What Does SEO Do for Your Website?
What does SEO do for your website? It changes four things: what a crawler can reach, whether a page matches the words buyers type, how many pages qualify for any query at all, and how often an impression becomes a visit. Nothing about it is mysterious once you score it. Here is the board, how to read each row, and how often to look.
Four rows. One board. No vanity metrics.
Fig. 1 — illustrative figures, not a forecast or a benchmark.
What does SEO actually do for your website?
SEO does four things: it clears friction so a crawler can read every valuable page, it aligns page copy with the words buyers actually type, it widens the set of pages that can qualify for a query, and it earns the click when a result already shows. Links, speed and markup exist only to move one of those four rows.
Why these four rows and not a ranking report
A ranking is an output. You cannot act on it directly, because it is produced by a system you do not control. The four rows are inputs: every one of them can be inspected on a Tuesday and changed by Friday. That is the whole reason for keeping this kind of board. When a client asks what work is buying, the answer is a line on the board rather than a keyword that moved in an unnamed tool.
There is a second reason. Ranking reports hide the two failures that matter most for a small commercial site: pages that never enter the index, and pages that get impressions but never a click. Both are invisible in a keyword table, and both are fixable. A board that scores crawl access, matching, coverage and click share surfaces those failures in the same frame as your wins, so nobody can pretend that a flat month was bad luck.
Finally, the board keeps strategy honest. If coverage is flat and clicks are flat, adding content is not the answer; the pages that already exist are not converting impressions. If coverage is climbing and clicks are flat, the snippets are the constraint. Each row rules out a different explanation, which is faster than arguing about algorithms.
Reading each row
Crawlable: can the crawler reach the page at all?
Most sites lose points before anyone judges the writing. A page behind a redirect chain, blocked in robots.txt, rendered only after a script that times out, or buried nine clicks from the homepage is invisible however good it is. This row counts the share of commercially important URLs that return a clean status, appear in the sitemap, and are linked from at least one other page. When it climbs, every other row gets cheaper to move.
Match: does the page speak the query?
Search engines do not rank pages, they rank answers. This row asks a blunt question: when a buyer types the problem in their own words, does one of your pages contain that phrasing, an answer near the top, and the proof around it? Page-level work lives here. A page can be perfectly crawlable and still score near zero because it talks about the company instead of the question.
Coverage: how many pages qualify at all?
Coverage counts pages that earn at least one impression in a month. A site with two hundred pages and twelve earning impressions does not have a traffic problem, it has a coverage problem. Every page that answers a distinct query adds a row; every thin page repeating a neighbour dilutes one. The healthy shape is a wide base of small, steady pages plus a short head that carries most clicks.
Clicks: does the impression become a visit?
Clicks are the only row a buyer ever sees, and the last to move. They rise when impressions rise and when the snippet earns the tap: a title that names the task, a description that answers before the click, a stable-looking URL, structured data that adds context. Track click-through rate separately from position, because a page can hold position four and still lose half its clicks to a weaker snippet above it.
What each number actually counts
Definitions matter more than dashboards. Two people can look at the same week and disagree about progress simply because one counts URLs and the other counts pages with traffic. The table below is the version we use in a kickoff call, written so that a marketing lead and a developer read the same meaning.
| Row | What it counts | Where to read it | Direction that helps | Typical lag |
|---|---|---|---|---|
| Crawlable | Share of priority URLs returning a clean status, listed in the sitemap, and internally linked | Crawl report, sitemap coverage, server log sample | Up, with no new orphan pages | 2 to 6 weeks |
| Match | Share of target pages whose opening block answers the query in the buyer's own words | Manual page review against a query list | Up, page by page | 4 to 8 weeks |
| Coverage | Pages earning at least one impression in a 30-day window | Search performance report, page dimension | Up, then stable | 1 to 2 quarters |
| Clicks | Visits from organic results, per page and per query group | Search performance report, query dimension | Up, with click-through stable or rising | 2 to 3 quarters |
| Click-through | Clicks divided by impressions for a page and query pair | Same report, segmented by page | Up after a snippet rewrite | 2 to 4 weeks |
Illustrative only. Figures describe how the rows are counted on a typical engagement, not results you should expect.
How to read the board without fooling yourself
Comparing a 30-day window to a 7-day window
Half the panic in SEO reporting comes from mismatched windows. A strong week against a weak month looks like a collapse. Lock every row to the same window and compare like with like, or compare four-week blocks and ignore the rest.
Reading an average instead of a distribution
A mean click count per page hides the truth on most sites: a handful of pages carry everything. Read the median and the top ten pages separately. If the median is zero, you have a coverage problem wearing a traffic costume.
Counting brand queries as demand
People who already know your name are not new demand. Split branded and non-branded queries before celebrating. A click row that grows only on brand terms means the board is measuring your reputation, not your reach.
Treating one dip as a verdict
Releases, seasonality and reporting changes all move rows. Give a dip two full weeks and one explanation before you change strategy. Most dips are noise; the ones that are not will still be there in a fortnight.
The three-month checkpoint
Ninety days is long enough for the first three rows to show a direction and short enough to keep a team honest. We review at three fixed points rather than whenever a report lands, so each conversation has one job and one decision.
Day 30 — hygiene and baseline
Fix status codes, redirect chains, sitemap drift and orphan pages. Freeze the four numbers as a baseline so later months are compared to a fixed line, not to last week.
Day 60 — match on money pages
Rewrite titles, opening blocks and subheads on the pages closest to revenue. This is where page-level work compounds fastest, and where a rewritten opening paragraph can lift click-through before coverage moves at all.
Day 90 — coverage or consolidation
Decide one direction: publish the query gaps that showed up in the data, or merge the thin pages that keep splitting one topic. Choosing both at once is how coverage stalls.
Where we start on a new site
We begin with a crawl and a query inventory, because both are cheap and both stop guesswork. Then we take the pages that already have impressions and improve their match before we add a single new URL. If the word itself is still fuzzy for your team, start with Whats Is SEO? so everyone shares one definition before a kickoff argument about tactics.
Page-level work — titles, openings, internal links, structured data — is the fastest lever on the board, and it is what our On-Page SEO Services exist to do at template scale rather than one page at a time. After those two rows stabilise, coverage expansion earns the right to be called a strategy instead of a content quota.
A crawl of every URL, a priority list of ten pages, and the four numbers written down as a baseline.
A position, a ranking date, or a click count. The board moves when the work is right, and we show you the row it moved.
Questions we hear before the first score
How long before the scoreboard shows movement?
Crawl and match rows can move within four to six weeks when the fix is on-site. Coverage takes one to two quarters. Click rows are the slowest because they depend on the first three, and they are also the noisiest week to week.
Do I need more traffic or better pages first?
Check the coverage row first. If fewer than one in five of your URLs earns an impression, new pages will not help. Fix matching on the pages you already own, then expand the base.
Which row should a small site watch closest?
Clicks per page, not total clicks. A ten-page site that lifts average clicks per page from two to six has done more than a blog that adds twenty posts nobody reads.
Can these scores go down on purpose?
Yes. Removing thin pages lowers coverage and often raises clicks per remaining page. A cleanup month that shrinks the board can be the healthiest result in the quarter.
Do you report keyword rankings too?
Yes, as context rather than as a target. Rankings explain why a click row moved, but a position without a click pays nothing, so they are never the number we optimise.
What do you need from us to start?
Read access to search performance data, a crawl, and one person who can approve page changes. If analytics is thin, we install a baseline before touching content, because an unmeasured fix is only a guess.
Want your own board before you commit to anything?
Send the domain. We will return four scores with the pages behind each one, and tell you which row is cheapest to move first.