How to Measure SEO Results and Know What You're Paying For
Picture Anna, who owns a small physiotherapy clinic. On the first Monday of the month, a PDF arrives from her SEO agency. There are green arrows beside "Keyword rankings" and "Organic traffic," and a page of completed work: technical audit, on-page optimisation, link building. She reads it twice. Then she opens her booking calendar and cannot tell whether a single new patient came from any of it.
Anna is a hypothetical example, and every number in this guide is illustrative. She runs through the whole article so each metric ties back to a decision an owner actually makes.
Anna's problem is not a lack of data. The report simply does not answer the question she is paying for: is this bringing in the right people?
The short answer: to measure SEO results, follow four things in order. Track visibility in search, relevant traffic, enquiries and sales. Compare each against a starting point and goals agreed before the work began. Rankings and traffic show that SEO is moving. Enquiries and revenue show whether the movement matters.
The rest of this guide shows how to set that up, how to read the numbers without fooling yourself, and how to check whether a monthly report backs up what it claims.
Before You Open Another Report, Define What SEO Should Deliver
You cannot judge a result without knowing what result you expected. Before any metric, decide what a good month looks like for your business, in your own terms rather than in the agency's.
Choose the business outcome that matters
The right outcome depends on how your business earns money:
- Local business: confirmed phone calls, bookings and form enquiries from people in your service area.
- Online shop: orders, revenue and profit from organic visitors, not just visits.
- B2B company: qualified leads, sales opportunities created and deals closed.
For Anna, the outcome is new-patient bookings from people within a 15-minute drive. A visit from another country is not a result, however good it looks in a chart.
Record your starting point
Write down five things before you judge anything: baseline numbers (clicks, enquiries, bookings), the priority pages, the target region, the date work began and what you pay each month. Without a baseline, every later number floats free of context.
Here is Anna's agreed goal, with illustrative figures:
Baseline (the three months before work began): 40 organic enquiries a month, 22 of them qualified, 9 booked. Goal by month six: 55 organic enquiries, 32 qualified, 14 booked, mostly through the sports-injury and back-pain pages. Spend: $1,200 a month.
The goal is not a promise. It gives Anna and her agency a shared yardstick, so "working" stops meaning "it feels like it."
Follow the Journey From Search Visibility to Actual Customers
SEO results arrive in a chain: someone sees your page, clicks, visits, acts, and sometimes buys. Each link answers a different question and supports a different decision. Reading only the first links is how green arrows end up next to an empty calendar.
Impressions: Are You Appearing for Relevant Searches?
An impression means your page appeared in a search result. You find it in the Performance report in Google Search Console. It tells you whether Google considers you a candidate for searches that matter, and it helps you decide whether to keep expanding into a topic.
Impressions need context. A jump from 8,000 to 20,000 could mean the site now appears for "physiotherapy jobs" and "what is a physiotherapist." Nobody who searches those is a patient. Filter by country and by query before you cheer, and check that the queries describe what you actually sell.
Rankings: Are the Right Pages Moving Up?
Ranking data is useful when it is specific: an agreed group of queries, each tied to the page that should answer it. "Sports injury physio" should lead to the sports-injury page, not the homepage.
Be careful with the average position for the whole site. It blends very different searches, so it can rise because a few easy queries improved, or fall because the site started appearing for many new ones. Google's documentation explains how positions are counted. Use rankings to see whether priority pages are moving, then check the steps after the click to see whether that movement matters.
Organic Clicks and Sessions: Are More Relevant People Visiting?
These two numbers come from different tools and are not interchangeable. Clicks (Search Console) count visits from Google results to your site. Sessions (GA4) count visits that GA4 recorded, which depends on tracking, consent settings and how it attributes sources. Expect them to differ a little, and treat a large, unexplained gap as a reason to check the setup.
Slice them by landing page, country and device. Anna wants to know whether the sports-injury page gets more visits from her city, and whether those visitors are on phones, where the "call now" button lives.
Leads and Sales: What Happens After the Visit?
This is the link that decides whether SEO paid for itself. In GA4, mark important actions as key events (for example form submissions, bookings and call-button clicks), as described in Google's help page on key events. Then confirm quality in your CRM or booking system.
Keep two things separate: a tap on a call button, and a conversation that really happened. Anna's phone log may show 12 call-button taps in a month, 7 answered calls, and 3 bookings. Those are three different numbers, and only the last one pays the rent.
Here is how the metrics line up:
| Metric | Where to find it | What it tells you | What it cannot prove |
|---|---|---|---|
| Impressions | Search Console, Performance | You appear in searches | That the searches are relevant, or that anyone chose you |
| Rankings | Search Console, a rank tracker | Priority pages are moving for agreed queries | That visitors arrive, or that results look the same for every searcher |
| Clicks | Search Console | People chose your result | That they were the right people, or that they stayed |
| Organic sessions | GA4, Traffic acquisition | People visited from search | What they did next, or that tracking is complete |
| Key events | GA4 | Visitors took an important action | That the action was a real, qualified enquiry |
| Qualified leads and sales | CRM, booking system, call log | The business outcome you care about | Exactly which channel deserves the credit |
Set Up a Measurement System You Can Trust
You do not need a complicated stack. You need three sources that agree well enough to rely on, and one habit: check that tracking works before you draw conclusions from the numbers it produces.
Use Google Search Console to Understand Search Performance
Search Console shows what happens before the click: queries, pages, clicks, impressions, click-through rate (CTR) and average position. Google's guide to the Performance report covers each field. The comparisons that matter most are the last three months against the previous three, and the same period a year earlier if history exists. Filter by page (the priority pages first), by country, and by query (branded and non-branded separately).
Use GA4 to Understand What Visitors Do
GA4 shows what happens after the click. Look at organic visits, the landing pages they arrive on, and the key events they trigger. The Traffic acquisition report describes where sessions come from, and it can be viewed for new and returning users, as Google's help page explains. Use it to find out whether organic visitors reach the pages that earn money and then do something there.
Connect Enquiries to Your CRM or Sales Records
Analytics counts actions. Your own records tell you which of them were real. Even a small business can keep a simple enquiry log, one row per enquiry:
| Date | How they found you | Landing page | Type (call, form, booking) | Qualified? (yes/no) | Duplicate or spam? | Became a sale? | Value |
|---|---|---|---|---|---|---|---|
| 3 Mar | Google search | /back-pain-physio | Form | Yes | No | Yes | $90 first visit |
Asking "how did you find us?" on the phone and on the booking form fills the second column with surprisingly useful answers. The log also exposes spam, duplicates and enquiries that were never a fit, all of which inflate analytics.
Check the tracking first. Submit your own test form. Make a test call. Confirm the key event appears in GA4 and the entry appears in your log. If a plugin update or redesign broke the form tracking in February, a "drop in leads" in March is a measurement problem, not an SEO problem.
Compare Results Without Fooling Yourself
Sometimes the data is accurate and the conclusion is still wrong. Four habits prevent most of those mistakes.
Compare Equivalent Periods and Account for Seasonality
Do not compare December with January, or a holiday week with an ordinary one. Compare the same month a year earlier when you have the history, and look at the monthly trend instead of a single data point. A physiotherapy clinic may see more back-pain searches after the new year. A tax adviser sees a spike every spring. A 30% rise that your market delivers every March is not an SEO win.
Separate Branded and Non-Branded Searches
Branded searches include your company name. Non-branded searches describe what you do, like "physio for runner's knee." Non-branded growth is the clearest sign that SEO is bringing in people who did not already know you.
Brand searches can rise for other reasons: a radio ad, a local news story, a strong referral month. Keep them in the report separately, but keep them in the overall business picture too. Someone who searches your name and books is still a patient. The point is to know which part of the growth SEO created.
Look at Priority Pages, Markets and Search Intent
A healthy sitewide total can hide a problem. Blog traffic may double while the service pages that bring bookings lose clicks. Check the pages that earn money, the markets you serve, and whether visitors came looking for information or looking to book. Informational visits help brand awareness, but they rarely become next week's appointments.
Keep a Record of What Changed
Keep a simple change log: content published, technical changes, redesigns, analytics changes, price changes and ad campaigns, each with a date. When a number moves, the log is the first place to look for an explanation. It also protects you from crediting SEO for a spike that coincided with a promotion.
Read the Numbers Together: Four Common SEO Scenarios
No single metric tells the story. These four patterns come up often, so each one is laid out the same way: what you see, what might explain it, what to check, and what to ask. All figures are illustrative.
Impressions Are Up, but Clicks Are Flat
What Anna sees: impressions grew from 8,000 to 14,000, while clicks stayed near 310.
Possible explanations: the site now appears for broader or less relevant queries; positions are low (page two); a competitor's listing or a map result takes the click; the listing is unappealing.
What to check: which queries gained impressions, their average position, CTR by query, countries, and what the results page looks like for the main searches.
Ask your provider: "Which queries added the impressions, and are they ones we want to win?"
Traffic Is Up, but Enquiries Are Not
What Anna sees: organic sessions rose 35%, while enquiries stayed at 40.
Possible explanations: visitors want information, not an appointment; the new traffic lands on blog posts instead of service pages; the form or phone button is broken or hard to find; the offer is unclear; tracking misses something.
What to check: landing pages that gained traffic, how visitors from each page behave, a test submission of every form, and the enquiry log against GA4 key events.
Ask your provider: "Which pages produced the extra traffic, and which of them lead to enquiries?"
Rankings Are Up, but the Business Impact Is Unclear
What Anna sees: twelve tracked keywords moved up an average of six places, but bookings did not change.
Possible explanations: the keywords have little commercial intent; they have low search volume; positions are measured from a location or device unlike her patients'; the pages ranking are not the ones that convert.
What to check: how each tracked keyword was chosen, its monthly searches, the clicks it sends, and what visitors do after landing.
Ask your provider: "Which of these keywords brought visits, and which of those visits turned into enquiries?"
Traffic Is Down, but Qualified Leads Are Up
What Anna sees: organic sessions fell 15%, yet qualified enquiries rose from 22 to 27.
Possible explanations: low-intent informational traffic dropped while service-page visits held or grew; a cleanup removed irrelevant pages; spam enquiries were filtered out; the mix of visitors improved.
What to check: traffic by landing page and intent, which pages lost visits, and whether enquiry quality was judged consistently in both periods.
Ask your provider: "Is the decline in pages that matter to the business, or in pages that never produced patients?"
This last scenario is why a lower number is sometimes a better result. Fewer visitors who fit the business can beat more visitors who do not.
How Long Should You Wait Before Judging SEO Results?
There is no honest universal answer, because three clocks run at different speeds: when the work is done, when search results change, and when the business feels it.
Google's own SEO starter guide notes that the impact of changes can show up anywhere from a few hours to several months, and that it is often worth waiting a few weeks before judging whether a change helped. A provider who promises results by a fixed date is promising something nobody controls.
That does not mean you must wait in the dark. In the first months you can already judge four things:
- Was the agreed work completed, and is there evidence of it?
- Is tracking working, and do the numbers reconcile with your own records?
- Do you have a baseline and a goal to compare against?
- Is the next step clear, along with how it will be evaluated?
Business results come later, and the trend over several months tells you more than any single month.
How to Tell Whether Your SEO Report Supports Its Claims
A report is a set of claims. Your job as the client is not to distrust it, only to check that each claim comes with enough evidence to be useful.
Every Growth Claim Needs a Baseline and a Source
Take the line "Organic traffic increased by 40%." On its own it says almost nothing. You need the period compared, the source (Search Console clicks or GA4 sessions), the absolute numbers (40% of 200 is 80 extra visits, and 40% of 5,000 is 2,000), the segment (all pages, or priority pages only, branded or non-branded) and what it meant for the business. A better version reads: "Non-branded organic clicks to service pages rose from 210 to 294 between Q1 and Q2 (Search Console), and form enquiries from those pages went from 11 to 15."
Completed Work Should Be Traceable
If the report says a task was done, you should be able to see it. Useful evidence includes the URLs of updated pages, the date the change went live, a short description of what was fixed and why, published articles with links, and the live links for any placements. "On-page optimisation: 12 pages" is a claim. A list of the twelve URLs is evidence.
Repeated Tasks Need an Explanation
Regular monitoring can be perfectly reasonable. Checking indexing, uptime or crawl errors every month is routine care. But the reader should still learn what was checked this month, what was found and what was changed. If the "Work completed" page is word-for-word the same in March, April and May, ask what was different each time.
Declines and Missing Data Deserve Attention
A good report explains significant drops, flags gaps in tracking and says when the way of comparing has changed. If a metric you saw last month quietly disappears, or a chart switches from clicks to sessions, ask why. Silence about bad news is information too.
Use this table when a report arrives:
| Report claim | Evidence to request | Question to ask |
|---|---|---|
| "Organic traffic grew 40%" | Period, source, absolute numbers, segment | "Is that clicks or sessions, and is it growth on priority pages?" |
| "Rankings improved" | The keyword list, its target pages, dates and location settings | "Which queries, and which pages now rank for them?" |
| "Technical issues fixed" | URLs, before and after, the date of the fix | "Where can I see what changed?" |
| "New content published" | Live URLs and publication dates | "Which of these pages already bring visits or enquiries?" |
| "Backlinks acquired" | Live linking URLs, anchors and the relevance of each site | "Can I open each link and see where it sits?" |
| "Monitoring completed" | What was checked, what was found, what was changed | "What was different this month from last month?" |
Can You Calculate SEO ROI?
Yes, with honest limits. The basic formula is:
SEO ROI = (profit attributed to SEO − SEO costs) ÷ SEO costs × 100%
Four details decide whether the result means anything:
- Costs: count everything. Agency fees, content, tools, developer time and any internal hours belong in the total.
- Revenue versus profit: use profit (or gross margin), not revenue. A $50,000 sale with a $45,000 cost base is a very different result from one with a $20,000 cost base.
- Delay: a patient who enquires in May may book in June. Match sales to the enquiries that produced them, not to the calendar month they close.
- Attribution: people often find a business through search, then return through a referral or a direct visit. Analytics will not capture every step, and the log of "how did you find us?" answers helps fill the gaps.
An illustrative calculation for Anna: 14 new patients came from organic search in a month, each worth $260 in profit, so $3,640. SEO cost $1,200 that month. ROI is ($3,640 − $1,200) ÷ $1,200 × 100%, or about 203%. Treat a single month with caution. A longer window smooths out timing.
If your sales records are not reliable enough for this, start with a simpler number: cost per qualified organic lead. With $1,200 spent and 32 qualified enquiries, it is $37.50. Track it month by month. It tells you whether the channel is getting more efficient, and it needs only the enquiry log from earlier.
What to Ask Your SEO Provider at the Next Monthly Call
The aim of the call is not to catch anyone out. It is to turn a general conversation into a specific one. These six questions work for almost any business:
- Which pages generated more qualified enquiries this month?
- How much of the growth came from non-branded searches?
- What changed on our priority pages?
- Where can I see the work completed this month?
- What declined, and what have you checked so far?
- What is the next priority, and how will we evaluate it?
A good provider will usually welcome them, because clear answers make the relationship easier for both sides. Write down the answers and bring them to next month's call, so progress is measured against what was said rather than what is remembered.
Get a Second Opinion When the Reports Still Leave Questions
Sometimes you do everything above and the reports still leave you unsure. The numbers look fine, the task lists look busy, and you still cannot tell what changed. That is the situation Check My SEO Guy was built for.
You upload two to five of your recent monthly SEO reports, and the AI analysis looks across them for the things that are hard to spot one report at a time: tasks repeated month after month, figures that contradict each other, growth claims with no baseline or source, missing context, and statements that deserve a closer look. The aim is a clearer picture of what the reports actually show, so your next conversation with your provider is more specific.
There are limits, and it is better to know them up front. The service analyses the documents you provide. It cannot see inside your Search Console or analytics accounts. Confirming that a link is live, that a ranking is accurate or that a technical fix was deployed may need a separate check of the site itself.
Upload your recent SEO reports and get a clearer picture of what they actually show.
Your Monthly SEO Results Checklist
Come back to this list each month. It takes less time than reading the report twice.
- Check that tracking works: test a form and a call, and confirm both appear in GA4 and in your enquiry log.
- Compare the agreed metrics for equivalent periods, splitting branded from non-branded searches.
- Review the priority pages and the quality of the enquiries, not just the totals.
- Match each claimed piece of work to evidence: URLs, dates and live links.
- Write down the open questions, the next actions and how each will be evaluated.
Frequently Asked Questions
How do I know if my SEO is working?
You know when relevant visibility, relevant traffic and qualified enquiries move together over several months, compared against a baseline and a goal you agreed in advance. One number moving alone is only a hint.
Which SEO metrics should a small business track?
Start with five: non-branded clicks to priority pages, rankings for an agreed query group, organic sessions to service pages, qualified enquiries, and booked or paid customers. Add cost per qualified lead once you have a few months of data.
Can I measure SEO results with free tools?
Yes. Google Search Console and GA4 are free, and a spreadsheet works as an enquiry log. Paid rank trackers and CRMs add convenience, but the basic picture is available at no cost.
Why are rankings improving but leads are not?
Common reasons include keywords with little buying intent, low search volume, ranking pages that are not the ones that convert, broken form or call tracking, and an unclear offer on the page. The scenarios above show how to check each one.
What should a monthly SEO report include?
A useful report states the goal and baseline, shows the metrics for equivalent periods, separates branded from non-branded search, and covers priority pages and enquiry quality. It lists completed work with evidence, explains declines and gaps, and names the next steps.
Can an AI review verify everything in an SEO report?
No. An AI review of your reports can find repeated tasks, contradictions, missing context and claims that need checking. It cannot confirm live links, current rankings or technical fixes by itself. Those need a separate look at the site or the source accounts.