Prioritising SEO work when everything looks urgent
You have a long list of possible SEO work and no defensible way to order it.
Short answer
Rank by expected clicks gained divided by effort. Estimate the gain from impressions you already have and the click-through rate your own data shows at the target position. That makes striking-distance pages and snippet fixes rise to the top and pushes speculative new content down, which is usually the correct order and rarely the intuitive one.
Price the change before you make it
Take the page’s current impressions, its current position, and the click-through rate your own data shows at the position you are targeting. The difference is the expected gain, in clicks. This turns an argument about priorities into arithmetic.
The tier that almost always wins
Pages with substantial impressions, a position in the striking range, and a click-through rate below what that position should return. Demand is confirmed, the position exists, and the fix is a title. Highest return, lowest effort, and consistently under-prioritised because it feels too small.
The tier that looks urgent and is not
Technical findings on pages with no impressions. Fixing the schema on a page nobody searches for produces a tidy green tick and no traffic. Sort your technical backlog by page impressions before working through it and most of it disappears.
New content is a bet, not a fix
A new page has no demand data, no position, and no history. It might be the best thing you do this quarter and you cannot price it in advance. Fund it deliberately as exploration, sized as a portion of your time, rather than letting it crowd out priced work.
Effort estimates should include maintenance
A page is not written once. Every page you add is maintained forever, and a set of pages that each need occasional updating is a recurring cost that never appears in the original estimate.
The one exception to arithmetic
Anything blocking indexing outranks everything, whatever its expected value looks like. A page that cannot be crawled has an expected value of zero regardless of demand, so the unblocking work is a precondition rather than a line item.
Re-rank monthly, not continuously
Positions and impressions move. A list built three months ago is describing a site that no longer exists. Rebuilding it takes an hour and prevents a quarter spent executing a stale plan.
Questions
- What if I cannot estimate the gain?
- Then you are looking at speculative work, which is fine as long as it is labelled that way and capped. The point of pricing is to stop speculative work displacing the priced kind by default.
- Should quick wins always come first?
- Usually, because they compound: the traffic arrives sooner and funds patience for the slower work. The exception is anything blocking indexing, which comes before everything.
- How do I get the CTR curve for my own site?
- Export your queries, bucket them by position, and divide clicks by impressions in each bucket. Filter brand out first or the top bands will be wildly inflated.
Measured, not asserted
The arithmetic is not marginal. On ihatepdf.cv, non-brand pages at positions 1 to 3 returned 29.95% click-through against 1.47% at positions 8 to 10, and 93 of 160 of 160 pages earned under ten clicks. Pricing each candidate fix against that curve is what separated the work that mattered from the work that felt productive.
Free tool for this: Position Value Calculator. No account, nothing uploaded.
Where this goes deeper
Every number on this page comes from one complete dataset: one product taken from zero to 100K+ users on search alone, with nothing spent on advertising. The full argument is Chapter 15 and Chapter 30 of the book. Five chapters are free to read.