Products Solutions Templates Pricing Resources

Root cause analysis techniques: digging deeper than the numbers

Home / Everything About / Everything About Analytics / Root cause analysis techniques: digging deeper than the numbers

Your conversion rate dropped 25 percent last month. You know what happened (diagnostic analytics). But you still don't know why. Is it a site change? A market shift? A competitor move? A seasonal pattern? Root cause analysis is how you dig past the surface to find the real reason.

This article covers the techniques to move from "something changed" to "here's exactly why it changed" and "here's what to fix."

Root cause analysis is the practice of investigating problems systematically to find the underlying reason, not just the symptom. Most website owners stop at "traffic dropped" or "conversions fell." Root cause analysis asks: what's the actual cause underneath that symptom?

A symptom is surface-level. A root cause is fundamental. If your bounce rate is high, that's a symptom. The root cause might be slow page load, confusing headlines, or poor mobile design. Fixing the symptom (try to lower bounce rate generally) doesn't work. Finding and fixing the root cause does.

Why symptoms mislead you

Website owners chase symptoms because they're easy to see. Traffic dropped—obvious symptom. But why? Dozens of possible causes. If you guess wrong and fix the wrong thing, the problem stays.

Root cause analysis prevents this. It forces you to ask why repeatedly until you reach the actual cause. The "5 Whys" technique is a simple version: ask why five times. Each answer becomes the basis for the next why.

Without root cause analysis, you fix the wrong problems and waste time. With it, your fixes actually work.

The 5 Whys technique: simple but powerful

Start with the problem. Ask why it happened. Take the answer and ask why again. Keep going until you reach the fundamental reason.

Problem: Conversions dropped 30 percent.

Why 1: Email campaign traffic dropped. But ad traffic stayed stable.

Why 2: Email click-through rate fell from 12 percent to 4 percent.

Why 3: Email open rate stayed the same, but once people opened the email, fewer clicked the link.

Why 4: We changed the email template last week. The new design moved the call-to-action button below the fold.

Why 5: The button is now invisible without scrolling. Most email readers don't scroll.

Root cause: Button placement is causing people to miss the CTA.

Fix: Move the button above the fold.

This is root cause analysis. You didn't guess—you systematically found the actual problem. The fix targets the real issue, not a symptom.

The fishbone diagram: organize your investigation

The fishbone diagram (also called Ishikawa diagram) helps you map all possible causes in categories. It's visual, which helps you see patterns.

Draw a horizontal line (the spine). The problem goes at the head. Branches off the spine represent categories: People, Process, Tools, Data, External Factors.

Under each category, list possible causes. Where do multiple causes intersect? That's often where the root cause sits.

Example: Your conversion rate dropped.

People: Did the sales team change? Did the writer leave?

Process: Did you change the checkout flow? Did you require more form fields?

Tools: Did you switch analytics tools? Did the payment processor change?

Data: Is the data incomplete? Is there a tracking error?

External: Did competitors launch? Did Google algorithm change?

By mapping this visually, you see which categories have the most possible causes. That's where you investigate first.

Breakdown analysis: segment to find the problem

Instead of looking at overall metrics, break them down. Overall conversions dropped 30 percent. But:

Email conversions: down 40 percent

Ad conversions: down 10 percent

Organic search conversions: down 15 percent

Now you know the problem is email-specific, not universal. That narrows your investigation dramatically. You're not looking at all traffic sources—just email.

Go deeper. Email visitors who converted previously spent 4+ minutes on the site. Now the average is 1 minute. Why? They're bouncing faster. From which page? The landing page. What changed on the landing page? The headline.

By breaking down systematically, you isolate the root cause.

Correlation vs causation: avoid false leads

This is the most common mistake. Two metrics move together, so you assume one caused the other. Usually wrong.

Your bounce rate increased the same week traffic decreased. Correlation: both changed. Causation: does bounce rate cause traffic to drop? No. Bounce rate is what happens after visitors arrive. It can't affect how many visitors arrive in the first place.

To test causation, ask: is there a logical mechanism? Could X actually cause Y?

Can a high bounce rate cause fewer visitors to arrive? No.

Can a slow page cause high bounce rate? Yes.

Can a slow page cause fewer visitors to arrive from Google? Yes (rankings drop if page is slow).

Test the mechanism before claiming causation.

Timeline analysis: when did it change?

If a metric changed suddenly on March 15, something happened on March 15. If it declined gradually over two weeks, something has been changing gradually.

A sudden drop suggests an event: a site change, a Google update, a competitor launch. A gradual decline suggests a trend: seasonal pattern, slow rank loss, changing audience behavior.

The timeline shape tells you what kind of cause to investigate.

Counterintuitive causes: look beyond the obvious

Most website owners blame the obvious: "We changed the site, so that must be the problem." Sometimes yes. Sometimes no.

Look for changes you didn't make. Did Google update? Did a competitor launch? Did the season change? Did iOS update and break tracking? Did your email list degrade?

Root cause analysis means questioning everything, not just the obvious suspects.

How to document your findings

When you find a root cause, document it. Write down:

The symptom: "Conversion rate dropped 25 percent"

The root cause: "Email CTA button moved below the fold"

The evidence: "Email click-through rate dropped from 12 percent to 4 percent after template change"

The fix: "Move button above the fold"

The result: "Email click-through rate returned to 11 percent after fix"

Documentation prevents you from making the same mistake twice and helps your team understand the problem.

When to stop investigating

Root cause analysis can go infinitely deep. At some point, you reach a cause you can't control or shouldn't spend more time on.

You found: "Button placement caused lower conversions." You could dig deeper: "Why was the button placed there? Because the designer thought it looked better. Why did they think that? They prefer that aesthetic."

At this point, stop. You've reached actionable cause (button placement). The why behind the why (aesthetic preference) is less important than fixing the actual problem.