How Do I Find SEO Opportunities?

Start with data you already own

Most opportunity hunting starts in a keyword tool, which is the wrong end. A keyword tool tells you what exists; your Search Console data tells you what is already happening to you, which is a much shorter path to something actionable.

The distinction matters because your own data comes with proof attached. A query you already earn impressions for is demonstrated demand and demonstrated relevance. A keyword from a tool is a hypothesis about both.

Five patterns worth looking for

1. Positions 11 to 20. Pages one improvement away from earning clicks. Highest return on effort in SEO, covered in striking distance keywords.
2. High impressions, low CTR. Google shows you and nobody picks you. The fix is the title and description, not the content, and it is the fastest feedback loop available.
3. Queries with no dedicated page. You get impressions from a page that mentions the topic in passing. A page that addresses it properly usually outranks the accidental one.
4. Pages that used to perform. Decline is easier to reverse than growth is to create, because the page already proved it could rank.
5. Competitor gaps. Queries they rank for and you do not. The only pattern here that genuinely needs a paid tool.

Note that four of the five need nothing but Search Console, and the fifth is the one most tooling is sold on.

What competitor analysis is actually for

Competitor keyword gaps are useful and routinely over-weighted. A query a competitor ranks for is not automatically an opportunity: they may have authority you do not, the query may not relate to what you sell, or it may convert badly for them too.

The genuinely useful competitor question is narrower: which queries do they earn that you already have partial relevance for? A gap where you rank at 40 is a different proposition from one where you do not appear at all.

Using an assistant to do the sorting

This is where an assistant earns its place, because the work is genuinely analytical rather than creative. The rule that makes it reliable: give it your data rather than asking what it remembers.

Here is my Search Console query and page data for the last 12 months. Find: (1) queries at average position 11-20 with over 100 impressions, (2) queries with over 500 impressions and CTR under 1.5%, (3) queries where the ranking URL does not obviously match the query topic, (4) pages whose clicks fell more than 30% comparing the last 3 months to the prior 3. Output each group sorted by impressions, with the URL involved.

Every one of those is checkable against the export, which is what separates this from asking a model for keyword ideas. Connecting an assistant directly to the data through an MCP server removes the export step entirely.

The two opportunities that hide from every tool

Two genuinely valuable patterns will never appear in an opportunity report, because the data that would reveal them does not exist in one place:

• Pages nothing links to. A crawler follows internal links, so a page with none is unreachable by definition. Finding them requires comparing your crawl against Search Console, GA4 and your sitemap, which is a join no single tool performs by default.
• Queries you should rank for and do not appear for at all. Absence leaves no trace in your own data. This is the one genuine argument for a competitor index, because their presence is the only evidence the demand exists.

The first is the more valuable of the two on most established sites, because an orphaned page that still earns clicks is ranking with none of the internal support every other page receives.

How often this is worth doing

Quarterly, and after any event that redistributes opportunity abruptly: an algorithm update, a migration, a redesign, or a competitor publishing heavily in your space.

Monthly opportunity hunting is a common habit and mostly produces a longer list rather than a better one, because the underlying data has not changed enough in four weeks to reveal anything new. The exception is the CTR pattern, which is worth checking monthly because it responds fast and the fix is cheap.

One discipline worth keeping: record what you decided not to do, and why. Opportunity lists get rebuilt from scratch every quarter, and half of each new list is usually things a previous version already rejected for reasons nobody wrote down.

The one to run first

If you do only one of these, run the CTR check: pages with substantial impressions and a click-through rate under about 2%. The page already ranks, the fix is a title and description rather than a content project, and the result is visible within a crawl cycle rather than a quarter.

It is also the check most likely to change your mind about what is wrong. Teams frequently arrive convinced they need more content and leave having found that a third of their existing pages rank perfectly well and simply are not being clicked, which is a much cheaper problem and a much faster fix.

Turning a list into an order

Opportunity lists are easy to produce and useless unsorted. Three factors decide the order, and only one of them is volume:

• Proximity. How close are you already? Position 12 is a different project from position 60.
• Effort. A title rewrite and a new pillar page are not comparable units of work.
• Commercial relevance. Does the query relate to what you sell? A large opportunity in the wrong topic is a distraction with a big number attached.

For the metrics that tell you whether any of it worked, see which SEO metrics actually matter.