Semi-Automated SEO Auditing and Reporting

SEO agencies conduct their audits in a specific way. Some rely on automated reporting through sites like SEMrush or Brightedge, while others have a completely manual process to QA the site. Ultimately, the most effective way to conduct an SEO audit is to combine these methods into a semi-automated process. 

Considering I’m a data scientist, it might seem strange that I don’t advocate for 100% automated SEO audits and reporting. My job depends on my being able to automate tasks and streamline manual processes - but there’s a concept in machine learning known as “human-in-the-loop” that applies to SEO services.

What is Human in the Loop (HITL)?

In Machine Learning, HITL systems involve humans and machines working together to maximize performance. For instance, data labeling is a key component of text classification or image classification that requires manual effort. In these examples, humans are using their best judgment to QA the training data for ML models. Basically, they’re looking for out-of-the-box cases that a machine-learning algorithm might not be able to identify on its own.

So, how does human-in-the-loop relate to SEO?

One of the major SEO tools that companies rely on is SEMrush. Whether it’s keyword research, auditing, backlink competitor analysis, or on-page optimization - SEMrush has some sort of automated process for generating the output you’re looking for with little-to-no manual effort.

Unfortunately, some SEO agencies have taken this too far by simply generating SEMrush reports for their clients without adding anything of their own. By relying on a 100% automated process, SEO analysts aren’t learning how to spot anomalies, or what is known in data science as “edge cases”. 

Examples of SEO Automation Gone Wrong

Let’s say that an SEO agency simply runs the monthly SEO audits and standard reports for their clients in SEMrush. Perhaps there are entry-level analysts in charge of pulling the reports, formatting them, and sending them out to the team. Here’s the problem - even though everything in that report will be accurate in terms of the statistics, without a person to analyze the findings and add a narrative, the report is essentially meaningless. 

Here are a couple of examples:

Overestimating Traffic

 A client’s website saw a 75% increase in traffic month-over-month. Great! You report this to the client and say that your content marketing campaign is working. Unfortunately, that might not be the case. If an automated report shows a change that drastic, it’s the analyst’s job to look into what happened or why it happened.

For instance, if I saw a 75% increase in traffic, first I’d look at the traffic by day to determine if the month’s traffic was higher overall or if a specific day or week drove the increase. If it was a single day that drove the increase, what happened that day? Did you release new content? Was there an event? Did Google experience an algorithm update? These are things that your standard SEMrush wouldn’t pick up, but can make the difference between a descriptive report and a prescriptive report.

False Alarm

Looking at this from another angle, the holidays just passed, and depending on the business, traffic could significantly decrease due to a change in demand and overall search volume. In a recent SEO report I was completing, I found that month-over-month traffic was down 30% but a closer look showed that this was driven solely by the drop during the week of Christmas. Because I was conducting a manual QA of the SEMrush report, I could catch this and conduct a similar analysis that looked at the first two weeks of December compared to November, which demonstrated that traffic was relatively stable.

The Importance of Personalization in SEO

When an SEO agency scales up, it might resort to a more standardized, automated reporting and auditing process. While certain repetitive tasks can be automated successfully, there’s something to be said about the human touch and judgment required to produce insightful reports. Even if a company leverages data science and machine learning for SEO analytics, there is no way to completely take humans out of the loop. Looking for an SEO agency with a combination of technical knowledge and diligence to complete semi-automated projects? Contact Hire a Writer today to find your next SEO team.