No audit ever finds only a handful of things to do. A competent one finds more than any team has capacity for, so the real product of an SEO engagement is not the list of issues but the order in which they are tackled. That order decides how soon results arrive, how much risk is taken on, and whether the client trusts the plan. It is also where weak engagements show: a list of fifty recommendations with no ranking is a way of avoiding the decision.
Impact: how much could this move?
Impact depends on three things: how many pages are affected, how valuable those pages are, and how directly the issue stands between a page and its performance. A fault that stops pages being crawled or indexed sits at the top, because nothing downstream can work around it. Google’s guide to debugging traffic drops defines technical issues as errors that can prevent Google from crawling, indexing, or serving your pages
, which is a useful test: if a problem can stop a page being served at all, it outranks one that makes a served page a little better.
Effort: what will it cost to do?
Effort is more than development time. It includes content work, design, testing, approvals and the time other teams need to fit it in. A change that is small to write but must pass a release window and a legal review is a bigger piece of work than it looks. Estimates should come from the people who will do the work, not from the SEO who proposed it.
Confidence and risk
Two further questions keep a ranking honest. How sure are you that the change will help? An established fix to a clear fault deserves more confidence than a speculative improvement. And what could go wrong? Changes to URL structure, templates, robots rules or canonical logic have a wide blast radius and need a plan and sign-off before they are scheduled, which is the subject of the next article.
A simple scoring approach
A common way to make the discussion concrete is to score each item from one to five for impact, confidence and effort, then divide impact times confidence by effort. The number is not an oracle. It is a prompt: when two items score similarly, you talk about why, and when the client disagrees with an order, you can point to which assumption they are challenging. Record the scores with the reasons, so the ranking can be revisited when something changes.
The dependencies people forget
The highest-scoring item is not always the right first step. Some work cannot start, or cannot show results, until something else is done.
- Indexability before improvement. Better content on pages that search engines cannot index changes nothing visible. Check that important pages can be crawled and indexed before investing in them.
- A redirect map before a migration. Google’s guidance on site moves with URL changes says that in more complex moves you need a list of the old URLs mapped to their new destinations. It also suggests, where possible, moving a part of the site first to test the effects on traffic and indexing. Both are dependencies, not afterthoughts.
- Duplicate-handling decisions before consolidating content. Google explains that it groups duplicate pages and chooses a canonical version. Deciding which page should be the canonical one has to come before merging, redirecting or rewriting the others.
- A baseline before the change. Without a record of where things stood, a later improvement cannot be shown. Capture the starting figures first, and say what is unavailable.
- Template fixes before page-level work. Editing individual pages that share a faulty template just creates work that the template fix will overwrite or make redundant.
A worked example (hypothetical)
Suppose an audit produces five candidate items for a site with a few thousand pages.
- A template change leaves product category pages with a noindex tag.
- Product pages have duplicate titles.
- Thirty old blog posts could be merged into six stronger guides.
- Internal links to the best-converting pages could be strengthened.
- Alt text is missing on some images.
Scored for impact, confidence and effort, item one comes out clearly first: it blocks a whole section from search, the fix is well understood and the effort is small. It is also a dependency for item four, because internal links to pages that are not indexable do nothing for them. Item three, the content merge, has the biggest potential and the highest effort, and it depends on a decision about which URLs should be canonical, so it needs planning rather than a place in this month’s sprint. Item two is a template fix with moderate impact and low effort. Item five is real but small, and goes at the end of the list.
Notice that the order is not the order in the audit, and not the order of the easiest items. It is the order the dependencies and the scores give.
Talking to the client about trade-offs
A client usually has views on what should come first, often based on what they can see rather than what is blocking performance. The scores help with that conversation, because they turn “I think we should do the blog first” into “the blog depends on pages being indexable, which is item one”. Show the reasoning, ask which assumption they would challenge, and record the decision. If they choose a different order, that is their right, and it should be recorded together with the trade-off they accepted.
When everything is urgent
Teams under pressure often label everything high priority, which makes the label meaningless. A useful discipline is to limit how many items can be in progress at once, and to require that every new urgent item names the item it displaces. This makes the cost of a change of direction visible, and it keeps the team finishing things instead of starting them.
Quick wins that are not quick
The change that looks like the cheapest win, such as a sitewide title tweak, can be one of the riskiest, because it touches every page at once. Treat the size of the blast radius as part of the effort. A change that is easy to make and hard to reverse deserves the same care as a big project.
Keep opportunities and tasks apart
It helps to hold two lists. Opportunities are possibilities with an estimated impact and effort, ranked and revisited. Tasks are commitments with an owner and a date. Mixing them turns the plan into a to-do list nobody can reason about, because a task that was never a good idea sits beside one that is overdue. SEO Ops keeps them separate and lets an opportunity become a task when it is chosen.
Re-prioritise on purpose
The order is a decision at a moment in time. New data, a failed release, a competitor’s move or a change of business priority can all reorder it. When it changes, say so and say why, rather than letting the plan drift. A ranking that moves for stated reasons builds trust; one that moves silently destroys it.