Do some testing

Andrew says: “Test, test, test.

Test a lot. Test lots of different things. Do some testing.

How do you know what to test?

“You don't want to test everything. What I normally advise is testing what your customers use, or what your target customers use.

I used to work with a company where the boss would come and say, ‘This doesn't work, the website looks awful on my computer,’ but we couldn't work out what was going on. We were testing all these things, checking in different browsers, and checking on different devices. Eventually, I went into their office, and they were using an extremely old version of Internet Explorer (not even Edge), and they had the magnification set to 150%.

They were the edgiest of edge cases. None of our current or target customers were using Internet Explorer, and certainly none of them were using it zoomed in at 150%. It looked a bit clunky on their device, but that wasn’t something we needed to test.

With analytics and things these days, you can see what devices your customers are using. Are they mostly on phones? Are they mostly on desktops? For most companies and businesses these days, the majority of your audience is going to be on mobile, but not all. I work with lots of B2B companies that talk to people like HR teams who are based in an office, and they're all on desktops. That's where all their business comes from, and that's where all their conversions come from.

You need to be testing and looking at things the way your customers would look at them.”

Do you need to test to identify where issues are coming from, or can you just use analytics for that and then test the problems you find?

“There are things that would be a problem across multiple devices or browsers. If your conversion rate is at 3% across the board, but everything seems to be fine, that's where you need to do some testing. You might be able to up your conversion rate by another 2-3% if you could fix one issue.

Maybe there's some friction in the checkout process, there’s a button that doesn't quite work, or there's a process that's a bit clunky. Those are the kinds of things where you still have a conversion rate – things still work, and your analytics show that it’s working – but you could make it better.

There's a difference between testing for functionality and whether or not it works, QA testing, and A/B or split testing. That’s where you’re testing things like the red button versus the blue button. You’re not testing whether it works, you’re testing whether you can make it better. What if you tweak this? What if you change this?”

How do you determine which tests to try initially?

“You can look at competitors. Are your competitors doing things that you're not doing? That doesn't necessarily mean that they're doing it better; they could be doing it worse. You can test those kinds of things.

SEO and digital marketing have always moved on those expectations of what people want and expect to happen. Those user patterns and behaviours change, and those moods and demographics shift. The way my kids browse the internet is very different from the way my parents browse the internet. People’s expectations and requirements change over time, so you can start to build up a good idea of what to test from there.

Curiosity is a good thing. You might think, ‘I wonder if this might work.’ If it's not a completely suicidal idea and you can test it in a small way, then test it. Find the data to prove it.

Clients will always say to me, ‘What should we do about this?’ and I have a lot of very strong opinions, but I prefer to back them up with data. My instinctive response is, ‘I think we should do this, but we should test it.’ I’ve assumed it’s the right thing to do (because I’m a gnarled digital marketing veteran), but I might be wrong. I quite often am.”

How important is direct testing with customers, seeing what they're doing, and getting feedback in real time?

“I've been lucky enough to work with some really good QA teams who have helped me with that.

Back in the day, I thought that testing would just be logging onto my browser, clicking the buttons, and seeing if it worked. When you work with proper QA teams, you have proper processes written down. You clear your browser history and cookies so that you're starting in the way that a brand-new user would, you're going to do it on a specific device, you'll scroll this much, you'll click on this button, etc. It’s meticulously broken down, step-by-step.

As an old school digital marketer, I was rolling my eyes, thinking, ‘Do we really need that? How detailed do you have to be?’ However, they would often identify issues that I didn’t notice. Because you are so familiar with your website, you assume that your customers will be able to find things as easily as you can.

If you don't have a dedicated QA team, one of the best ways to test stuff is to go to a café, find somebody who's sitting on their computer, and say ‘I'll pay for your coffee, would you mind just going to this website and testing it for me?’ You just say, ‘I want you to go to this website and try to buy a green iPhone.’ Then, you shut up and just watch. It is enlightening.

Inside, you'll be screaming because, to you, it's so obvious. You built this beautiful user flow, but as soon as you put real users into it, they never go the way you expect them to. They go to the search bar, or they click on things that pop up on the homepage. They're easily distracted, and they can't find the thing that you think is blindingly obvious.

You don't have to have a huge QA team. Just go down to a coffee shop, be prepared to put a few quid into a few cups of coffee, and ask people really basic questions: ‘Can you find the red ducks on this website?’”

How relevant are these testing tasks to SEO, and what measurable impact do they have on SEO success?

“The longer I'm in SEO, the more I find myself doing things that aren't strictly SEO. Back in the day, I used to think my job was to sprinkle as much traffic onto your website as I possibly could, make you rank number one on Google, and your traffic would increase to three times as much overnight.

However, it's so much more than that. If I can keep your traffic the same, but increase your conversion rate by 10%, that's better. Then, everybody wins. That is really important because, if you're getting the wrong kind of traffic (they're having a terrible experience, they're not finding what they want, and they’re just bouncing off), then those aren't your customers.

If you think that your job is just traffic, you're missing the point. Most of my clients see all these reports about DA rating and rankings, and they want to know what I’m going to do as an SEO. They want to hear me say, ‘I'll get you more traffic,’ but I always push back on that. I’ll say, ‘What if I kept the traffic the same, but your revenue went up? Would you be happy then?’”

Will testing also help you improve your visibility on AI search engines?

“Everything around AI visibility is changing so much and so quickly at the moment that anything I'm about to say will probably be out of date by the time you read it.

You can test that. You can measure your visibility in some of these tools and see the journey that people are going on. A lot of people are still a bit worried about AI and these kinds of tools replacing Google search. From what I've read and seen, AI is where people are doing their first exploration, and then they come back to traditional search engines afterwards.

They're having a conversation with an AI model, and they're asking questions, and then the models are recommending a shortlist of things or helping them find a way through things. Then, people want to go to a traditional search engine. They’ve found out about something they didn't know about before, they've gained a bit more knowledge, and they're coming back into the search engine.

It’s taking over the top-of-funnel a little bit. That's where things are getting shaved off. Therefore, you want your content to match where they're joining the journey a bit later. They’ve found out about this thing that they didn't know anything about before, so what do they want to know next?

That’s where things like query fan-out and AlsoAsked.com (the tool that Mark Williams-Cook built), which is great for this kind of stuff. It's about finding those things that your customers are going to ask next. That's probably less about testing and more about positioning and researching around your content.”

Why should you test your check-out processes to remove friction, and how can that improve conversions?

“I was recently working with a company that does wedding and event catering. They work at lots of different venues, and their venue list has grown organically over the years.

They have this query form where you can book a call to ask about catering for an event, and they had check boxes for the locations you could choose, but there were about 30 of them, so you had to scroll a very long way to get to the submit button.

We simply changed it to a drop-down menu. It went from taking up a huge part of the screen to a single box. Then, you could see the submit button immediately, and the process was much more straightforward on mobile. The conversion rate shot up by 30-40%, just from that one change.

Try to make it as straightforward as you can for people to give you money. It sounds really simple, but it gets unnecessarily complicated. The number of times you see forms where people are asking the default questions for your name, your email address, your date of birth, etc., but you’re just buying a pair of socks. Why do you need to know how old I am?

I know you want that for your marketing, but shift that to later. Let me give you the money now so that I can buy the socks. If you want to demographically target me later, put that in your email follow-up stuff and ask me those questions then. Don't make it difficult for me to give you money. That’s the kind of friction you want to take out.”

If you sell seasonal products, how do you know when to shift your menus and navigation to ensure the most relevant products are uncovered as easily as possible?

“Some of that just comes from experience in the business.

I work with companies that have a big push around Christmas, and lots of them even start that push in September/October. As a parent, I'm desperately trying not to think about Christmas because I'm trying to put it off as long as possible. In America, Christmas starts as soon as the pumpkins from Halloween go away. For a business, it's way before that. People are already planning their Christmas campaigns at the beginning of autumn.

If your clients have already started to get queries about Christmas, you've missed the boat. You should have been pushing earlier. If they’re getting queries from the people who had to jump through all the hurdles and get past all the summer stuff to find Christmas, and they're asking about it, you should have been pushing the Christmas stuff a bit earlier.

Some of that is just experience, but you can also do things like look at your internal search data. If you're not tracking the internal search data on your website, you really should be. If people are going to the search function on your site and looking for ‘woolly jumpers’ when it's 90 degrees outside, they're planning for their winter wardrobe, and you need to be there when they’re ready for it.

It's a balancing act. You want to be selling the stuff that everybody wants right now, but be ready for when they switch over – and that switch can happen really quickly, particularly in the UK, where the weather changes at the drop of a hat.

There's never a hard and fast rule for the cutoff date. Some of it is experience; some of it is looking at those clues in the data. Google Trends is always a great one to look at. When does that switch take place, where people stop looking for sunglasses and start looking for woolly gloves?”

Andrew, what's the key takeaway from the tip you shared today?

“Do some testing. You don't have to test everything, and you don't have to test all the time, but do some testing.

You won't regret it, and it could save your business.”

Andrew Cock-Starkey is Founder and SEO Consultant at Optimisey. Find out more over at Optimisey.com.