What is Information Gain?

At a glance

• Information gain is measured against what already exists, so the bar rises every time a topic gets covered again.
• Rewriting the current top-ranking pages produces no gain, by definition.
• The cheapest sources of gain are your own operating data, your own tests, and the customer questions nobody answers in public.

Redundancy is easy for a machine to spot

A retrieval system holds many passages saying the same thing. When it assembles an answer it needs a small number of them, and any additional source that repeats what it already has contributes nothing to the answer.

This is not a judgment about writing quality. Your explanation can be clearer than the three beside it and still be skipped, because restating a known fact does not change the answer. A fact the other sources lack does.

Google’s guidance on creating helpful, people-first content asks a version of the same question: does this page offer substantial value compared with other pages in the results. Both systems are trying to avoid paying attention to the same thing twice.

The tenth explanation of a concept has nowhere to go

Most content calendars are built by looking at what ranks and producing a better version of it. That produced results when the reward was a ranking position, because a list of ten had room for a tenth good page.

An answer does not. It names a few sources, and there is no benefit to including the fourth page that says the same thing as the first. The tenth explanation is not ranked low. It is absent.

That is what makes commodity content worthless now rather than merely inefficient. A definition page with nothing extra on it is absorbed into the model’s general grasp of the topic and returns nothing to whoever paid for it.

Find the gap before you write the brief

Reverse the usual order. Research what is missing, then decide whether you can supply it.

1. Read the sources an assistant currently cites for your question. Not the search results, the sources named in the answer.
2. List what all of them leave out. Prices, failure cases, regional differences, what happens after two years of use, who this is wrong for.
3. Check what you can answer from internal knowledge. Support data, returns data, your own testing, the objections your sales team hears weekly.
4. Write the brief around that gap and treat the background explanation as context, not as the point of the page.

If step three comes back empty, do not write the page. Commission the test, run the survey, or pick a different question. Original research is slower than another explainer and it is the only version of this work that compounds.