How to Test SEO Strategies – 7 Ways + Bonus!
“SEO is dead.” That’s a frequent statement from people whose SEO strategies are indeed dead. For your SEO to be effective, you need to understand what is actually working. You can two that in one of two ways.
- Listen to experts regurgitate the same nonsense that worked 3 years ago.
- Identify and test a strategy.
Based on the title, we’re going to look at how to test SEO strategies and methods.
If you read and apply these tips, you will be able to effectively test and implement a SEO strategy that is based on real world results.
That means you can offer real SEO services that you’re confident will work, because you’ve already tested each method and seen them work!
Clients often come to me with a specific need. They want to rank for a particular keyword or have a problem. My first step is to diagnose what’s wrong with their rankings. Here are a couple issues that I’ll frequently encounter.
- Spammy anchor text
- Over-abundance on exact match anchors
- On-site issues
- Landing page without enough content
The client in the screenshot came to me after they received $7 million in Series B funding. In 12 months we’ve tripled their organic traffic and they recently completed Series C funding for $40 million.
From the director of marketing, “Brandon and his team were the driving force behind our valuation that elicited a $40 million C round. Without Brandon I can’t imagine where we would have been.”
These results were achievable because I knew what would work. I knew what would drive SEO rankings because I had already tested everything before I ever met the client.
Here are 7 ways to effectively test your SEO strategies and a bonus at the end.
1. Embrace the scientific method.
Hypothesis > Test > Analyze
Your hypothesis is simple, this test will push up a page in the search results. If you don’t start with that hypothesis, you’re already at a disadvantage.
How do I find a hypothesis? If your SEO knowledge isn’t deep enough to already have your own ideas about what will cause rankings, find someone who is successful at SEO. This is NOT a guru on Moz or SEJ. Those guys are great if your client is a billion dollar brand. Most clients want results, not an agency telling them how to write ad copy.
This expert is NEVER selling a course. Someone selling you a course for $19.95 per month has already admitted defeat. He’s willing to sell secrets that are worth millions for the price of a cheap dinner.
Forums are a great place to find people who know what they’re talking about. Some blogs can be great as well. I remember reading Eli Aloisi at BlueHatSEO in 2005 (Read original posts here). I would literally check his site every day to see if had a new post. I emailed him telling him to post more often. I stalked followed him on forums. I couldn’t get enough. Eli’s posts not only gave great tips, but his approach expanded my thinking which was limited by people teaching keyword stuffing as the key to success. I attribute much of my outside-of-the-box thinking to what Eli. He was excelling when I was still learning.
2. Document everything.
Documenting all of the data is important because it gives you a guide to understanding what is working and what is not working.
My documentation process is simple. I use a spreadsheet to keep track of what was done, the date it started, cost, vendor, expectation, current ranking, and any other relevant data. I also add that keyword and URL to my rank tracker of choice (SerpBook and SERPWoo). I keep all of my tests in one group and one spreadsheet. Testing needn’t be difficult, just keep track of everything.
If you just try to remember what you’ve done, you may forget something and you will have a successful test that you can’t replicate.
3. Have a control.
Having a control group means that you expect nothing to happen with that control. This ensures that your test is not a result of a big Google shakeup.
When I’m testing for SEO I will usually grab a few keywords that are low-medium competition. If I can push up a random page then I should be able to push up a page of my choosing. This is not always the case, but is a good place to start.
When choosing a keyword or two, and subsequently testing a theory, make sure you document everything in the top couple pages. SERPWoo does this natively and makes it worth the cost of a subscription.
When you’ve selected a keyword, you should only try to push up one or two URLs in that search term. I like to keep the volatility low for any given term.
Testing for reputation management is different and depends on your type of client. I work with a lot of individuals and small to medium sized businesses (under 50 employees). My tests reflect what my typical client looks like.
When I set up an ORM test I will search for a random individual’s name and see what the results look like. I often have to search for 20 names before I find one that I like. I want to see similarities between the test and my normal clients. If I don’t see any similarities then I search again until I find URLs in the top 20 that are familiar.
I still use the same method of having a control and pushing up a URL or two.
Having a control is mandatory, without a control, you could see results that weren’t of your doing. This will leave you with the belief that you influenced those results and you’ll waste countless hours trying to replicate success that wasn’t yours to claim.
4. Use non-clients for testing.
My preferred way of testing is a set of random results. This ensures that I’m not biasing the search results with frequent searching and clicking, building out other properties, and other activities that could influence the test.
A non-client based test gives you the ability to setup a test and then come back later to see the results. If you do your testing this way, you will spend extra money pushing up a result that doesn’t matter, that’s the drawback to using non-client names and keywords for testing.
5. Use clients for testing.
Using an existing client for testing is great if you’re sure the method is safe. I hope you wouldn’t use a client’s business as a test-bed if you’re unsure of the safety of the type of link or activity that you’re doing.
I will often use an existing client for reputation management work because we’ll create great content that will find it’s way to the second page. Then I can take that content, set up a test and see if it moves up in the SERPs.
Once again, make sure you are certain that your test won’t harm your client.
6. Test it again.
Now you’ve had a test and it was successful. Since there are so many different things that can influence your test, you should always repeat a successful test. If you can get similar results twice, I consider that a success. It’s at this point, and only at this point, that I feel comfortable offering the results to my clients.
I always treat the budget of a client with the utmost respect. If I wouldn’t utilize a strategy on my own site, I won’t use it on theirs. That’s why testing SEO methods is so important. If a SEO test is overly complicated or extremely expensive, it’s relevant for a small percentage of clients. I always try to base my tests on broad factors and then get narrower if necessary.
Having a second successful test ensures that your theory was correct and not an anomaly.
7. Consider other factors.
There are many factors that contribute to rankings. What we want to do with each ranking factor is isolate and eliminate that from influencing the results.
Here are a few example factors that can contribute to rankings that you should consider when testing new strategies.
- Social media numbers – Do higher numbers mean pages rank higher?
- On-page metrics – Does great on-site mean a page will rank higher?
- Word count – Does a 1000 word article rank better than a 200 word article?
- Domain metrics – Will a domain with DA 60 rank better than a brand new domain?
- Images – Will images help a page rank?
Those are just a couple of things that I would want to consider when trying to create a link building test. If one or two of those factors are skewed, your results could be skewed. Controlling the variables will help give you true results.
Without considering factors outside of the SEO strategy you implement, you are not going to have true results.
8. *BONUS* Example Tests!
As I’ve mentioned, I do a lot of testing. Here are a couple tests. You’ll need to replicate these tests to get the results, but hopefully, like Eli did for me over 10 years ago, this will change your thinking.
Hypothesis 1 – Social media shares can increase page rankings. Can social media shares along drive rankings?
Hypothesis 2 – Will spammy links to internal page penalize the entire domain? I used to build hundreds of thousands of links to press releases and other internal pages, why didn’t those pages get penalized? Does domain age play a role?
Hypothesis 3 – Will a page rank with good on-site keywords if you don’t build any anchor rich keyword links?
Hypothesis 4 – Google bombing can be a quick way to get churn and burn results.
Hypothesis 5 – How much is too much for word count on a page? Is 5000 better than 2000? 10,000?
Hypothesis 6 – Guest posting enough to drive rankings. I love guest posting, do you need social media or other links?
Testing can either prove or disprove your hypothesis. Don’t force the results one way or the other. Just because you really want something to work doesn’t mean that it will actually work.
I typically find success in less than 10% of my testing. Don’t give up if you don’t find a successful test, just keep theorizing until you land on something successful!
If you have any questions, leave me a comment and I’ll answer!
Latest posts by Brandon Hopkins (see all)
- How to Test SEO Strategies – 7 Ways + Bonus! - October 3, 2017
- When Money Is No Object. Online Reputation Management On Steroids! - July 18, 2016