The author’s views are entirely his or hers (except for the unlikely event of hypnosis) and may not always reflect Moza’s views.
SEO testing expert Emily Potter joins us again to end this season on the whiteboard Friday! Today, she takes you through some tests that created unexpected results for her SearchPilot team, and what those results mean for an SEO strategy.
Enjoy and follow next season of Friday’s Whiteboard episodes, which is expected later this summer!
Hello, Moz fans. Not Emily Potter. I am the head of the customer success department at SearchPilot. If you haven’t heard of us yet, we’re an A / B SEO testing platform. We conduct extensive SEO tests on company websites.
So these are sites in industries like travel, e-commerce, or posting sites, anything that has a lot of traffic and a lot of suggestion pages. I am here today to share with you five of our most surprising test results we conducted at SearchPilot. Part of a successful SEO testing program is getting used to being very surprised, and that’s because something you really thought would work won’t work in the end, an SEO best practice test that eventually it actually harms your organic traffic, or something like that. which you tried just because you could, that’s the winner in the end.
All our customers and we at SearchPilot are always surprised, but that’s why testing is so important. If you’re a big business website, testing is what gives you a competitive edge. It helps you find those things that your competitors might not, especially if they don’t test, and it helps you prevent changes that would harm your organic traffic, which you might if you couldn’t try. him.
Sometimes, though, it’s as simple as offering you a business case to get support, that you need to introduce something on your site that you’ll do anyway, but you may not have had a buyout from other stakeholders. To learn more about how we run tests at SearchPilot and how to monitor things like seasonality, algorithm updates and all that, visit our website and there are many resources.
1. Use ‘data-nosnippet’ to force Google to display custom meta descriptions
Okay, the first test I’m going to share with you today is a customer who used the data-nosnippet attribute to force Google to respect its meta descriptions. As you probably know, Google now overwrites meta descriptions and title tags, which can be really frustrating. In the case of meta descriptions, it sometimes yields text that is strung together with ellipses, not very readable, lacks good grammar, and a lot of SEOs find this frustrating.
In order for Google to display our meta descriptions instead, our customer added the data-nosnippet attribute to the body tag. The data-nosnippet attribute is that it tells robots like Googlebot that I don’t want you to record any of this content. So by putting it on the body tag, we actually forced Google to use what was in the head, that’s the meta description.
As you can see, this was a negative. This resulted in a 3% loss of organic traffic. When it comes to SEO tests, it’s actually a pretty small loss, but it’s still not something you’d want to introduce and why would you lose traffic at all if you know it’s going to hurt something. So in this case, it turns out that Google may actually be better at writing meta descriptions than we are.
So maybe meta descriptions aren’t something you should spend as much time on as SEO. The meta descriptions we find in SearchPilot make it very difficult to find something that is positive, and we’ve often run this a few times in different industries and different sites, in fact Google is better at writing than we are.
So maybe let the robots do the work.
2. Increasing the number of related articles
Our second test was on an e-commerce site. This was in the blog section of their website, where they had blog content related to their products. At the bottom of each article were two related links to the articles. In this test, we increased this from two to four.
At this time, internal linking experiments are complex because we influence both the pages to which we add links and the pages that receive links. So we need to make sure we have control over both. Once again, to learn more about how we do this, you can check out our website or contact me later. Now, in this case, it was an 11% increase in organic traffic, which may not seem surprising because the connections we know work.
Why then did I include this in the five surprising test results? I have this included because it was actually for donor sites. By that I mean the page where we added the links. In fact, we did not notice any perceptible impact on organic traffic on the sites receiving the links. This was really surprising and it shows that links don’t just transmit the property of links.
In fact, they help robots better understand your site. They can be a way to connect different pieces of content together. So we could actually benefit both sides. This is also why it is so important to perform a controlled experiment if you are running internal connectivity tests. First, if you were just measuring the impact on the pages you received, you wouldn’t find them at all.
Or often, not often, but sometimes we at SearchPilot have actually seen this as positive for one group and negative for another. Therefore, it is very important to determine the net effect.
3. Localization of product content on the US e-commerce website
Our third test, which I will share with you today, is when we localize content on product pages for an e-commerce site in the U.S.
It was turning things like pants into pants. It was a website that was originally based in the UK. They appeared in the US market and kept content in the UK when they did. So we wanted to find out what would happen if we updated this and adapted it to the market we are in. This was a 24% increase in organic traffic.
Now, for me, it was a surprising size, how much of a difference it made. But I don’t think it’s surprising when you think about it. If pants don’t get a lot of searches per month in the U.S., but pants, then I guess you’d expect the localization of that content to improve your organic traffic.
So there were places where this content existed, like meta title, meta description, H1 and things like that. If nothing else, this is just a nice indication that sometimes the usual SEO recommendations actually work, and it was a great example of how they could come up with a business case to encourage their developers to make a change they might not be able to. convincing them was very important to the contrary.
4. Add prices to title tags
Test number four, adding prices to titles. Again an e-commerce site. You would think a best practice recommendation has a price in the title. This is something that users want to see. But, as you can see here, this was actually negative and the 15% drop in organic traffic was so considerable.
Here, however, the context is important. One of our hypotheses was that our competitors in the SERP did not use prices in the title code, but had price snippets derived from structured codes. So maybe users just didn’t respond well when they saw something different from what other competitors in the SERP had.
It’s also possible that our prices weren’t as competitive, so it didn’t help us to put them first and center in our title tag because it was clear that some of the other competitors in the search we had had better prices. In any case, we have not introduced this change. But this is an important lesson, as no two sites are the same.
At SearchPilot, we performed this test many times and saw positive, negative and unconvincing results. So there is no one-size-fits-all approach to SEO and nothing is the absolute truth and even something as simple as adding price to your title tags.
5. Add alternate keyword-rich text
The last test I will share with you today was when we added alternate text rich in keywords to the images on the product page. As you can see, this did not have a perceptible effect, which is a common SEO recommendation. This is a common thing that occurs with things like technical audits or the great results you give to a potential new customer.
Here, we found that it didn’t actually have much of an impact. This suggests that the replacement text does not have a significant impact on the ranking. However, there are other really important reasons why we should implement alternative text, and we decided to introduce it anyway. Number one is accessibility.
Replacement text helps make your images more accessible to those who may not see them, and helps bots to be able to explain what’s on the page. However, if something is simply not depicted, then it helps people to still know what they are watching. So alternate text, while it may not be a big winner for SEO traffic, is still an important implementation and not something we want to forget.
That’s all I have to share with you today. If you found this interesting and want to get more such case studies, you can sign up for our case study email list, which we issue every two weeks with a case study email that includes another case study. You can also find everything we have done in the past on our website. So, even if you can’t run SEO tests or if you’re not a big business website, you can still use the insights we have to help you create some business cases in your business.
Thank you for accepting me. Goodbye, Moz.