SEO is a queue, and the queue has owners
Most explanations of SEO inside marketing describe the funnel: awareness at the top, consideration in the middle, conversion at the bottom. Accurate, and completely useless on a Tuesday when three drafts are stuck waiting for someone's approval.
What actually determines whether SEO works inside a company is throughput. How many briefs turn into published, indexed, internally linked pages per month, and how many technical fixes reach production. Everything else — the strategy, the audit, the keyword map — is inventory sitting in a warehouse.
So this page is about the pipeline as an operations problem, which is the version nobody puts in a pitch deck and the version that decides your result.
From keyword brief to published page
Here's the real sequence, with realistic elapsed times for an Indian SMB or mid-market team. The times assume nothing goes wrong, which is optimistic.
| Stage | Who owns it | Realistic elapsed time | What stalls it |
|---|---|---|---|
| Query research and page brief | SEO lead or agency | 2–4 days per cluster | No access to Search Console, analytics or the CRM |
| Draft | Writer, in-house or agency | 3–5 days per page | No subject-matter access, so the writer guesses and the draft reads like it |
| Subject-matter and editorial review | Product, sales or the founder | 2 days if scheduled, 3 weeks if not | A single approver with a full calendar |
| Legal or compliance review | Legal, in regulated sectors | 0 days to 3 weeks | No pre-approved library of claims, so every page is reviewed from scratch |
| Build and on-page QA | Content ops or agency | 1 day | A CMS that can't set titles, canonicals or schema independently |
| Publish, internal links, index request | Whoever owns the CMS | Same day | Nobody assigned, so the page ships orphaned and never gets linked |
| Structural changes — schema, speed, redirects | Engineering | One sprint to never | SEO tickets competing with the product roadmap on the same board |
The engineering dependency nobody budgets for
Roughly a third of technical SEO work cannot be done inside a CMS. Redirect rules, sitemap generation, hreflang, pagination, template-level schema, Core Web Vitals, server response times, robots rules, how JavaScript renders — all of it lives in the codebase.
The structural problem is that SEO tickets arrive as small, unglamorous, individually low-priority items, into a backlog scored on product impact. They never win a sprint on merit, and they never will, because 'add BreadcrumbList schema to the category template' loses every argument against a feature customers asked for.
Three things actually fix this, and none of them are a better-worded recommendation.
- A standing allocation. One engineer-day per sprint, permanently assigned, no negotiation per ticket. Small and constant beats large and hypothetical.
- Batch by template, not by page. One ticket that fixes the schema on 340 product pages is a real piece of work. Three hundred and forty tickets is a reason to ignore you.
- Acceptance criteria a developer can test. 'Improve page speed' is not a ticket. 'LCP under 2.5s on the category template at p75 on mobile' is.
- A named engineer. Not a team, not a queue. One person who knows why the redirect map matters.
Approvals, and how legal quietly kills the calendar
In most Indian SMBs the approver is the founder, and the founder is the bottleneck by design — they're the only person who knows whether a claim is true. That works at four pages a month and collapses at twelve.
In regulated categories the second gate is real and slow. Lending, insurance, healthcare, pharma, education with outcome claims — legal review isn't bureaucracy, it's the thing that stops a page becoming a regulatory problem. But it's usually run badly: every page reviewed from scratch, no memory between reviews, three weeks per round.
The arithmetic is worth stating plainly, because nobody does. A two-week approval loop on a four-article month gives you two published articles. You paid a full retainer either way. Nobody bills you for the delay; the delay just eats the deliverable, and the monthly report shows drafts submitted rather than pages live.
- Build a claims library once. Get the twenty sentences you use repeatedly pre-approved, in writing. Then most pages need no legal review at all.
- Separate fact from opinion in the draft. Mark the claims. Legal reviews the marked lines, not 1,400 words of context.
- Book a standing 30-minute review slot every week. Reviews expand to fill the absence of a calendar entry.
- Set a default-approve rule after five working days for anything with no marked claims. Silence should mean yes, or silence becomes your content strategy.
Who measures, and what gets reported to whom
Three data sources, usually three different owners, and nothing works until they reconcile. Search Console belongs to whoever owns SEO. GA4 belongs to marketing ops or analytics. Lead source belongs to the CRM owner, usually sales ops.
The number that matters is the one furthest downstream: qualified leads from organic search, recorded in the CRM, measured against a baseline that was frozen before any work began. Traffic reports are easy to make look good and easy to inflate with queries nobody buys from.
A cadence that holds up under pressure:
- Weekly, internal: pages shipped, pages confirmed indexed, dev tickets opened and closed. Throughput only. Five minutes.
- Monthly, SEO owner to head of marketing: impressions and average position for the target query set, organic sessions, organic leads in the CRM, what changed and why.
- Quarterly, head of marketing to founder: leads against the frozen baseline, cost per organic lead against cost per paid lead, and what the next quarter's plan is.
- Anything that only ever gets reported as a chart with no decision attached should be deleted from the report. It's costing someone an hour a month to produce.
The three stall points, and the fix for each
Almost every stalled SEO programme we've looked at is stuck in one of these three places, and usually the client and the agency each think it's the other one's fault.
- The brief queue. Briefs get written faster than drafts get approved, so work-in-progress piles up and the agency looks busy while nothing goes live. The fix is a WIP cap: don't commission brief nine until page four is published. It feels like slowing down. It doubles throughput.
- The dev backlog. Technical recommendations become a document that gets praised in a meeting and actioned never. The fix is the standing allocation above. A recommendation that can't be scoped into a testable ticket isn't a recommendation, it's an observation.
- The post-publish void. The page ships and nothing happens to it — no internal links, no promotion, no check that it got indexed, no refresh at month six. The fix is a definition of done: a page isn't finished until at least three relevant existing pages link to it and Search Console confirms it's indexed. Internal linking is the cheapest ranking work available and it's the first thing dropped when a month gets busy.