Search Engines Stopped Matching Strings. Most Content Plans Did Not.
What changed, and when
Search stopped being a string-matching problem a long time ago, and most content practice has not caught up. The shift was from matching the words on a page against the words in a query to working out what the page is about, what the query wants, and whether the publisher is a thing the system recognises. All three of those are about meaning rather than text.
The practical consequence is that two sites can publish near-identical coverage of a subject and get entirely different results, with the difference sitting in whether the search engine can say who the publisher is. That is a ten-hour job that nobody schedules, because it does not look like content work and it does not look like technical work.
Open most content calendars and you find a spreadsheet of keywords sorted by volume, each assigned to an article. It is tidy, it is measurable, and it produces sites that cover a hundred phrases and no subject.
What coverage means now
A topic is covered when the questions a knowledgeable buyer would ask next are answered somewhere on the site, and the pages connect in a way that shows the relationships. Not when a phrase appears at a target density.
Mapping a subject properly
- Write down the decision a buyer is actually making, in their words
- List the questions that decision generates, in the order they arise
- Group the questions into the pages a person would expect them on
- Attach search demand to the map afterwards, as sizing rather than as the source
- Find the gaps between the map and what you have published
Why this matters more for AI answers
Retrieval works on concepts attached to entities. A site that covers a subject completely gives a model many relevant passages and a clear sense of what the brand is authoritative about. A site with disconnected keyword articles gives it fragments with nothing behind them.
What resolution looks like in practice
An entity is a thing a search engine holds a record about: an organisation, a person, a place, a product. Until your organisation is one, every mention of your name is a string that might refer to you or to the four other businesses with similar names. Resolution is the moment the system stops guessing, and it happens because the same facts about you appear in the same form in enough places that one reading becomes obviously correct.
The work is therefore mostly consistency rather than volume. One legal name, one trading name, one address format, one description of what you do, used identically on your own site, your business profile, your social accounts and anywhere else that gets read. Where those disagree, every disagreement is a reason to keep guessing.
The resolution checklist
- One home page sentence that says plainly what the organisation is, without metaphor.
- Identical name, address and phone number in identical formatting everywhere they appear.
- Organization markup on the site, with sameAs pointing at the profiles you actually control.
- A named author on content, with a page describing who they are and what they have done.
- At least a few third-party mentions that describe you the same way you describe yourself.
Writing for a topic you have not earned
Teams read that semantic coverage matters and respond by publishing forty pages around a subject nobody associates them with. Coverage without resolution produces a library of pages a search engine cannot attribute to a recognisable publisher, which is close to worthless. Establishing who you are comes first, and it is a week of unglamorous work rather than a quarter of content.
Common questions
How do we know if our entity is resolved?
Search your brand name and look for a knowledge panel, and ask an assistant who you are. If the answer is confidently wrong or confused with another business, the entity is not resolved and that is the first job.
Does schema markup create an entity?
It states a claim. Corroboration elsewhere is what makes the claim credible. Markup alone, contradicted by inconsistent listings, achieves nothing.
Is this only relevant for large brands?
The opposite. A large brand is resolved whether it works at it or not. A small business with an ambiguous name has everything to gain and usually nobody looking after it.
Do keywords still matter at all?
As evidence of what people want, very much. As the thing you optimise a page towards, much less than they did. The useful shift is from placing a phrase to covering a subject.
