Search · the written read

The honest written read on your store.

eight subsystems · your own data · every section allowed to say “not enough evidence”
  • It reads eight things about your store — your Search Console connection, your traffic trend, conflicts, declines, splits, technical access, your content corpus, and what past fixes actually did.
  • Each section can fail on its own — one dead subsystem returns an error for its own section and the report says so, rather than the whole audit collapsing or quietly omitting it.
  • It is allowed to reach no conclusion — and does. Sections routinely say the history is too short or a check was blocked, instead of manufacturing a finding.
  • Findings come with the evidence attached — the term, the volume, the position, the impressions, and every URL involved. Not “improve your internal linking”.
  • It won’t promote a fix it can’t support — where the data lacks what an action would need, the recommendation stays in prose instead of becoming a numbered item.
the opening of a real report

Executive summary

a specialty retailer · production report

This store does not look like it has a clear technical SEO failure. Search Console is connected correctly, the live host matches the property, robots rules are OK, and the sitemap points to the right host. The main problem in the evidence is internal keyword overlap: too many collections, products, pages, and blog posts are competing for the same commercial searches.

Verdict

Not a technical problem

Note what it did there. The obvious thing to sell a merchant is a technical crisis. This report opens by ruling one out.

What it reads

Eight subsystems. Any of them may return nothing.

Each section is gathered independently and wrapped so that a failure inside it produces an error for that section alone. One broken subsystem never sinks the report — and never quietly disappears from it either.

  1. 01

    Connection

    Is Search Console connected, current, and pointed at the host that is actually live?

  2. 02

    Traffic trend

    Month over month and against last year, with the verdict stated separately from the rows.

  3. 03

    Conflicts

    Your own pages competing for one search — joined with any verdicts you’ve already given.

  4. 04

    Declines

    Pages losing clicks, with a mechanical liveness check before any explanation.

  5. 05

    Splits

    One topic spread across several URLs. Your own “these are the same product” calls override the analysis.

  6. 06

    Technical access

    Robots, sitemap, host redirects, and how the site answers common search crawlers.

  7. 07

    Corpus

    How much content exists, where it came from, how much is enriched, how fresh it is.

  8. 08

    Receipts

    What previous fixes measured. On a first report this is usually empty, and it says so.

The part worth paying for

Three things a real report
refused to say.

These are verbatim from one production audit. An SEO tool that always finds something is not diagnosing your store; it is describing its own incentives. Every one of these could have been dressed up as a finding.

  • On declines

    “There are no decline rows yet. Decline comparisons are still accruing, history depth is 4 days, and full 28-day comparisons will be available from 10 August. So I cannot honestly name page-level losers from this subsystem yet.

    It could have called any wobble a decline. Four days is not twenty-eight.

  • On technical access

    “What we could not verify automatically is how the site responds to common search bots, because the site rate-limited automated requests. That is not evidence of a problem; it just means this subsystem could not prove those items either way.

    A blocked check is not a failed check. Most audits sell you the difference.

  • On its own recommendations

    “This section does not supply full URLs for those duplicate pairs, so I am keeping those merge candidates in narrative form rather than promoting them to line items.

    It found something real, and declined to turn it into an action it couldn’t specify.

A finding

Every item carries its own evidence.

A line item is not advice. It’s a named problem, the numbers that establish it, the URLs involved, what to do, and what you should expect if you do it. This is one, real, from the same report.

fix-003cannibalization

Stop splitting one broad category term across collections, products, and blogs

Evidence

Search volume
3,600
Best position
39.8
Impressions
1,202
Ranking on
8 pages

Pages involved

  • /collections/all-espresso-machines
  • /collections/<a sub-category>
  • /products/<a machine>
  • /products/<another machine>
  • /blogs/buying-guide/how-to-choose-a-machine-for-home
  • + 3 more

Recommendation

Assign one broad espresso-machine landing page to own this head term, and pull exact-match targeting away from model pages and editorial guides. The guides should support buying intent; they should not compete for the same generic phrase.

Expected impact

This is one of the clearest ways to improve visibility on a very large head term that is currently diluted across mixed intents.

Read the last block again. It says “one of the clearest ways to improve visibility” — not “this will win you the term”, and not a projected traffic number. A report that quantifies an outcome it cannot know is telling you something about itself.

Honest by construction

The report can’t hide its own gaps.

Each of the eight sections is gathered in isolation and wrapped so that any error inside it becomes that section’s content. There is no path where a subsystem fails and the report simply reads as though that topic were fine.

That is a structural property, not an editorial promise. The narrative is written from whatever the gatherers returned — so when a gatherer returns “blocked” or “not enough history”, that is what there is to write about.

What you get when something goes wrong

A subsystem errors
that section says so
History is too short
the date it becomes available
A check is blocked
stated as unproven, not as a fault
A finding lacks its URLs
stays prose, never a line item
No prior fixes to measure
the receipts section says empty

five ways to say “I don’t know” · all of them used in one real report

What it’s for

The thing you’d pay an agency a month’s retainer to find out.

Most merchants have never had anyone read their whole search position at once. The value isn’t the fixes — those are in the app anyway. It’s the sentence at the top telling you where the problem actually is, so you stop spending money on the place it isn’t.

The report in this page’s examples opens by ruling out a technical failure and pointing at internal overlap instead. That’s a month of misdirected effort avoided, in a paragraph.

  • It reads your whole position at once — not one page, not one term
  • Findings link to the fixes — the actions are already in the app, priced and reversible
  • Run it again later — the receipts section fills in as fixes get measured
  • A full diagnosis

    50 creditsAll eight subsystems, the narrative, and the line items
  • Acting on a finding

    priced per fixFrom 2 credits for links to 25 for a merge — each quoted before it runs
  • Doing it yourself

    freeThe report is yours. Nothing obliges you to buy the fixes from us
Before you ask

The honest answers.

  • What if my store is fine?

    Then the report says so, and that is a legitimate result. The example on this page opens by ruling out the technical failure a merchant might have been braced for.

  • How long does it need to be useful?

    The conflict and split analysis works immediately. The decline analysis needs 28 days of history to say anything honest, and the report will tell you the date it becomes available rather than guessing in the meantime.

  • Do I have to buy the fixes?

    No. Every finding names the pages and says what to do. Plenty of merchants take the report and do the work themselves, and that’s a fine outcome.

  • Is this written by AI?

    The narrative is, from structured evidence the app gathered. The numbers in it are not — they come from your Search Console and your own catalog. That split is the whole design.

  • How is this different from the issues list?

    The issues list is a live queue you work through. The report is a single read of the whole store at one moment, with a narrative that says which of those issues actually matters most.

Most SEO audits find a problem
because they were paid to.

50 credits · eight subsystems · every section allowed to say it doesn’t know