Prioritizing Fixes
Prioritizing fixes is turning a long list of findings into an order of work. A crawler easily finds hundreds of remarks, and without an order a team either drowns in trifles or fixes what will change nothing. There are two working tools here: four levels of severity and an "impact to effort" estimate. This is practice, not a Google rule, so it is important to separate two things: what Google documents as a reason to rank a task higher and what is merely convenient for planning. Figures such as "ten problems give 80% of the result" do not appear in the documentation, and cannot be promised to a client.
The anchor pages are Google's technical requirements and the page on core updates and your website. How to assemble the list of problems is in the lessons on the audit system and the crawling and indexing audit.
What in the documentation justifies high priority
- Eligibility to be shown. A page can be indexed if it meets the technical requirements: Googlebot isn't blocked, a 200 response, indexable content. While an important page does not meet them, nothing else on it matters;
- Policy violations. A page or a site may rank lower or not appear at all; after human review a manual action is possible. See the lesson on manual actions;
- Meaningful, sustainable changes. Google advises avoiding "quick fix" changes and making changes that make sense for users and are sustainable in the long term;
- Small drops. With a small drop in position (for example from 2 to 4) no drastic action is needed; better not to change content that is already performing well.
Four levels of severity (practice)
| Level | Criterion | Examples |
|---|---|---|
| 1. Critical | a page does not meet the technical requirements or the whole site's display is at risk | noindex or a block on important pages, persistent 5xx errors, a hack, a manual action |
| 2. High | a risk of policy violation or noticeable loss of impressions on important pages | hidden text, unlabeled paid links, mass duplicates, 404 on needed pages |
| 3. Medium | affects how Google understands and shows pages | canonical and redirects, redirect chains, titles, duplicate descriptions, structured data |
| 4. Low | minor improvements | minor technical remarks, formatting |
Two rows of the familiar matrix need moving. A missing meta description is not an error and not a "medium priority" by itself: Google mostly takes the snippet from the page content and uses the description when it describes the page better. The problem is identical descriptions on all pages, but even that depends on the site. See the lesson on the meta description. And alt attributes are not always "low priority": Google names descriptive filenames, titles and alt text among the ways to understand an image's subject, and it is an accessibility matter too. For a site that lives on image search it is no trifle. See the lesson on image SEO.
The "impact to effort" estimate (practice)
For each task give two scores from 1 to 5:
- Impact. How many important pages are affected and what will happen to them if fixed. Rely on data: clicks and impressions of the pages the problem affects (see the lesson on SEO KPIs);
- Effort. The time and resources of development and content.
The ratio "impact ÷ effort" then suggests the order within one level. It is a convenient scale, not a calculation: scores are subjective, and Google does not promise a fix will give predictable growth. So the critical level always goes first regardless of "effort": removing noindex from an important page is, for example, the cheapest and most important thing.
The Pareto principle and "80% of the result"
The rule "20% of problems cause 80% of the loss" is often repeated in audits, but it is not in Google's documentation. It may be true of your site or not: check by data, not by a formula. Nor can you promise a timeline like "we'll fix the critical ten in a week and that will give 80% of the result": Google writes that some changes can take effect in a few days, but it could take several months for the systems to confirm lasting improvement, and there is no guarantee of a noticeable effect.
How to put a plan together
- Record everything found without fixing as you go;
- Group by cause, not by address: one template error can explain hundreds of crawler remarks. Such templates are found with the site crawler and the overall picture with the SEO audit;
- Assign a level and scores, giving the source: what confirms it is a problem, for example a Search Console report or a documentation page;
- For each task write the expected result and how to verify it: which metric to compare and over what period. It is hard to tell whether a specific change led to growth, so compare with similar periods; the two-period comparison helps;
- Fix in batches and record the dates, so changes can later be tied to the trend. If a dip occurs, analyze the causes by the lesson on traffic drops.
What is not in the documentation
- The rule "20% of problems, 80% of the result" and promises of timeline and effect from fixing "critical" problems;
- A universal severity scale: the levels above are practice;
- An indication that a missing meta description or alt text belongs to a particular level;
- The formula "impact ÷ effort" as a Google method.
Checklist
- First the problems that stop important pages meeting the technical requirements are closed.
- Then policy risks and manual actions.
- Tasks are grouped by cause, not by address.
- Each task has a source, impact and effort scores, an expected result and a way to verify it.
- Changes are meaningful and sustainable, not "quick fixes".
- The client was not promised percentages or timelines of effect.
- The dates of changes are recorded.
Mistakes
- Fixing trifles while important pages are closed to indexing.
- Treating every crawler recommendation as an error.
- Promising "80% of the result in a week".
- Fixing every address separately instead of one cause.
- Not recording dates and not comparing with a similar period.
What comes next
Next is the diagnosis of traffic drops: what to do when a dip has already happened.