Sit down in front of it once. The Performance tab is open, the query table is glowing, and four big numbers run across the top. Clicks down 11% this month. Average position 6.2. The operator's eye goes straight to position, reads "6.2" as the verdict, and concludes the restaurant is losing the ranking war. That reading is wrong, and it's the most expensive misread in the dashboard — it sends a Tuesday afternoon chasing a number that moves over months, while the number that moves this week sits two rows up, untouched.

So let's not read it that way. Let's tear the report down row by row — what each metric actually is, the misreading an operator reaches for first, and the one move each row should trigger.

Google Search Console is free, official, and tells you exactly which queries Google shows your restaurant for, how often visitors click through, and which pages are working. The interface is the problem. It's built for SEO professionals; the labels assume terminology a restaurant operator doesn't have. So we'll supply the terminology — and read the artifact in front of us instead of the marketing tour of it.

You're thinking: I already have Google Analytics. Yes. And it tells you what they did once they got there. Search Console is what they typed before they got there. Different question.

This is the plain-English version. Four reports do the work. Skip the rest.

Set it up first (10 minutes, one time)

Go to search.google.com/search-console, sign in, add your domain. Verify ownership via DNS or via a meta tag your developer can drop in (most modern hosts have a one-click verification). Wait 24-48 hours for the data to populate.

If you've already been verified for a year and never logged back in, your account already has months of data — it's been collecting in the background.

Four reports, four cadences — Performance weekly, Queries monthly, Pages quarterly, Coverage only when the alert fires. Everything else in the sidebar is ignorable.

Report 1: Performance — the headline numbers

This is the only screen most operators ever need. Four numbers across the top:

  • Total clicks — how many people actually clicked through to your site from Google search.
  • Total impressions — how many times your site appeared in any search result, whether anyone clicked or not.
  • Average CTR — clicks ÷ impressions. The percentage of times your snippet won the click.
  • Average position — your average rank across all queries that showed your site.

Now read them in the order the math runs, not the order the screen lists them. Walk the four rows top to bottom the way you'd walk a line check — one station at a time, asking what each one is telling you and what it is not:

  1. Total clicks (the row your eye lands on first). Misread: "clicks are down, the site is broken." Correct read: clicks are an output, never a cause. This row tells you a result, not a reason. Don't act on it — use it to decide whether to walk the other three rows at all. A flat clicks row means close the tab.
  2. Total impressions. Misread: skipped entirely, because it has no obvious action attached. Correct read: this is the row that tells you whether a clicks drop is a demand problem or a snippet problem. If impressions held steady while clicks fell, Google still showed you just as often — people just stopped clicking. That points the finger one row down.
  3. Average CTR — the only row you can move this week. Misread: treated as a fixed property of the site, like square footage. Correct read: CTR is the share of appearances your snippet converts, and it's written by the title and meta description — text you can rewrite this afternoon. A branded-query CTR (your own name) below 40% is the loudest single signal in the report: the snippet is broken.
  4. Average position (the row your eye lands on second, and over-weights). Misread: "6.2 is bad, we're losing." Correct read: position is the slowest number on the screen — it moves over months, and a #4 with a sharp snippet out-earns a #2 with a generic one. It sets a ceiling on impressions; it does not set your clicks. It's the row to note and the last row to act on.

The figure below is the centerpiece of the whole teardown. It's the same four numbers, re-stacked into causal order — the order the report would print them in if it cared whether you understood it. Read it once and the misread above stops being available to you:

  1. 1

    Average position

    Where you rank decides how often you’re shown at all. Slow to move — months, not weeks.

  2. 2

    Total impressions

    Every appearance position earned, whether anyone clicked or not.

  3. 3

    Average CTR — the lever

    Clicks ÷ impressions: the share of appearances your snippet converts. Branded CTR below 40% = broken snippet.

  4. 4

    Total clicks — the output

    Impressions × CTR. Nothing moves this number except the two steps above it.

Not the dashboard order — the causal order. When clicks drop, walk the chain backwards: impressions held but clicks fell means the leak is at step 3, and the fix is a snippet rewrite, not a ranking campaign.

The dashboard lists the four numbers in the order that panics you. The math runs them in the order that fixes you. The teardown is just learning to read the second order off a screen printed in the first.

The leverage is almost always in CTR, not in position. A restaurant ranked #4 with a strong title + meta description outperforms a restaurant ranked #2 with a generic one. If your CTR for branded queries (your restaurant name) is below 40%, the snippet is broken — almost always the meta description. The pattern shows up sharpest when branded-query CTR is healthy (a customer searching for you by name finds you and clicks) but the cuisine-query CTR is in the low single digits. Same site. Same domain authority. Two different snippets — one written by a human who knows the room, one autogenerated from the page title.

The number to anchor on

Branded-query CTR below 40% is the single loudest reading in the whole report — a customer typed your name and still didn't click the result with your name on it. That threshold reflects observed median patterns, not a controlled study (see Sources), so treat it as a tripwire rather than a target. When it trips, the fix lives in the meta description, not the ranking. Everything else in this teardown is a footnote to that one row.

Put the two readings side by side. Same screen, same four numbers — the only thing that changes is which column you're reading from:

The misreading

Eye lands on clicks down, then on position 6.2, and stops. Verdict: "we're falling in the rankings." The Tuesday gets spent on a ranking campaign — backlinks, keyword stuffing, a call to someone who "does SEO" — chasing a number that moves over months. Impressions and CTR go unread. The snippet never gets touched.

The correct reading

Start at impressions: did Google still show you as often? If yes, this isn't a ranking problem — it's a clicking problem, and that's CTR, and CTR is the title and meta you can rewrite this afternoon. Position is noted and set aside as the slowest, ceiling-setting number. The Tuesday gets spent rewriting one snippet, and the clicks row answers next week.

Report 2: Queries — what people typed to find you

Below the headline numbers, click "Queries." This is the gold. Sort by impressions descending. The first 20-30 rows tell you exactly what Google thinks your restaurant is for.

What to look for:

  • High-impression / low-CTR queries. These are queries Google thinks you should rank for but visitors aren't clicking. Almost always a snippet problem — the title or meta doesn't match what they expected.
  • Surprise queries. Phrases like "best happy hour brookline" that you didn't deliberately target — those are clues for content you should add.
  • Missing queries. If "tacos al pastor near me" doesn't appear at all, your menu schema is missing or broken.
  1. 1

    High-impression, low-CTR queries

    Google thinks you should rank for them but visitors aren’t clicking. Snippet problem — rewrite the title and meta.

  2. 2

    Surprise queries

    Phrases you didn’t target but Google has decided you might be for. Add a small section to meet them halfway.

  3. 3

    Missing queries

    Searches you’d expect that don’t appear at all. Usually a menu-schema gap — the dish names aren’t in real text.

Three patterns to scan for in Queries, in the order they tend to surface. The fix is different for each — one rewrites a snippet, one adds a paragraph, one rebuilds a menu.

The category of surprise queries worth watching closely: phrases Google has decided you might be relevant for, where you're already at page two without having tried. Adding a small section to the page — not a full pivot, just enough to acknowledge the query's intent — frequently moves position into clickable territory within a few weeks. Most of the time the gold isn't the queries you're trying to rank for; it's the queries Google has decided you might be for and you haven't met halfway.

Run the SEO Grader on the page that's getting the impressions but no clicks — it'll surface the title + meta issues directly.

The Queries report is the only place a customer's actual words land in your dashboard. Read it like a server reads a comment card.

Report 3: Pages — which pages are doing the work

Switch the tab from "Queries" to "Pages." Sort by clicks descending. The pattern most restaurants see:

  • The homepage gets 60-70% of clicks.
  • The menu page should get 15-25%.
  • The reservation page typically gets 3-8%.
  • Everything else is rounding error.

Healthy click distribution by page

Homepage

60–70%

Menu page

15–25%

Reservation page

3–8%

The shape of a healthy Pages report. Bars use the midpoint of each range. If your menu page sits below ten percent, the menu page is the issue — usually a PDF, or buried in a subdirectory that ranks badly.

If your menu page is below 10% of total clicks, you have a menu-page problem — usually it's PDF (invisible to Google) or hidden behind a subdirectory that ranks badly. The fix is converting to a real HTML menu.

Report 4: Coverage / Pages indexed — the silent killer

In the left sidebar, under "Indexing," click "Pages." This tells you how many of your pages Google has actually added to its index. If 50% of your pages are "Excluded" or "Discovered — not indexed," you have a foundational problem — usually slow load, broken canonical tags, or thin content.

Run Restaurant Audit if this number looks bad. The audit's technical-SEO section diagnoses the most common indexing blockers.

Three Coverage red flags · what each one actually means before you go fix it

The three weekly habits

Search Console cadence · under 20 minutes a month
  1. Mondays — open Performance, look at week-over-week. Two minutes. If clicks dropped 20%+ week-over-week, something broke (Google update, site change, holiday weirdness). Investigate. Otherwise, close the tab.
  2. First-of-month — sort Queries by impressions. Five minutes. Note any new queries appearing in the top 30. Decide whether to add content for them.
  3. Quarterly — Pages report, sort by clicks. Ten minutes. The page distribution should match your business's actual conversion path. If it doesn't, you have content-architecture work to do.

Total weekly burden: under 20 minutes per month. Result: actionable signal you can't get anywhere else, including from any agency that "manages your SEO."

What Search Console doesn't tell you

Search Console only covers Google search clicks. It doesn't tell you about Google Maps clicks, Find a Table bookings, or AI-search citations from ChatGPT and Perplexity (those don't pass referrer headers). For the AI-search side, set up the FAQ schema mentioned in the AI article and watch your first-touch attribution in Plausible — that's where AI-driven visits show up.

Save the four-number screenshot to your Workshop monthly. The trend over six months tells the story; the single weekly check tells you when to investigate.

Position is what you can't easily move. CTR is what you can move on a Tuesday afternoon.

Tell us

Be the first field note on this piece.

Tried this in your own restaurant? 100–400 words, your name on it. Don reads every one. Your note shows up here once approved.

Sources & further reading

Google Search Central — Search Console documentation

Google Search Central — The four reports highlighted (Performance, Queries, Pages, Coverage) are Google's own canonical organization of Search Console. The plain-English glosses in this post follow Google's public documentation for each report.

Advanced Web Ranking — CTR by position research

Advanced Web Ranking — AWR's annual CTR-by-position studies consistently find that snippet quality (title + meta) explains 30-50% of the variance in CTR within a single position. The "leverage is in CTR, not position" framing reflects this body of research.

Mozcast / SERP feature studies

Mozcast / SERP feature studies — Mozcast's ongoing tracking of Google SERP feature appearance shows that branded queries with healthy CTR almost always have a tailored meta description, while underperforming branded queries are correlated with autogenerated meta. The 40% branded-CTR threshold in this post reflects observed median patterns, not a single controlled study.