Is Your SEO Actually Working? Metrics That Matter Most
Your SEO report looks great. Rankings are up. Sessions are up. So why does the pipeline still feel flat? That gap is exactly why reporting literacy matters. It helps you judge whether an agency is driving real business results, and it helps you know when it might be time to fire one.
In this post, we'll look at the metrics that actually tell you SEO is working, not just looking busy. Rankings and raw session counts can flatter performance, get manipulated, or both. They rarely tell the whole story. Instead, we'll focus on four truth-tellers: qualified organic conversions, non-brand impression growth, assisted conversions, and branded search lift.
You'll also learn how to spot vanity metrics, read a baseline, question attribution, and evaluate local SEO visibility without getting dazzled by map pack noise. We'll cover what a credible monthly report should disclose, including the business outcome behind each metric, the baseline it is measured against, and what did not move. Finally, you'll get a simple stay, probation, or fire framework and three questions to ask every month.
This is Watt Consulting's point of view. Less flattery. More truth. Let's dig in.
Why a Flattering SEO Report Is a Warning Sign
You're not grading SEO as a discipline. You're grading whether the report in front of you is honest. Different skills, and only one is your job.
Flattery is the default because it's cheap. Rankings and raw traffic are easy to produce, easy to screenshot, and easy to move without helping your business at all. Movement isn't progress.
Every report owes you three answers: what changed, what baseline that change is measured against, and which business outcome it touched. Dodge one and you're reading a brochure. And yes, vanity metrics can be used the right or wrong way.
There's an incentive problem, and it deserves naming. Anyone selling SEO marketing services benefits from showing you green arrows. Your defense is literacy, not trust. Honest work also starts with a technical SEO audit that records what changed and when.
Vanity Metrics That Feel Great and Prove Nothing
Rankings aren't facts. In practice, position reports vary by device, location, and personalization, and they can shift when Google adds or removes a featured snippet you never touched. A screenshot of position 3 is a moment, not a measurement.
A ranking win on a term nobody searches is a participation trophy. Ask what volume and intent sit behind each keyword, and whether that searcher could become a customer this quarter.
Raw sessions hide quality. Traffic can climb from automated crawls, viral social spills, or a news mention, none of which represent purchase intent. 1,000 sessions producing two leads loses to 100 producing five.
In our experience, bounce rate, time on page, and third-party authority scores are directional at best. Layout and traffic source move them as much as content quality does.
Apply the decoration test: our rule of thumb is that if a metric rises while revenue stays flat for two straight quarters, treat it as decoration. Write down which numbers in your last report pass. Usually very few do. Not every number is junk, though; The Metrics That Actually Tell You Something Useful exist, and we'll cover them next.
Four Metrics That Actually Tell You SEO Is Working
Now for the good news. Each of the four numbers below is hard to inflate without doing real work, that friction is the point.
We pulled the definitions from Google's own Search Console and GA4 docs. The judgment calls about how much weight each deserves are ours, and we'll say so instead of dressing them up as industry consensus.
Every metric also has to arrive with context: a comparison period, a stated baseline, and a sentence about what would have made it move the other way. And since SEO takes time to compound, remember that in our view, a single month is noise; we look for direction across at least two quarters before judging an agency's impact.
1. Qualified Organic Conversions
Of the four, this one ends arguments. Qualified organic conversions are people from organic search who did something invoiceable: booked a call, requested a quote, sat a qualified demo, made a purchase. Not a form fill from a student researching a school project.
Ask how the agency defines qualified. If it says "any conversion event in GA4," push back. Most sites count newsletter signups, PDF downloads, and contact-page views alongside revenue-driving actions. Check vocabulary: GA4 has moved to calling these key events, so a report still saying "conversions" may count a loosely defined bucket.
Before you trust the number, map out where your customers actually convert.
Compare against the same month last year, not just last month, because seasonal patterns, not SEO, drive many month-to-month swings. A flat quarter on qualified conversions beats a rising quarter on sessions, and a report leading with the former is telling you something good.
2. Non-Brand Impression Growth
If qualified conversions are the scoreboard, non-brand impressions are the early warning system. In Search Console, split impressions into branded queries (people already searching your name) and non-brand queries (everyone else). Only that second bucket is new demand you're earning.
Non-brand growth shows up before conversions do, so it's a solid leading indicator that content and technical work are reaching strangers. But watch the ratio, not the raw number. Branded impressions climbing while non-brand stays flat means you're just getting better at being found by people who already knew you.
Impressions measure visibility, not intent or preference, so never let them stand alone. Pair them with conversions, and check how to find the AI performance reports in Search Console to see how your visibility is trending over time.
3. Assisted Conversions
Impressions tell you people are seeing you. Assisted conversions tell you whether that seeing turns into something later. Most B2B and considered-purchase buyers research across multiple sessions before converting, which means assists count organic touchpoints that showed up earlier in a path that eventually ended in a sale or lead, value that last-click reporting throws away.
Here's the catch: every platform calculates assists differently. If your report doesn't name the attribution model behind the number, treat it as unverified and ask directly. Note that Google sunset four models, linear, first click, time decay, and position-based, from Google Analytics 4 and Google Ads, with the forced migration completing in September 2023, though these models were already unavailable for new conversion actions months before that cutoff. Then read the assist number for funnel shape, not credit. A rising assist count with falling final conversions means you're helping people along without closing them.
If you have a CRM, push for the real payoff: tie organic assists to closed revenue instead of form fills. That one connection turns a soft metric into a financial one, and that's a number a decision can actually use.
4. Branded Search Lift
Count how many people search your brand name each month. When SEO, content, and social pull in the same direction, that number climbs, because people remember you and come back looking for you by name. It also rises when your content and brand voice stay consistent across channels.
Be patient. In our experience this metric lags everything else, we typically look for a pattern over two to three quarters before drawing conclusions. One month of movement is mostly noise and seasonality.
Read it carefully. Paid spend, PR, word of mouth, and offline advertising all lift branded search too, so treat it as a confirmation signal, not isolated proof.
Locally, watch for queries pairing your name with a neighborhood or service. Separating those out is worth the effort, since they point to genuinely local demand.
How to Read a Baseline, and Why No Baseline Means No Trust
Branded search is your slowest signal, so every other number needs a fixed reference point.
Insist on four things in writing before any chart counts:
A start date. When did the work begin?
A comparison period. Stated, not implied.
A seasonality note. What does this business normally do here?
A tracking change log. Migrations, consent banners, the Universal Analytics to GA4 switch, form rebuilds. Each can swing reported numbers with zero real change in performance.
That log matters most and gets skipped most.
Match comparison periods to your business cycle. Retail, B2B, and professional services peak at different times, so a like-for-like quarter from last year is the fairest lens.
Ask for the raw export alongside the summary slide. A credible partner hands it over without hesitation. Reluctance to share it is a clear early warning.
Finally, write your own baseline down and keep it. If a new agency cannot reconstruct where things stood, you cannot judge anyone, including yourself.
What a Credible Report Discloses About Attribution
Baseline set, attribution is the next tell. A trustworthy report names its attribution model, says in plain language what that model favors, and admits what it therefore undercounts. Last-click shortchanges early organic touchpoints; other single-touch models create their own blind spots depending on where in the funnel they anchor.
Check for footnotes on excluded traffic. Any changes to filters or traffic classification switched on mid-period need disclosure. It is the same discipline as a content audit triage checklist: log what you changed, or the numbers lie to you later.
Be suspicious of blended reporting. Rolling organic, paid, social, and email into one line hides which channel is actually carrying the result, and it makes smart budget shifts impossible.
One question, every month: which business outcome does this line map to? If nobody can answer, cut the line.
Local SEO Services: When Map Pack Visibility Is Real and When It Is Noise
Same test locally: which numbers earn a line on the report?
Anyone who has tracked local rankings knows they shift with the searcher's location and query context, and are easy to screenshot in a flattering way, so treat any single snapshot with caution. What holds up? Directions requests, calls, and website clicks from your Google Business Profile, plus conversions from local landing pages. Those describe someone taking a step toward visiting or buying.
Separate discovery from proximity. Proximity heavily influences map pack results, a business may appear simply because the searcher is nearby, not because it has earned wider demand. Ask any provider of local SEO services how they measure visibility outside your immediate radius. Real growth shows up across a defined set of target areas, not one lucky pin drop. Reviews are the supporting signal we watch: volume and recency are known factors in local visibility, and they reflect real customer experience in a way that is hard to sustain artificially. For the mechanics, see how to rank in the Google Local Pack in Seattle.
Stay, Probation, or Fire: A Decision Framework
Here's a simple way to decide.
Talk first. One soft month, a dip on a volatile keyword, or a traffic shift tied to a tracking update is not a firing offense. It's a question. GA4's direct traffic often signals lost attribution, not lost business, so dig before you decide.
Probation. Qualified conversions flat two quarters running while spend holds steady, stagnant non-brand impressions, or reports that keep skipping baselines. Put written goals and a review date in place.
Terminate. Repeated refusal to hand over raw data, metrics tied to no business outcome, undisclosed attribution changes, or a report that never admits anything stalled.
Exit cleanly. Request a final report in the same format, transfer Analytics and Search Console access, and document the baseline.
And stay fair. SEO compounds; technical work often shows nothing for a quarter. Slow results and absent evidence are different problems with different answers.
The Three Questions to Ask Every Month
Whatever call you land on, ask the same three questions every month. They take very little time and cut through most reporting theater.
Which business outcome does this number map to? If a metric can't answer that, it belongs in an appendix, not at the top of the report.
What's the baseline? Start date, comparison period, seasonality note, and a written log of any tracking changes, every single month.
What didn't move? An agency that answers this candidly is worth keeping. One that can't is telling you how it'll handle the next hard month.
Ask all three on a short call or in email. The answers give you a defensible reason for every decision you make about your SEO partner.
Conclusion
SEO reports should prove business impact, not just look impressive. If a partner cannot explain which outcome a number maps to, or admit what stalled, that silence is your answer.
A credible SEO partner earns trust through transparency, not theater. Run the questions, read the answers, and let real metrics guide your next move. Clarity is yours to demand, and it starts with the very next report you read.
Comments