When it comes to SEO, there's a lot of advice out there. However, with so much information available, it can be tough to know what will work best for your site. That’s where SEO research and development comes in.
In the podcast:
SEO in 2008 was both easier and harder. Today testing plays a key role in success. [02:16]
Just because Coca-Cola has a billboard doesn’t mean you need one. [05:18]
This works well for so-and-so – maybe you should give it a try. [07:22]
How’s that lead magnet strategy working for you? [08:56]
How you begin your article could impact your ranking. [10:24]
Does region affect SEO at all? [12:22]
This is how you can test without endangering your entire site. [15:20]
If it doesn’t work, can you put everything back? [18:05]
Switching domains was a calculated risk that’s worked for James. Here’s why. [20:01]
When James thought of moving his business onto a new domain, SEO was a huge consideration. He’d been working on his previous domain’s SEO since 2008. So he went to his SEO friend and advisor, Gert Mellak, to know if it was something he could do and ought to consider.
Now on the other side of the process, James and his team are still moving pieces. But with Gert’s help, everything is working.
It has prompted James to think, though, what else could they test in terms of SEO? Because it would be good to know what gains could be made, or to know in advance if challenges were to arise. And Gert has just the expertise to speak to that.
SEO then and now – testing is key!
It just makes Gert a bit nostalgic, he says, to look back to 2008. SEO then was almost the wild, wild, west.
James thinks it was a bit easier.
“SEO has come a long way. Google has come a long way.” – Gert Mellak
They didn’t have as much knowledge, says Gert, so it was still difficult. Knowing what they know today, 2008 would be easy. SEO has come a long way. Google has come a long way. Very often now, people publish content with the goal of being found on Google, which is already the first step.
What Gert also sees is people implementing best practices they’ve picked up over time, stuff they’ve seen on YouTube and elsewhere.
But a lot of information, Gert says, also comes from ongoing testing. And this means that you sometimes need to step out of your comfort zone and do that testing. At his agency, SEO Leverage, they have their own R&D department and testing sites. That’s where they try things they sometimes don’t initially think will have any impact, but that sometimes surprise them.
Testing should be something you do as a matter of course, says Gert. Too many people are content with practices they learned 10 years ago. And yes, it’s probably still good to have a keyword in your title and elsewhere. But don’t limit it to that.
Consider stepping back and thinking, if I was Google, would I value this? Would I assess this page differently if this happened? This is where testing comes in, and where it’s important to have a proven framework you can actually work with and actually track what is happening.
All the big companies do R&D, says James. You need it to stay ahead. He’s curious to know, what has Gert seen as a consultant when he gets people started on testing that yields results?
What works for them might not be for you
It’s interesting, says Gert. The best practices people generally implement are very one-size-fits-all. But there are very different types of sites in very different industries with aspects that are common and not common in between.
So you first need to recognize that what works for a big competitor may not be something you can apply to your own website and get the same results.
James agrees. If Coca-Cola has a billboard, it doesn’t mean you should go out and sponsor one.
Absolutely, says Gert. He has had people come to him wanting to replicate something Amazon does, for their small Shopify site. That’s not how it works – a lot of things need to be taken into account.
Very often, you need to be able to test, in a safe space with minimal risk, something you have seen working before, either on your test site or on Gert’s R&D test sites.
When something is worth trying out
Gert and his team work with roughly 30 to 40 clients at every given moment at SEO Leverage. And this gives them a lot of insight into what can work in different industries.
That’s why Gert can sometimes say to James, Here’s this phrase, this is where you’re ranking, here’s something that I’ve seen work well on another site, and it will probably be good for you guys, too.
James does the same in his coaching. When he recognizes a pattern unfolding for a client, he can recommend what he’s seen working for someone else, and more often than not be rewarded by results.
Consider this lead magnet tweak
Based on the data points Gert and his team have, they can suggest specific experiments to their clients.
One time, for instance, they saw a very low conversion rate on a very high-traffic site. It occurred to them that changing the lead magnet might make a difference. They tested the idea on one URL, and within days opt-ins rose from three to 30 a month. With that result, they could then scale up and apply it to other pages.
Were they aligning search intent with the offer, James asks?
Exactly, says Gert. They were trying to make the lead magnet always be the next logical step for the client who comes to a specific article. Very often people offer a general sort of opt-in – an eBook or a checklist that is one-size-fits-all. Whenever we say one size fits all, probably, Gert says, we should, take a break, take a step back and say, I’m not leveraging everything I can actually do.
So dive in, Gert says, look at your articles that get the most traction. Is there a better way to, maybe, name the lead magnet? Sometimes people better resonate with just a change in title.
How structuring content can have big wins
For another client, Gert and his team were able to achieve an interesting gain in traffic. In this case, they picked a sample set of URLs that were converting well, but not getting enough traffic.
On that sample set, they made a content structure chain. For Google very often – and they’ve tested this time and again – it’s important what comes first and what goes later in an article. This are SEO content best practices many people are not doing.
Gert’s team made a slight change on what happens at the beginning of an article, not only cueing, but really topic-mapping, and got a 22 percent gain within a couple of weeks, on the sample set of URLs. They’ve expanded the sample now, and are waiting on results to see if they can confirm the findings. It looks like they can.
“Nobody wants to scroll down and see if something is being answered at the bottom of the page.” – Gert Mellak
Google is really focused, says Gert, and they give importance to the beginning of the page. Gert thinks it really started with mobile-first development, where people are mostly navigating on their phones. Nobody wants to scroll down to see if something is being answered at the bottom of the page.
You want to at least let people know from the start that it will be part of the answer they’re trying to find, and have them jump there, scroll there or whatever. Also, Google reads the start of a page as a person would read the backside of a book, to determine if it’s worth buying.
Does region make any difference?
In migrating some of his podcasts, James was asked by a platform for a region. He wonders now if region has big impact. How does it relate to location-dependent or independent SEO terms? Does it make any difference if you want to target locations or not?
It’s an interesting one, says Gert. Since the pandemic, they’ve seen people who used to deliver in-person services switch to Zoom.
He specifically suggested to one client to move to Zoom only, because he could just see the SEO potential for it to actually work. This client was only delivering service in Australia in one city, and he thought he could only rank for his service plus city name. Gert and his team essentially then established him in multiple Australian cities.
Now he’s getting people from Brisbane, from Perth, from Sydney, from Melbourne, from everywhere, with specific SEO targeting. And he’s still delivering the service that’s traditionally delivered in-person physically. He’s delivering it in Zoom, and is getting more quality leads than ever. Because the site is good, the branding, everything is great. He just didn’t leverage the location aspect of it.
Speaking to that sort of general versus specific, James is curious, is Google clever enough to rank a single sales page for a course, for multiple keywords? Or do we still need to silo out our sales pages or landing pages to bring people in on a more relevant SEO pathway?
It’s definitely something worth testing for everybody, says Gert. They’re testing it consistently on every single client. Very often, they see people ranking with one really big article, getting a lot of traffic. But when they look a bit closer, we see many of the keywords the article is ranking for are ranking on page five, page eight, page seven.
Why? Because the article is a good fit for a handful of keywords, and a decent enough fit for a hundred or 200 other keywords to still sit on page five. But page five doesn’t help anybody. So now if you zoom in on those keywords, and say, Okay, if my channel article is already a good fit, let’s type in those keywords into Google, then you’ll see Google really wants specific articles for those keywords.
So then you can go in, grab them, put them in onto your website as a specific fit, and make sure to strategically link content together so Google can recognize what you’re doing. Then you actually get a lot more traction, because you just give Google what they really want, not what they kind of like.
Take some of the risk out of your testing
Is there an SEO testing approach Gert would recommend to minimize the potential harm of a failed hypothesis?
You definitely want to make sure that you test something you have seen working very often somewhere, maybe on your existing R&D sites.
At SEO Leverage, they have test sites on a range of industries. So when Gert gets an idea over the weekend, he might drop it to his team and have them test it on one of the sites.
If something works on one of your test sites, or on one of your competitor’s sites, where the competitor is very comparable to you, – and make sure you’re being selective there – then it makes sense to go to your website and say, Okay, now let’s pick something that doesn’t impact my bottom line negatively or too much, and pick a sample only of URLs.
Sometimes it makes sense to pick one URL and see if it does anything to this URL. But you want to make sure you test your hypotheses on a low-risk sample set. Implement it there, and wait at least four weeks, possibly even eight. Because Google needs to crawl and index all those pages, meaning putting the information into the database, and then synchronizing with all the data centers around the world, so Google everywhere knows what’s actually happening. And then certain changes need to mature.
“Always be testing, but have the patience to wait for the results to show what is actually going on.” – Gert Mellak
Sometimes you see even a slight downward trend, until it goes back up much faster. You need to give this a little bit of time. It’s important to always be testing, but have the patience to actually wait for the results to show what is going on.
Then once you see a positive improvement, you can then go back and say, Okay, let’s expand the sample size a little bit. And let’s test this again. And see, can this really, again, work for a larger sample set?
If it is proved to work again, then you can consider just continuously expanding the sample set. Gert is very much against rolling out something sitewide if it’s not necessary. Just slowly expand the sample set, to be safe.
Are some experiments reversible?
James recalls some crazy experiments, short of link-building which he knows could really damage a site if done wrong. But he wants to talk about on-site. He recalls there are some things you can do on site that will drop you out of Google, but then you can reverse what you’ve done and come back up. It’s like when website masters turn their site to noindex, nofollow, and then switch it back on. Are some on-site tests reversible?
Definitely, says Gert. He would probably guess, though, that you can only reverse so many changes on your site to get back credit from Google. You can absolutely, if you run a test on 10 URLs and get worse performance, undo the change and recover.
They do see sometimes, when it comes to Google indexing pages, if you do a change on a large sample set, it can take really, really long to get all those pages indexed. So if you test something across 500 URLs, you might have to wait for a month to have those pages be taken into account. Just make sure you don’t reverse something that hasn’t even been rolling out yet. Google is very cautious of what page they crawl, because crawling costs them money.
Why switching domains has worked for James
How risky was it for James to switch his entire property across to another domain name?
In James’s case, Gert says, there were factors that contributed to its success. So far, it’s seemed to have worked really well. Some of the rankings have even improved, at least everything around revenue share. And it looks like James has switched to a domain that’s even having more searches than the original domain.
And it’s older, says James.
It’s older, says Gert, it has more searches, it probably gets more clicks. People searching for James and seeing his name in the search results are going to click there.
James pretty much has the same website, so just a few components changed on the site, like the homepage and the logo. But everything is still in place. There weren’t too many changes. So a lot of the signals that Google has been picking up over time, they can apply completely to the new website.
James’s team informed Google and said, Look, we switched from this domain to that domain. So Google even knew what was happening. It’s not something they had to figure out. So the risk was really minimized.
Gert had a client inquire about switching domains, and he advised against it. The new name was a made-up name rather than the speaking domain name they had before, which would automatically impact the clickthrough rates.
And then, he says, you wouldn’t have your keywords that are often there. The links can get redirected, but might also have a little bit less context. So you want to be careful, you want to switch for the right reasons. A site migration can really bring you down when the signals change drastically for Google.
Apart from all the technical things that can possibly go wrong, if the redirects and everything are not in place, you want to get someone to guide you. And it’s something SEO Leverage does on a consistent basis. They help people migrate, and do as much as they can in preparation for the process.
It certainly made James and his team feel safer, having an advisor.
If you want help with your own SEO, maybe run some different tests, get some ideas, have an analysis of how your site’s performing, look up Gert at SEOLeverage.com.
Liked the show? Leave us a review on iTunes
Access a treasure trove of coaching, content and community inside the James Schramko membership