The problem with standard SEO reporting

Most SEO reports are built to look good, not to prove causation. Impressions go up because Google indexed more of your pages — not necessarily because rankings improved on queries that matter. Rankings improve on low-volume keywords that drive no traffic. Traffic goes up in Q4 because it always goes up in Q4 for your industry. The report attributes all of it to SEO work.

This isn't always dishonest. Most SEO consultants genuinely believe the metrics they're reporting. But belief isn't proof. A business that grew 15% year-over-year in a market that grew 20% actually underperformed — but without a benchmark, the growth looks like success.

The same problem exists in SEO. If your organic traffic grew 25% after six months of SEO work, that sounds good. But if the average business in your category grew 30% organically over the same period due to market growth and algorithm changes, your SEO underperformed — and nobody told you.

The metric most reports lead with

Impressions are the number of times your pages appeared in search results — including at positions 40–100 where nobody clicks. Impressions can double while your traffic stays flat. If your monthly report leads with impressions as the primary success metric, ask for clicks instead.

Metrics that actually prove SEO impact

Not all metrics are equal. Here's how to think about which ones actually demonstrate that SEO work caused a business outcome:

Metric What it actually measures Verdict
Impressions How often your pages appear — including at positions where nobody clicks Weak
Average position Average ranking across all queries, including low-volume ones that don't drive traffic Weak
Domain Authority / DR Third-party scores that correlate with link profiles — not Google ranking signals Weak
Clicks (GSC) Actual visits from organic search — people who saw your listing and clicked Meaningful
Clicks on target queries Clicks specifically on the keywords your SEO work targeted Strong
Organic leads / form submissions Contacts generated from organic traffic — tied directly to revenue potential Strong
Organic phone calls Calls attributed to organic search visits — the hardest to fake Strongest

The further down that table you can get your reporting, the more confidence you have that SEO is producing real business impact. This is the framework behind measuring SEO ROI correctly — connecting organic performance to revenue, not just to rankings.

The holdout method — separating SEO from everything else

Even clicks and leads don't prove causation on their own. If your organic leads doubled in the six months after you hired an SEO consultant, three things might explain it:

Without controlling for the second and third explanations, you can't be sure the first is what caused the result. This is the problem quantitative analysts solve with holdout groups — a set of observations that don't receive the treatment, used as a control.

In SEO, a holdout works like this: when you start an optimization program, identify a set of pages that will receive work (the treated group) and a comparable set that won't (the holdout). After 60–90 days, compare how the two groups moved. If the treated pages improved significantly more than the holdout pages, that's evidence the SEO work caused the improvement rather than external factors.

A practical example

A plumbing company has 40 service area pages. An SEO consultant optimizes 20 of them — new title tags, improved content, schema markup added. The other 20 are left untouched as a holdout. After 90 days, the optimized pages show 35% click growth. The holdout pages show 8% click growth. The difference — 27 percentage points — is the estimated SEO contribution. The 8% the holdout pages gained tells you what would have happened anyway due to market growth or algorithm shifts.

Most consultants don't run holdouts because it creates accountability they'd rather avoid. If the holdout pages improve just as much as the treated pages, the SEO work didn't cause the results. That's an uncomfortable finding to deliver to a client paying a monthly retainer. But it's the honest answer — and it's the only way to know whether the budget is working.

How to run the analysis in GSC yourself

You don't need a data scientist to run a basic version of this analysis. Here's how to do it in Google Search Console with no external tools:

01
Identify your before and after periods
Your "before" period should be the 90 days immediately before SEO work started. Your "after" period should be 90 days at least 60 days after work began — giving Google time to crawl and re-rank. Use GSC's date comparison feature to pull both periods simultaneously.
02
Pull clicks by page — not by query
Go to Performance → Pages. Add the date comparison. Sort by click change. This shows you which pages gained and lost organic clicks between the two periods. Focus on clicks, not impressions — impressions can move without any business impact.
03
Separate treated pages from untouched pages
Take the page list from GSC and tag each row: did this page receive SEO work (new content, title tag changes, technical fixes) or not? Calculate average click change for treated pages and average click change for untouched pages separately. The difference is your signal.
04
Control for seasonality
If your business is seasonal, compare year-over-year rather than period-over-period. A landscaping company's organic traffic will always be higher in April than February — that's not SEO, that's spring. Use GSC's year-over-year comparison to strip out seasonal effects before attributing changes to optimization work.
05
Check for confounders
Did a Google core update run during your analysis period? Did you launch new pages or change your site structure? Did a major competitor close or open? Any of these can move rankings independently of SEO work. Cross-reference your timeline with Google's update history before attributing all movement to optimization.

The verdict framework — what the data is telling you

After running the analysis, you'll land in one of four situations:

Strong positive signal
Treated pages significantly outperformed holdout pages
Click growth on optimized pages exceeds holdout page growth by 15+ percentage points, and the gains are on queries that produce leads. The SEO work is likely causing the improvement. Continue.
Weak positive signal
Treated pages slightly outperformed, but gap is small
Treated pages gained more than holdout pages but the difference is less than 10 percentage points. The work may be contributing, but external factors are doing most of the work. Investigate which specific changes drove the largest gains and double down on those.
No signal
Treated and holdout pages moved similarly
Both groups gained (or lost) clicks at roughly the same rate. The movement was caused by external factors — algorithm update, market growth, seasonality. The SEO work didn't cause the results. This doesn't mean the work was bad, but it means you can't credit it with the outcome.
Negative signal
Treated pages underperformed holdout pages
The pages that received optimization lost more clicks (or gained fewer) than pages that were left alone. This is a serious finding. Either the optimization approach is wrong for these pages, the changes introduced technical problems, or the targeting was off. Investigate before continuing.

What to do with the answer

If the SEO is working

Document specifically what worked — which pages, which changes, which queries — so you can replicate the approach. Ask your consultant to apply the same methodology to the next set of pages and set expectations for what "working" looks like before they start, not after. This is how you build compounding SEO results instead of one-off wins.

If the SEO isn't clearly working

Don't immediately fire your consultant. First, check whether enough time has passed — meaningful SEO results typically take 4–6 months, and the holdout analysis is only reliable with at least 90 days of post-work data. If the timeline is right and the signal is flat, have a direct conversation: what specifically has been done, what was the expected outcome, and why does the data show treated pages moving similarly to untouched pages?

A consultant who has been doing genuine work will have specific answers. A consultant who has been sending reports without executing will struggle to answer that question concretely. The vetting framework applies retroactively as well as prospectively.

If you want someone to run this analysis for you

The analysis above is doable in GSC with a spreadsheet. But separating signal from noise accurately — controlling for seasonality, identifying confounders, building the holdout comparison correctly — takes time and familiarity with the data. If you want an independent, honest verdict on whether your past SEO investment produced real results, that's exactly what the SEO Impact Audit delivers.

Related
Already know SEO is working? Calculate the revenue impact.

The SEO ROI Calculator estimates what better rankings could mean in dollars — across three growth scenarios, based on your actual traffic and conversion numbers.

The honest bottom line

Most SEO reporting is not lying to you — it's just not designed to answer the question you actually care about. "Did this work?" is a causal question. Standard SEO reports answer a correlation question: "did things get better while we were working?" Those are different questions with different answers. The holdout method is how you get the honest one.