How to Tell If Your SEO Is Actually Working
The bottom line
SEO works the way a savings account works, not the way a slot machine works. You judge it by the slope of the line over months, by whether qualified traffic to pages that make money is growing, and by whether the people doing the work can explain their reasoning and show their checking. A single ranking on a single day tells you almost nothing. If your report only celebrates that you rank first for your own company name, you are being managed, not informed.
If you sign the invoices for SEO, you are being asked to judge a discipline that is deliberately opaque to outsiders. The vocabulary is dense, the feedback loop is slow, and almost everyone reporting to you has an incentive to make the picture look good. This briefing gives you the small number of things that actually separate real progress from theatre, so you can read a report in ten minutes and know whether to trust it.
The mental model: trajectory, not snapshots
The single most expensive mistake executives make is checking a ranking on a Tuesday, seeing it slip, and concluding the work failed. Search is probabilistic and lagged. Google itself, in its official Search Engine Optimization Starter Guide, states that changes may take anywhere from hours to several months to be reflected, and that you should expect to wait at least a few weeks to assess whether work is helping. Results are also personalised and localised, so the rank you see is not the rank your customer sees.
So judge SEO the way you judge a pipeline or a fund: by the trend line over a rolling window, typically three to six months, against a baseline you recorded before the work started. One good week proves nothing. One bad week proves nothing. The slope of the line, sustained, is the signal.
This also changes how you should react to volatility. Google rolls out broad core updates several times a year, and it confirms each one on its public Search Status Dashboard. During and just after those windows, rankings move for reasons that have nothing to do with your team's competence. A mature operator anticipates this, annotates the report when an update lands, and waits for the dust to settle before drawing conclusions. Panic on a single bad day is a tax you pay for not understanding the model.
Metrics that matter versus vanity metrics
Most SEO reporting is padded with numbers that move impressively and mean nothing. Here is how to tell them apart.
| Metrics that matter | Vanity metrics |
|---|---|
| Organic sessions to your money pages (the pages that actually sell or capture leads) | A single keyword rank, screenshotted on a good day |
| Qualified organic conversions, tied to revenue or pipeline | Total impressions with no context or trend |
| Share of non-brand visibility (are people who do not know you finding you?) | Ranking first for your own brand name |
| Indexation health (are your important pages actually in Google's index?) | "Pages published this month" reported as a result |
| Growth in valuable keyword coverage (more queries returning your pages on page one) | Domain authority going up by a point, in isolation |
| AI and assistant referral traffic (ChatGPT, Perplexity, Google AI Overviews sending visitors) | "Number of backlinks built" with no quality or relevance test |
Brand search deserves a special warning. You will almost always rank first for your own company name, with or without SEO, because Google's job is to find the thing you literally typed. A report that leans on brand terms to show growth is hiding behind demand it did not create. The honest measure of acquisition is non-brand: visibility among people searching for the problem you solve, not for you. The newest line item worth tracking is AI and assistant referral traffic. As people increasingly start with ChatGPT, Perplexity, or Google's AI Overviews, those surfaces send real visitors, and your analytics can now show them as a referral source. A team paying attention will already be reporting it.
How to read a report without being snowed
A good report does four things. It states a baseline and compares against it. It connects search activity to business outcomes, not just traffic. It separates what was done from what resulted, and is honest about which is which. And it is specific about what it does not yet know.
Signals it's working
- Non-brand organic traffic trending up over months
- Conversions and revenue cited next to traffic
- A baseline shown, with clear before-and-after
- More money pages ranking on page one
- Honest notes on what slipped and why
Red flags
- Only brand terms reported as wins
- No baseline, so nothing to compare to
- No business metric anywhere in the report
- Activity ("we did X") sold as outcome
- Charts with no axis labels or date range
The honest bit about AI-accelerated reporting
Reports are increasingly drafted with AI assistance, and this triggers two opposite and equally lazy reactions. One is to assume that anything produced quickly must be shallow. The other is to assume that a long, dense report must be rigorous. Both are wrong. A report produced in an hour can be sound if the underlying data was pulled correctly and the conclusions were verified. A forty-page report can be worthless if its charts are decorative and its claims were never checked against reality.
Judge the thinking, not the turnaround time or the page count. Does the report show its sources? Can the author tell you where a number came from and how they confirmed it? Does it distinguish what the data proves from what it merely suggests? Length and speed are not quality. Verification is quality.
How to tell if your team or agency is good
The strongest signal is not the cleverness of the tactics. It is intellectual honesty. Good SEO people behave in five recognisable ways.
They set expectations on timelines. They tell you up front that meaningful movement takes months, and they do not promise page-one positions by a date. They tie work to business outcomes. Every initiative connects back to traffic that converts, not to abstract scores. They tell you when not to do something. A team that only ever says yes is selling, not advising; the good ones will talk you out of low-value work. They show their checking. They can walk you through how a number was produced and verified. They admit uncertainty. Search has no guarantees, and anyone who claims certainty is either inexperienced or not being straight with you.
Questions to ask your team or agency
- Strip out brand search. What is non-brand organic traffic doing over the last six months, against the baseline?
- Which of our money pages improved, and how did that change conversions or pipeline?
- Show me one number in this report and walk me through exactly where it came from and how you verified it.
- What did you decide not to do this quarter, and why was that the right call?
If you get clear, specific, slightly uncomfortable answers to those four, you are probably in good hands. If you get deflection, brand-term victory laps, or charts without context, the problem is not your understanding of SEO. It is the work.
Want an honest second opinion on your SEO?
We read the report behind the report: real baselines, real business metrics, and a plain-English verdict on whether the work is moving the needle.
