combination of the change history in AdWords, some rough notes and my memory to keep track of my testing. It worked but it wasn't great. It became more difficult as my business grew. I started to feel like I was losing control. I couldn't answer questions about an account without having to root around in Google docs, AdWords and email. Read on to see how I fixed this.

Should you use Target CPA bidding for lead generation campaigns?

Google wants you to use Target CPA bidding. Is it right for lead generation campaigns? Read on to find out

Managing AdWords? Want a quick win?

This is for you if you’re managing AdWords for efficiency - you want to reduce you cost per lead (CPA). The golden rule when managing AdWords for efficiency is to do more of what works and less of what doesn’t. The concept is simple but actually doing it is more difficult because you have to ...

© Peter Bowen 2018 | Isle of Wight