We hear about SEO testing a lot, and I’ve spoken about it a great deal in my career, trying to educate on the best practices to make data-informed decisions. One of those best practices is known as “test and learn,” testing ideas and hypotheses on a small section of a site to see how it performs. While I often focus on telling people how I do “test and learns” for big websites, one question from the audience that stuck out at one of my most recent presentations is why do I even bother doing it?
Table of contents
I realised that they did have a point. The reasons might be very obvious to me due to the nature of my role and the clients I work with. But, one thing I failed to recognise was that my specific experience was probably different than some of the audience listening in. Testing SEO initiatives presents many benefits that are not dependent on the company or industry you work for, however. I am going to expand on this shortly.
In general, I could summarise the answer in one sentence: I test to avoid the stuff of nightmares in my day-to-day job. Something like a slow but inexorable decline in traffic after a technical change, like the one below, for example.
(A screenshot from Causal Impact displaying a loss in traffic over a period of 30 days)
The best part about it? This could have been prevented by testing the change in advance. We’re going to get back to it in a second.
Three reasons I run SEO tests (and you should, too)
Testing can mean different things in different arenas, but for SEO, it refers to the process of investigating how a change can affect a metric of interest (for example clicks) as compared to an untreated variant, or to previous performance.
So, why do I bother testing? For me, it all comes down to three reasons (plus a bonus one).
1. To inform the SEO strategy and figure out what works best for your audience
As SEO professionals, we are lucky to have a network that shares new strategies and best practices all the time. I am someone who gets excited by shiny things and new tricks, but I have also learned that I need to have some tangible results at hand to validate strategy changes.
Additionally, one of the things that I hear (and see) a lot is ‘Competitor X does this, so we should do it, too’. I had this conversation many times and, while competitive research is a necessary part of informing the strategy across all channels, it cannot be the sole driver of SEO. What works for them might not work for your brand when it comes to getting your ideal audience to visit and convert.
One example?
Even if you’re competing with Apple, you cannot expect their audience to make the same decisions on your website that they would make on Apple’s. That is, unless you carry the very same brand authority as Apple and deliver on the expectations that come with it. Taking the time to figure out what streamlines your audience’s decision-making (be it the navigation, the elements on page or how products are displayed on your website) will allow you to laser-focus your efforts beyond SEO, and take out most of the guesswork. We normally test for UX, and we now know how intertwined user satisfaction is with SEO, so why treat them differently?
2. To show stakeholders the importance of SEO
If you’re lucky, SEO is a massive part of your organisation; if not, you might have to fight your battles with data points and forecasts. Small changes that lead to tangible wins can buy you more space on the backlog for bigger projects.
I recently went through this exact situation. I am not particularly fond of the site architecture of a website I work on, and I’ve been very vocal about the changes I would need to make in order to make it SEO-friendly. I designed a new structure, polished it according to best practices and business priorities and discussed it with the tech team. The team, seeing the amount of effort required to do such a change, politely told me it was never going to happen. And I get it: it’s like I wanted to be invited to their dinner party by busting down their door. Sometimes you just want to knock instead.
So I suggested some small changes within the structure already available, and tested against a control group of pages. The best result wasn’t even the higher clicks of my test group against the control, it was the fact that showing them the exact impact proved my concept worked, and it bought me a seat at the table at the next prioritisation meeting.
In my experience, showing the impact of the work you do, it is likely that more people in the organization will get excited about the mission and get on board. This is especially true if you share results regularly and transparently with the wider team, In turn, testing can get you more resources for bigger projects, too.
3. As a risk mitigation strategy
You know that dreadful downward curve at the beginning of the article? To think it was only a ‘small’ change that led to all that traffic lost…and it was a migration. Of course it was.
Every SEO reading this will know that migrations are a massive pain: there are always moving pieces and even when you have everything in place to make it work (benchmarking, a tiered approach, early involvement of the SEO department) there is always the real possibility that something will go wrong. And I know where everyone’s mind is going: the dreaded redirects.
However, what happened here is something even simpler: the migration enabled two separate subfolder routes to access the same selection of products, on two separate pages (to better support UX). Naturally, they were aware of the risks for SEO, so they made only the preferred version indexable, to avoid duplication. The thing that went wrong? They linked the non-indexable version of the pages in the navigation, as it was a user flow that ‘seemed more natural’.
Bonus: to settle a tech debate
But this seems rather obvious, right? Should I really test something that I know for sure works for SEO and it is a known best practice? Well, if you’re lucky, no – the business should just trust you on that. But even when they do, there might still be some challenges.
- Different departments might have different KPIs and approaches
Particularly on enterprise websites, the tiniest change will have a number of people and teams that are stakeholders on the same pages you work on. While the ultimate goal might be the same (a transaction, a sign-up, or any other sort of conversion), their KPIs and backlog will look different from yours. Different teams will also have different ways to approach a solution.Testing ensures that the best one for users and search engines can be implemented with confidence.
- Data speaks louder than words
Even when we have our arsenal of guidelines, competitor research and our own years of expertise to back our strategy up, sometimes data is what speaks louder than all, especially if it’s a David VS Goliath situation and the SEO team tends not to get much space on the prioritization calendar for a reason or another.
Also, if you’re like me and confrontation works you up.
Final thoughts
I do love testing and that’s the reason why I wrote an entire article trying to convince you all to run more tests. Not testing for the sake of testing, of course: every test should end in finding actions to deliver (or avoid) to get the business more successful.
However, while testing provides you the confidence to spearhead your strategy and changes, don’t forget that SEO is part of the bigger mission of leading a user to become a customer. Even when best practices and results suggest one change that might benefit SEO, we might need to take a few losses and do some compromises, particularly if it clashes with UX and CRO tests.
If you’re not convinced yet on why you should run SEO tests, then it means you are lucky and do not face any of the challenges I wrote about. I salute (and envy) you.
