Modernise your understanding of analytics

Marco says: “Get familiar with basic analytics. SEO is now becoming more dependent on data.

It always has been, but with the most recent technologies and changes from Google, plus the advancements within companies that are getting more familiar with IT, it’s possible to do much more compared to, say, three years ago.”

What analytics knowledge do SEOs need?

“Understanding GSC and GA4 data is the bare minimum, because you must report on the results of SEO and how organic traffic is contributing to the business.

Also, custom events. In GA4, you can create custom events, or custom dimensions, which allow you to enrich your analysis. For example, if you have to label or cluster some pages on your website, you can use GA4 and create some custom dimensions. Or, if you want to track specific actions on your website, you can use Tag Manager in GA4 to track these specific actions and then use them in your reports.

This is important. What you get out of the box is usually very basic and poor. Most companies want something more custom. Getting more familiar with how these platforms work – or even with the more abstract side of data like strategy, processes, and even communication – can get you far.”

Are SEOs responsible for learning everything about GA4?

“You shouldn’t go into too much depth. You shouldn’t be a master of GA4. Even I don’t use all of it. Even if you are specialised in GA4, you will never use most things (depending on your clients and the job you are doing). You should just find an equilibrium with what is useful.

I recommend linking BigQuery with GA4, especially if you work for a medium or large-sized company. BigQuery is a data warehouse solution from Google, where you can get raw data from GA4 that is not affected by the usual limits. In standard GA4, there are some limits – supposedly to avoid overuse but I think it’s actually to make you pay for the paid solution or use BigQuery.

For example, thresholding, sampling, and a lot of negative effects of your data – even cardinality, where you have a dimension with a lot of unique values. Standard reports in GA4, or even explorations, are not as complete as BigQuery.

It is something I recommend learning, although it is hard at the beginning. It’s not user-friendly at all. It requires you to know some SQL and understand a bit about how the cloud works, but it’s not too hard. It should take about one month of study. Most of it you can do yourself. There are a lot of tools online, even AI, that can help you write basic code.

This is useful to have, instead of learning everything about GA4. Medium/large companies should just use BigQuery because it’s better in terms of data. They have a free tier that gives customers up to 10GB of storage and 1TB of queries per month, as well as other resources. Most companies never go beyond that but, even if you do, storage is not that expensive.

Pulling data and doing queries can be expensive, depending on how you do them, but that is mostly for the engineers to guide you.

Also, with GA4, you can only go so far back in time with your data. If you automatically import it into BigQuery, then that’s not an issue. In GA4 it gives you 2 months as standard for explorations – which is kind of a scam because, if you’re not technical and you don’t set up your data retention to the maximum 14 months, you only get the last 2 months, which is terrible.

Even if you do set it up for 14 months, you can’t compare year over year. It’s easier to just start loading your data into BigQuery. It’s not retroactive but it’s better than nothing.”

Should SEOs be doing this themselves, or should they get a GA4 professional to do it for them?

“I think the difficulty of GA4 is overrated. A lot of the stuff I’ve mentioned can easily be done by yourself. It’s not that hard. To be honest, in SEO, the priority goes to GSC over GA4 anyway, because GSC is the first SEO tool.

GSC also offers a bulk export, so you can put that data into BigQuery as well. It’s much easier than GA4 in terms of actual data. When you see the data in BigQuery, it’s easier to understand. Start from there. It doesn’t take much, to be honest.

The issue is that GSC is removing a lot of anonymised data – the famous anonymised queries. BigQuery gives you this anonymised data, so you get many more rows compared to what you get with the API. You can get 200% more, depending on how big your website is.

These are technicalities, though, and people stick to technicalities too much, because businesses don’t care. Of course, if you have a discrepancy, you have to communicate it because it’s important. In terms of numbers, though, I don’t stick too much to this trivia. Just Google it.

The most important thing is understanding how to put all of this into a process or system, and how to communicate it effectively. Otherwise, it’s all fluff. It’s technical knowledge: interesting facts without any real-world application. That’s why it’s important to understand the processes, and how a business can actually use this data.”

What data should SEOs be looking at regularly and how should that impact the way they’re carrying out their SEO activities?

“The data you should look at is largely reliant on the model of the website. If you have an e-commerce site, traffic is less important. It’s still important, but you care about the sales, conversion rate, average order value, and other metrics.

If you are a publisher, and you’re monetising with ads, you want a lot of traffic from the US because it pays more. You are not selling any products, so you cannot track sales.

Therefore, the metrics are largely reliant on what you’re doing. You should always track the financial metrics because any marketing channel should always be tied to the money-making aspect of a business.

You should also track content decay, to see if the content on your website is losing traction. If you are Amazon, there is no content. It’s not the New York Times. In that case, you could even track which products usually get sold together, if your CMS is not offering that by default.”

How would you track content decay?

“If you are only doing it for SEO, take clicks, because clicks come straight from Search Console. It’s all the organic traffic from Google (not all the organic traffic, just the organic traffic from Google).

Otherwise, you take your favourite GA4 traffic metric: users, sessions, views – anything goes. You can filter to select organic traffic or maybe all channels. It depends on your goal, and the question you’re trying to answer.

Every page has a trend over time. Say that it’s been getting clicks for 6 months. You can create an imaginary line for every page and then, based on the slope of that line, you can tell if there is decay or not. According to middle school/high school mathematics, this slope measures growth or decay, or if there is nothing happening. If it is close to 0, nothing is going on. With proper filtering, this value can help you understand whether something is decaying or growing over time. It’s a quick and efficient method.

You can also just use good old pivot tables, with some conditional formatting, to see how each page – or even each cluster – is changing over time, in terms of performance. If you are good with conditional formatting, some cells will be highlighted. It’s a very simple method, but it works because it’s visual. You can see where the biggest variations are or where you had the most clicks, users, or any metric you choose.”

Is it possible to automate this to send you alerts?

“There are many ways to do that, depending on the stack we’re working with. You can even do it in Google Cloud. You can do your calculation and then, if you have some values that exceed the threshold, you can get an email via automation, or decide any action you want to do.

You can also do something more complex with your database/spreadsheet. With the help of Make or Zapier, which are tools to automate things, you can send this data to other platforms – even to Google Sheets. Automation is infinite. You can automate whatever you want.

The important thing is understanding the process from a business standpoint, and then you can think about the technical stuff. Think about the business first. Most of this stuff is easy now because we have the tools. You don’t need to create it yourself; you can look for solutions.

Even without them, it’s quite simple. You can read tutorials online. Where people struggle is not the operational things, it’s the strategic side and the business side. Many people don’t want to study the business side, for whatever reason.”

How do you help a business decide what the key metrics are?

“I usually start from their goals or what they want to achieve. Some people start from the metrics first, but you should start from what they want to achieve and where they want to be. Then, see where it overlaps with your strategy, and you can also decide your KPIs and your most important metrics.

I don’t want to overcomplicate it, though, because I hate overcomplicating simple stuff. If you are a publisher and you just have subscriptions or ads, it’s very simple. You make money if people read your content. The more people read the more money you make. It’s very simple. There is nothing to understand.

If you’re in e-commerce, it can get a little bit more complex. Yes, you can make more sales and make more money, but not all products have the same markup. Not all products are equal. Sometimes it’s better to sell a lower quantity of products with a higher margin. It’s better to sell one Ferrari or Lamborghini compared to 1,000 glue sticks.

Ask what their goals are. Try to understand what they want to achieve in practice, because sometimes the goals are very muddy, or very generic. For example, ‘We want to be the best in the automotive industry’ is not a tangible goal. It’s not very realistic and it’s not measurable. However, if your goal is to increase your sales by 20%, your KPI is sales. You can gauge your progress by checking if your sales are actually increasing by that much.

Sometimes it’s a little bit harder, but I try to stick to the basic common-sense metrics, which are usually financial. People usually talk about revenue and profit. Then, you can also have conversion rates, or CTR on a page – how many people are clicking this link, clicking this CTA, etc.

In some cases, if you are talking about automation, the most important thing to consider is how much money you’re saving. If you have a process that takes 100 hours every month, that’s 1,200 hours per year. It’s a lot of hours, and every hour requires people working. That’s wasted productivity because they’re using their productivity on these 1,200 hours.

If you find an automation process that converts this from 1,200 hours to just 12 hours, you are saving a lot of money, even though it’s not revenue. You are not selling anything; you are just saving money. You’re just optimizing your processes, and these people can work on other tasks that are more profitable, or they can increase their productivity overall, which will make you more money.

Automation is not just about the basics of revenue and profit. Selling can also be about your efficiency and improving what you’re doing – the costs you don’t or cannot see.”

If an SEO is struggling for time, what should they stop doing right now so they can spend more time doing what you suggest in 2025?

“Stop focusing on theory, and stop being too siloed. Omnichannel is particularly attractive now. It’s always been better, but now it’s clear that you shouldn’t be doing only SEO. You should be a little bit more of a generalist.

Focus less on theory, patterns, and the theoretical aspects of SEO because, if you check the SERPs, they are terrible. There’s some stuff that doesn’t make any sense. We should be more practical, more realistic, and embrace reality.

I also don’t think you should focus too much on coding or the techie side, because it’s still about business. Focus on how you can use this knowledge (how you can use BigQuery, Looker Studio, or anything to do with data), and make it more relevant to the business in terms of their processes and systems. That’s what it’s all about that.

Many people focus too much on coding. The problem with that is, if I am a company, I don’t want to hire you. You’re not a developer, you’re just an SEO. I would just hire a developer.

These complementary, supporting skills are good because they make you a better SEO, marketer, engineer, etc. They are complementary skills that you are using to reinforce yourself – but the end goal is always the business: creating processes and putting them into practice. When you’ve talked about SEO, then you need to see how you can execute SEO and make something faster compared to other tools.

The process I propose is faster and cheaper, and you’re actually using your assets in your data. It’s more consistent for a company, and it’s more reasonable than telling them that you can code. That’s not very actionable. However, if you tell them that you can create a process and can make use of these technologies to achieve a specific goal, that’s much more realistic.”

Marco Giordano is a Data and Web Analyst, and you can find him over at SEOtistics.com.