Attribution Analysis: 4 Techniques Every Marketer Should Know

Content marketers are under more pressure than ever to prove the value of their work. You need to know which content drives results, what’s worth investing time and money in, and how your efforts contribute to wider company goals. 

That’s where attribution analysis comes in. It helps you pinpoint what is and isn’t working with your content so you can make stronger, smarter decisions. 

This guide covers four key attribution techniques every marketer should know: cross-sectional vs. longitudinal analysis, factorial analysis, A/B testing, and ROI modeling. Each one includes detailed examples and step-by-step tips you can apply using tools like Google Analytics 4 (GA4) and Looker Studio.

1. Cross-sectional vs. longitudinal analysis

Cross-sectional analysis looks at data at a specific point in time, helping you see what’s performing well right now. 

Meanwhile, longitudinal analysis tracks changes in data over a longer period, making it easier to understand how and why things change over time.

Let’s take a look at both below.

Cross-sectional analysis: Assessing data in the moment

Cross-sectional analysis gives you a snapshot of your content’s performance at a specific moment, helping you see what’s performing well right now. This makes it easier to shift resources toward what’s working best currently and ensures you don’t overlook the highest-performing content at any given time.

Cross-sectional analysis is especially useful when:

• Testing (and comparing) new strategies
• Performing A/B tests
• Conducting competitor analysis
• Making decisions without needing historical data

Consider this example:

Cross-sectional analysis also helps you evaluate the financial impact of your content.

For example:

Longitudinal analysis: Tracking changes over time

Longitudinal analysis shows how a metric changes over time. This method helps you see patterns and trends, so you can understand the impact of:

• Industry shifts
• Algorithm updates
• Audience behaviour changes
• Market competition
• Seasonal events or news

Consider this example: 

Company Z, a B2B SaaS brand in the business automation sector, has been steadily growing its traffic month after month. Its content focuses on three main clusters: Reporting, Customer Onboarding, and Employee Development.
Hoping to accelerate growth further, the marketing team looks at attribution data to see which topics are performing best. Here are the results:

Cluster	Page Views	Demo Sign-Ups	Sign-Ups per Page View
Reporting	14,343	62	0.43%
Customer Onboarding	12,211	57	0.46%
Employee Development	8,119	33	0.41%

Of the three, Reporting performs best in terms of age views, while Customer Onboarding performs best in terms of sign-ups per page view. Based on this, the team decides to distribute content production as follows:

Cluster	Content Breakdown
Reporting	35%
Customer Onboarding	55%
Employee Development	10%

When the team views the performance of each cluster month by month, they notice something interesting: Employee Development may be smaller now, but it’s growing rapidly. This trend suggests it might become a top performer soon. Based on this insight, the team updates their content plan accordingly:

Cluster	Original Breakdown	New Breakdown
Reporting	35%	30% (-5%)
Customer Onboarding	55%	40% (-15%)
Employee Development	10%	30% (+10%)

Had the marketing team relied solely on cross-sectional analysis, they may have missed this opportunity—and a potentially significant increase in conversions.

2. Factorial analysis

Factorial analysis helps you understand how two or more variables (like content type and distribution channel) affect performance, either on their own or together.

This method is advantageous when you want to:

• See how different parts of your strategy interact
• Control for variables
• Discover what really drives user behaviour

For example, suppose your attribution report provides these insights:

‍

At first glance, you can see that Reporting content has the most page views and Customer Onboarding content converts best per view.

These are good insights, but the data doesn’t tell you why. How do you know the differences in page views and demo sign-ups are related to the subject matter and not something else, like the author or the call to action (CTA)?

For example, let’s say two authors created this content. If that’s the case, it could be that the differences are due to their differences in writing style and skill.

To understand whether performance is due to the topic or the writer, you compare the same topics by author. If the numbers stay mostly the same regardless of author, then the topic is likely the main factor. If the numbers change a lot between authors, then authorship may be influencing the results. This is called controlling for a variable—you hold one thing steady (e.g., the author) and see how it affects performance.

From here, you can:

• Review editorial quality
• Provide training to the weaker of the two writers
• Adjust author assignments based on strengths

This is known as exploratory factor analysis, a way to reveal the “underlying structure” of your data.

You can also use factorial analysis to understand which combinations of factors offer the best results. 

For example, suppose you’re using two CTAs in your Reporting, Customer Onboarding, and Employee Development content. You can create a table in Looker Studio with both CTAs to see how they work together and affect the data.

This is a 3 x 2 analysis: three content clusters and two modals, creating six combinations.

The data suggests that Reporting content paired with Modal A leads to the most sign-ups per page view—even more than Customer Onboarding, which usually performs better overall.

Based on this, you might switch all Reporting pages to Modal B. You’d also want to investigate why these modals are performing better, such as the colour scheme, page position, and the language used, etc. You could assess all these through A/B testing, which we’ll explore below.

3. A/B testing

A/B testing involves trying out two versions of something to see which one performs better. It’s great for:

• Landing pages
• Product pages
• Email campaigns
• CTAs
• Page titles
• Headers
• Meta descriptions
• Meta titles
• Blog design
• Content length
• Article structure
• And so much more

When combined with attribution data, A/B testing helps you understand not only what’s working, but also why.

Here’s how to run a strong A/B test:

Step One: Formulate a hypothesis

A strong hypothesis will include a specific change and a specific outcome—for example, “Using call-out boxes to highlight product features in blog posts will increase demo sign-ups.”

Your findings will either validate or reject your hypothesis.

Step Two: Change only one thing

With A/B testing, it’s important not to make more than one change at a time. If you test multiple things, it’s hard to know which one caused the result.

Also, the more changes or versions you make, the larger the test and control populations you’ll require and the more time you’ll need to achieve statistical significance.

In other words, simpler A/B tests = faster turnaround times. 

Step Three: Analyse the results

Lastly, check the results against your hypothesis, which you’ll either validate or reject.

Even if your hypothesis isn’t validated, look at other changes. For example, you may not see an increase in demo sign-ups, but you could see an increase in social media shares, which is an interesting result and merits its own A/B test. 

If needed, break down results by audience, topic, or geography—but don’t over-segment. Doing so can reduce statistical significance.

4. ROI modelling 

One of the biggest challenges in content attribution is accurately measuring the ROI for each dollar spent. At Eleven Writing, we’ve found a straightforward way to overcome this hurdle using your existing data from GA4. Here’s how to do it: 

Step One: Get your conversion weighting in GA4

When configured correctly, GA4’s Path Exploration report shows the percentage of users who take a specific action, such as signing up for a demo, and later become paying customers. That percentage becomes your conversion weighting.

Step Two: Assign weightings to each conversion type

To quantify the impact of different conversions, assign weightings based on how often they lead to a purchase.

For example:

• 0.5% of demo sign-ups become customers → weighting = 0.005
• 0.8% of free trial sign-ups become customers → weighting = 0.008

Step Three: Create a formula field in Looker Studio

Import your GA4 data into Looker Studio, and then create a calculated field to apply your conversion weightings.

The formula for our example conversions is:

(SUM(CASE WHEN Conversion_Type = "Sign up for demo" THEN Conversions * 0.005 ELSE 0 END)) +
(SUM(CASE WHEN Conversion_Type = "Sign up for free plan" THEN Conversions * 0.008 ELSE 0 END))
‍
‍Important: Ensure the Conversion_Type values match exactly as they appear in Looker Studio.

Step Four: Normalise ROI for comparison

Since different pages have different visitor counts, normalising the scores allows for a fairer comparison.

To normalise, divide the weighted conversion score by the total number of visitors. This scales all scores between 0 and 1, where:

• 1 means every visitor converted.
• 0 means no visitors converted.

Let’s assume that Page A had 2,000 visitors and Page B had 4,000 visitors.

Conclusion: Page A has a better ROI per visitor, even though Page B had a higher overall conversion score.</p>

Step Five: Factor in Lifetime Customer Value (LCV) and production costs

To calculate the monetary value of each page’s performance, multiply the weighted conversion score by the Lifetime Customer Value (LCV).

Assuming LCV = $5,000, you’d calculate:

Then, calculate ROI using production costs (e.g., $500 per page):

Revenue - Cost / Cost

Final conclusion: Even though Page B generates more total revenue, Page A delivers a stronger ROI per visitor.

Conclusion

Attribution analysis gives you the clarity you need to make better marketing decisions faster. Whether you’re experimenting with A/B tests or digging into ROI, these techniques give you a clearer picture of performance and potential.

At Eleven Writing, we use these methods every day to help teams grow their content with confidence. If you'd like support applying these techniques to your own strategy, get in touch with us today.