Pull the queries
Export the queries with impressions and no clicks. Pick the two closest to a commercial phrase.
Owner: contentYou work SEO the way you work a shift: five fixed tasks, one owner each, and a written acceptance you can check on Friday. Nothing heroic, nothing mysterious. The reason most SEO fails is not talent 鈥?it is that the work happens once, in a burst, and then stops for four months while everyone returns to shipping product. A board with names on it fixes that.
No dashboards to babysit. The board is the dashboard.
Eight hours a week. Same five slots, every week, owned by name.
It requires five repeatable shifts: read the queries, edit one page deeply, fix one technical item, add one piece of proof, and record what changed. Eight hours total, assigned by name. If a week passes with no shift worked, the ranking does not punish you immediately 鈥?it waits, then charges interest.
Each card below is one shift. The owner is a role, not a person's name, so the board survives holidays and resignations. The acceptance line is what Friday checks 鈥?if it cannot be checked, the shift was not worked.
Export the queries with impressions and no clicks. Pick the two closest to a commercial phrase.
Owner: contentOne page, not five. Add the missing section, the specification, the comparison the buyer asked for.
Owner: contentOne item from the fix list: status code, canonical, orphan link, image weight, schema error.
Owner: technicalOne photograph, test result, tolerance table, certificate, or a paragraph showing who did the work.
Owner: contentDid the two URLs move? Write one line in the log. No meetings, no deck.
Owner: decisionTwo things about this board are deliberate. Tuesday is the only day that writes, because writing is the expensive shift and it needs a clear brief from Monday. Friday is one hour, because measurement that costs a day becomes a project of its own and dies.
The board fails when one person is asked to be all three roles at once, which is what "we do SEO internally" usually means in a small export company. Split them and the shifts take eight hours instead of twenty.
| Role | Owns | Never does | Time per week |
|---|---|---|---|
| Content | Query review, the one deep edit, proof collection. | Decides which market to enter. | About four hours. |
| Technical | Crawl, index, speed, schema, redirects, backups. | Writes buyer-facing copy. | About two hours. |
| Decision | Chooses the market, approves the schedule, reads the log. | Rewrites pages on a hunch. | About one hour. |
Which country, which product family, and which month. Three decisions a quarter, not thirty. A team that keeps re-choosing its market never accumulates anything, because every shift starts over on a new target.
The board assumes the site can be crawled and rendered. If that is false, every shift writes into a hole. One weekly check on coverage and one on template speed is enough to catch a regression before a quarter is lost.
"We updated some pages" is not acceptance. Acceptance is a sentence a stranger could verify without asking you. Four tests we use with client teams:
If the note cannot point to a live page, the shift is a conversation, not work.
Impressions, clicks, enquiries, or milliseconds. One number, recorded before and after.
Undated work cannot be compared with an algorithm update or a release. Dates are the only defence against superstition.
If the change cannot be undone, it needed a second signature. That rule prevents most panic edits.
A board makes work visible; it does not make a weak offer strong. If the product page cannot answer a buyer's first three questions, eight hours a week will produce impressions and no enquiries. The honest use of the board is to find that out in six weeks rather than six months, and then change the offer or the market rather than the copy.
It also cannot substitute for a one-off deep audit. The board maintains a site; an audit finds what the board has been unknowingly working around. That is why Whats Is SEO? exists as a separate step before the shifts begin.
A weekly SEO meeting has one job: choose next week's two URLs. Everything else is status theatre. Three questions get you there.
If the first question has no answer, the property is not receiving impressions and the problem is upstream 鈥?indexation, market choice, or the offer. Do not proceed to question two. Teams that keep writing pages while question one is unanswerable are paying for production, not progress.
One page in a shared document, five columns: date, URL, query, change, number. That log becomes the most valuable asset the team owns, because in a year it is the only honest record of what this site responded to. It is also the first thing a new agency should read, and the first thing we ask for.
Some clients want us to run all five shifts. More often we take the technical and the Tuesday edit, and the client keeps Monday and Friday because those need internal context nobody outside can supply. Either way the board looks the same, and the log stays in your hands.
We start by reading what already exists, then we take one shift and prove it for four weeks before expanding. If you want the underlying vocabulary before the board starts, SEO Audit Checklist covers the four moves in plain language. Teams that understand the loop write better briefs, which makes Tuesday cheaper.
A furniture components supplier ran the same five shifts for twelve weeks with two people. No new pages were published in the first month; all work went into six existing product pages and one broken template. By week nine, impressions on those six pages had roughly doubled and two buyers had mentioned a specification table in their enquiry. The template fix was the shift nobody wanted to do, and it produced most of the movement.
Send us the site and the team size. We will tell you which shifts to keep in-house.
Eight is enough for one market and one product family. Under four, the shifts never complete a cycle. Above twelve, most small teams are producing pages faster than they can prove them.
Whoever can change the plan. If the person reading the log cannot cancel a shift or move a market, the log becomes decoration and the board quietly dies.
Only if the tracker is empty. It rarely is. Speed, canonicals, and broken links accumulate quietly, and they are the cheapest insurance on the board.
Then outsource Tuesday and keep Monday, Thursday, and Friday inside. The brief from Monday is what makes an outside writer useful rather than generic.
A technical fix can show within days. A page edit usually takes three to six weeks to appear in impressions. Judge the board by the log, not by a single week.
Yes, with one change: Thursday's proof becomes a case note or a named engineer rather than a specification table. The five slots stay the same.