SEO KPIs that matter

Results first, signals second

A KPI is a measure you would act on. If it went up or down, would you change the budget, the plan or the team? If not, it belongs in an appendix, not at the top of the report.

For SEO, the measures that pass that test are commercial: sales, leads, revenue, gross profit and what each sale costs. Everything else, from rankings to clicks, is a signal that helps explain those results. Signals matter, especially early on, but they are not the point.

This matters most when you defend a budget. A finance lead does not fund impressions. They fund return. If your business case promised a payback month, the report should track progress toward it.

The three tiers of SEO KPIs

Tier 1 goes at the top of every report. Tier 2 explains it. Tier 3 only needs a line unless something has broken. A technical audit covers tier 3 in depth when it does.

Separate branded from non-branded

People who search for your company name were going to find you. Their visits rise with brand awareness, ads and PR, not with SEO work. Report non-branded organic clicks and conversions as the SEO measure, and keep branded figures as context. Search Console lets you filter queries to do this.

Worked example: from leads to cost per deal

A B2B firm spends $4,500 a month on SEO. Here is one month of its funnel from non-branded organic search. The figures are illustrative. Use your own CRM rates, not these.

Cost per lead looks excellent at $15. But only the $500 cost per deal can be compared with the deal's margin and with other channels. If gross margin on a $6,000 deal is 40%, each deal adds $2,400 of gross profit. Nine deals add $21,600, against $4,500 of spend that month.

Two cautions. First, deals close later than the leads arrive, so match this month's spend with the leads that are now closing, or report on a rolling three-month basis. Second, count the uplift above your baseline, not every organic lead, when you calculate ROI.

Tying GA4 data to revenue

GA4 can show organic sessions and conversions out of the box. Getting from there to revenue takes some setup.

1. Mark the right key events. Purchases, qualified form submits and booked calls. Not scrolls or page views.
2. Send values. E-commerce purchase events should carry revenue. For leads, assign an expected value based on your lead-to-deal rate and deal size, and label it as an estimate.
3. Report by channel and landing page. Filter to Organic Search in the traffic acquisition report, then look at landing pages to see which content earns.
4. Link Search Console. This adds query and position data next to landing pages, so you can connect rankings to the pages that convert.
5. Carry the source into your CRM. Save the session source or UTM data with each lead. Then closed revenue can be reported by channel, which GA4 alone cannot do for long B2B cycles.
6. State the attribution model. Use the same model for every channel you compare, and say which one it is in the report.

If conversion rate is the weak link, test page changes properly. A lift that looks good after a week may be noise. The A/B test significance tool tells you whether a difference is likely to be real before you roll it out.

Vanity metrics to drop

These numbers are easy to grow and easy to report, and they say little about return.

• Total impressions. They can rise on queries that never convert.
• Number of keywords ranked. Thousands of positions on page five earn nothing.
• Third-party authority scores. Useful for comparing sites. They are a vendor's estimate, not a Google metric and not a business result.
• Total sessions, branded included. Mixes brand demand with SEO effect.
• Links built, as a count. One relevant link to a page that sells can matter more than dozens elsewhere.

Here is why it matters. Imagine organic sessions rise from 8,000 to 10,400 in a quarter, up 30%, while organic leads go from 120 to 118. A traffic-first report calls that a win. A results-first report asks why the new visitors do not convert, and finds the problem: maybe the new pages target the wrong intent, or the forms need work. On-page CRO or a rethink of content targets would be the next step, not more of the same.

A monthly report that works

Keep it to one screen.

1. Results: organic sales or qualified leads, revenue and cost per sale, against forecast.
2. Progress to payback: cumulative profit against cumulative cost.
3. Leading signals: non-branded clicks and target-group positions, with one line on why they moved.
4. Issues and next actions: what broke, what is being fixed, what ships next month.

In the first months, results will be small and the leading signals will carry the story. That is fine, as long as the report says so plainly and shows when results are expected to take over. A report that switches its headline metric each month, to whichever one went up, loses trust quickly.

Compare each month with the forecast range agreed at the start, not with a hope. Model the expected numbers in the ROI calculator, then see how our analytics and reporting service builds this report from your own GA4 and CRM data.