Five Things Your SEO Report Isn't Showing You.

1. Whether the traffic is human.
Sessions, page views, time on site - these all get reported as if they represent real people. They don't. Automated traffic, bot networks and low-quality referrals inflate every metric. The agency celebrates growth. You wonder why leads are flat. We built BotExorcist to close this gap. Most agencies haven't.

2. Which pages are actually driving enquiries - not just visits.
Most reports rank pages by traffic. That tells you what's popular, not what's working. A "Contact Us" page with 50 visits and 10 form submissions is worth more than a blog post with 2,000 visits and zero conversions. We report on commercial value, not readership.

3. What Google can't reach.
Standard SEO reports show what's indexed. They don't show what Googlebot tried to crawl and failed. Blocked resources, soft 404s, orphan pages, redirect chains - these are invisible in most reporting. They're often the fastest fix available.

4. What happens after the click.
SEO agencies report the click. They don't report whether the landing page loaded fast, whether the contact form submitted, or whether the phone rang out. That's "conversion optimisation" - a separate budget. We can't separate it because we also build the sites and answer the calls.

5. Whether the scope matches the problem.
Most reports list what was done this month. They don't connect it to a business outcome or explain why it was prioritised over something else. A good report shows a backlog ordered by impact, not a list of tasks ordered by convenience.

Why rankings alone tell you very little

A position is not a result

A rank tracker tells you where one page appeared, for one keyword, on one device, from one location, at one moment. It does not tell you whether anybody searched that phrase, clicked your listing, or did anything useful after landing. A page can sit at position six and carry most of a business's enquiries, because the phrase was commercial and the page answered it properly. Another can climb into the top three and produce nothing, because the phrase was informational. Position is an input. Enquiries are the output. A report that stops at position stopped early.

Impressions and clicks are not enquiries

Impressions tell you how often you were shown; clicks tell you how often somebody chose you. Neither tells you what happened next. The pairing worth reading is the trend in clicks against the trend in enquiries over the same period. If both rise, the work is compounding. If clicks rise while enquiries stay flat, something between the search result and the form is leaking: the page is slow, the offer is unclear, or the traffic was never human. If clicks fall but enquiries hold, you lost the wrong kind of visit. None of that is visible in a rankings table.

What a useful report actually contains

Which keywords actually converted

Not which keywords ranked — which ones led to an enquiry, a quote request, a call or a sale. That means joining search data to the form submissions, calls and chat conversations you already have, and reporting the few phrases that carried commercial weight. Most reports list the top twenty queries by impressions. I would rather see the ten that produced contact, with the landing page each visitor arrived on. Where that join is not possible, I say so and we instrument it.

Indexing, crawl and page-level detail

Two things go missing here. The first is what Google could not reach: pages blocked in robots.txt, URLs marked noindex by accident, soft 404s, orphan pages nothing links to, sitemap errors, and pages that dropped out of the index. A dropped page is invisible everywhere else. The second is page-level performance — not which pages get the most visits, but which earn the enquiries. A contact page with a handful of visits and a steady trickle of submissions beats a blog post with thousands of readers and no enquiries.

What padding looks like

Activity dressed up as outcome

Padding is what fills a report when there is nothing much to say. It looks like a page of green arrows pointing at activity: posts published, links built, fixes closed, directories submitted, words written, tasks ticked off. Every one may be real work, but none is a result. The test is to cover the activity columns and ask what is left. If the answer is "not much", you are paying for a timesheet with a logo on it. Activity is what you buy, and it earns its place only when the report ties it to an outcome.

Numbers that move while nothing else does

Watch for the metric that only appears inside the report. Total keywords tracked is the usual one: it can grow while clicks and enquiries fall, because it counts phrases you rank for somewhere, including phrases nobody searches. Average position is another — it improves when one long-tail phrase jumps to number one while your commercial terms slide backwards. Third-party authority scores are tool opinions, not Google metrics, and rise on link volume alone. One number tells you whether the month was good, and it lives in your inbox: how many people asked to buy something.

How to read a report in five minutes

The order I read it in

• Enquiries first. How many real enquiries arrived, and from where? Find it in your own records first.
• Clicks and impressions next. Read the period-on-period trend, not the total. Rising clicks with flat enquiries is a conversion problem.
• Converting keywords and pages. A short list, with the page each one landed on. If it is long, ask which ten produced contact.
• Work done versus work reported. What actually changed on the site, and how much of it was aimed at the numbers above?
• Indexing and crawl movement. Added, lost, blocked or broken pages. Expensive problems hide here.
• What happens next, and why. One or two priorities with reasons, not a task list.

The question to ask before you close the tab

"What did we learn this month that changes what we do next month?" A report that cannot answer that is a statement, not a decision document. Suppose a report shows 4,000 impressions and 120 clicks this month, against 3,200 and 95 last month. That looks like growth. Then the enquiry line says nine, unchanged. Last month that was roughly one enquiry for every ten clicks; this month, one for every thirteen. The honest conclusion is not "traffic is up" — the extra traffic landed on pages that do not convert, so next month's work belongs there, not in chasing more traffic.

What to ask your provider

Questions they should answer without flinching

• Which keywords produced enquiries this month, not just traffic?
• Is this traffic human? Show me how bot and junk traffic is filtered before it reaches these numbers.
• Did anything drop out of the index, get blocked, or start returning errors?
• Which pages earned the most contact, and which earned the most visits?
• What did we do this month, and what did we deliberately not do?
• What is next month's priority, and what would make you change it?
• Can I see the raw data — Search Console, analytics, call records, form submissions — not just your summary of it?

A checklist you can run on any monthly report

• Real enquiries lead the report, in a number I recognise from my own records.
• Impressions and clicks are shown as a trend, beside that enquiry trend.
• At least one named keyword and one named page are credited with producing contact.
• The report says what was done, what was skipped, and why.
• Indexing and crawl changes are reported, including anything lost.
• Page-level performance separates visitors from enquirers.
• Bot and low-quality traffic is filtered, or the report explains why not.
• Next month's priorities are decisions with reasons, not a task list.
• The underlying data is available on request.

If a report passes that list, the provider is probably doing honest work, whatever the rankings say. If it fails three or more, the numbers in it are decoration.

---

We don't claim our reporting is perfect. But we start from the premise that if the data is polluted, the decisions built on it are wrong. Clean the data first. Then measure what matters.

See what an honest SEO report looks like →