Collapse 41 redirect chains to one hop
Owner: DevAcceptanceCrawl shows zero chains. No money page returns 404.
WebSEOGrow
Work board · tasks, owners, acceptance
Work board
How to Do SEO Work: Tasks, Owners and Acceptance Criteria
SEO work is not a strategy document. It is a queue of small, specific tasks that move across a board: a redirect chain collapsed, a product template re-titled, a page rewritten around one query, a report someone reads. Each card carries an owner and an acceptance standard, so "done" is a fact rather than an opinion.
No vanity dashboards. A board you can screenshot in a Monday call.
Answer block
What does SEO work actually consist of?
SEO work is a queue of small, verifiable tasks across three lanes: technical fixes, content rewrites and earned links, plus the reporting that proves each one landed. Every task gets one owner, one acceptance criterion and a retest date, so the team ships in order instead of arguing about priorities every Monday.
Three lanes of work, three different skill sets
Teams get stuck when they treat SEO as one job. The board only works if each lane has the right person and a standard that person can actually verify.
Technical fixes
Indexability, redirects, rendering, internal links, page speed, structured data, sitemaps. Small in number, large in leverage, and usually blocked by a release cycle.
OwnerDeveloper, briefed by the SEO leadContent rewrites
One query per page, answer above the fold, proof where the reader expects it. Slower than people plan for, because the writing has to be better than what already ranks.
OwnerWriter with category knowledgeLinks and reporting
Prospecting, a linkable asset, the pitch, the follow-up, then a monthly read of what changed. Unglamorous, repeatable, and the easiest lane to fake.
OwnerOutreach specialist, reviewed by the leadThe board: three columns, and nothing else
Three columns are enough for a small team. More columns create status theatre; fewer columns hide blocked work. Cards move left to right, never backwards without a note.
To do4
Rewrite pump category intro for one query
Owner: ContentAcceptanceQuery in H1, spec table above fold, 3 internal links.
Add product schema to 120 SKUs
Owner: DevAcceptanceZero errors in a structured data test, price and stock valid.
Build a spec comparison page for two models
Owner: ContentAcceptanceEvery claim has a source line. Page linked from both product pages.
In progress3
Merge six duplicate blog posts into one
Owner: Content + DevAcceptanceOne canonical URL, five 301s, sitemap updated, impressions flat after 14 days.
Compress hero images on eight templates
Owner: DevAcceptanceMobile LCP under 2.5s in field data, layout shift under 0.1.
Re-title 32 pages in four batches
Owner: ContentAcceptanceEach title 50 to 60 characters, query first, CTR compared at day 21.
Awaiting retest3
Sitemap rebuilt after migration
Owner: DevAcceptance1,204 URLs, correct lastmod, zero 404s in the index report.
301 map for retired category URLs
Owner: DevAcceptanceEvery old URL resolves to the closest live equivalent.
Internal link pass on the buying guide
Owner: ContentAcceptanceSix contextual links from pages already earning impressions.
The "awaiting retest" column is the one most boards are missing. Work that ships without a retest date is work nobody can prove, and it usually gets re-done from scratch six months later.
How to write an acceptance criterion that survives review
If two people can disagree about whether a card is finished, the card is badly written. Compare these pairs.
Too vague
"Improve the category page."
"Fix site speed."
"Do outreach for the guide."
Verifiable
"Primary query in H1, spec table above the fold, three internal links from pages with impressions."
"Mobile LCP under 2.5s in field data on the eight top templates."
"Ten specific pitches sent, five follow-ups logged, two placements live."
A good criterion names the thing to check, the place to check it, and the threshold. It also sets the retest window, because SEO results arrive after a delay and everyone forgets what was changed by then. Cards that pass this test are the reason a weekly board stays honest.
Who does what, and who signs off
One owner per card, one reviewer per column. The reviewer is not a second owner; they only confirm the criterion.
| Role | Owns | Signs off |
|---|---|---|
| SEO lead | Audit, query map, priority order, briefs | Every card before it enters "to do" |
| Developer | Indexing, redirects, speed, schema, internal links | Technical cards, with a crawl export |
| Writer | Rewrites, headings, proof, internal link suggestions | Content cards, against the query |
| Designer | Templates, above-fold layout, mobile reading order | Layout changes that affect LCP |
| Founder or PM | Budget, release slot, final call on scope | Anything that needs a deploy window |
How to pick the next card
Priority is a rule, not a mood. Score every candidate on three factors and take the highest total.
Factor 1Money at stake: does the page or template carry revenue, or only traffic?
Factor 2Blocking power: does this unblock other work, such as indexing or a template fix?
Factor 3Time to verify: can the result be checked in 14 days, or does it need a quarter?
A redirect fix that unblocks forty product pages usually beats a new article, even though the article feels more like marketing. If two cards score the same, take the one that can be retested sooner, because a fast feedback loop is worth more than a marginally better idea. This is also how a busy team avoids the classic trap of starting five things and finishing none.
The weekly rhythm: fifteen minutes, then work
Four short checkpoints, and nothing scheduled outside them. Meetings are for unblocking, not for status reading.
Three cards in progress is the ceiling for a small team. The fourth card is the one that quietly kills the first three, and no amount of standup discipline fixes an overloaded board. If a card sits in progress for two weeks, split it or kill it; long-lived cards are usually hiding a decision nobody wants to make.
Your board is probably missing a retest column
Send us your domain. We will return a card list with owners and acceptance lines, ready to paste.
Where the work usually starts
Two engagements cover most of the board: an audit that produces the card list, and on-page work that clears the rewrite lane. If your index is a mess, the audit comes first. If your pages are indexed but losing to better answers, the rewrite lane comes first.
Our On-Page SEO Services produce the initial board, and our SEO Audit Services take the rewrite and internal link cards from there. You keep the board either way.
Questions about the board
How many cards should a board hold?
Enough for four to six weeks of work, no more. A hundred-card backlog is a wish list, and it makes prioritisation impossible because nobody can see the whole thing.
Who writes acceptance criteria, the owner or the lead?
The lead writes them before the card starts, the owner can refine them, and neither can change them after the work begins. Changing the standard to match the output is how boards lose credibility.
What happens to cards that fail retest?
They go back to "to do" with a note explaining what the data showed. A failed retest is information, not a mistake, as long as the note is specific enough to prevent a repeat.
Do we need a dedicated SEO person?
A part-time lead plus a developer and a writer is enough for most sites under a few thousand URLs. What you cannot skip is the review step, because unreviewed SEO work is how sites get accidental noindex tags.
How do we stop the board becoming admin?
Keep it to three columns, one weekly review, and no mandatory fields beyond owner and acceptance. The moment updating the board takes longer than doing the work, the board is the problem.
Can a card span two teams?
Yes, but one person must still own the outcome. Shared ownership is the most reliable way to guarantee a card stays in progress for a month.