Information gain
Information gain is what a page adds that other pages on the same subject don't: new data, first-hand experience or an original view.
Information gain is what a page tells you that the other pages on the same subject don't. If ten articles already list the same five tips, an eleventh with those five tips adds nothing. One with test results nobody else has adds a lot.
It matters because AI platforms read several pages for one answer. A page that repeats the others gives the platform no reason to use it. A page with something only it says does.
Where the idea comes from
The term comes from information theory, and Google holds a patent on scoring how much new information a document gives someone who has already seen other documents. A patent doesn't prove Google uses it. But the idea matches what Google asks for in its guidance on helpful content: original information, research or analysis.
What adds information
The things that add information are the things only you have:
- Your own data, like what your customers order most or how long your product lasts
- First-hand testing, with photos and numbers
- A clear opinion, with your reasons
- Answers to the follow-up questions other pages skip
For example, most pages about cold brew give the same ratio of coffee to water. Fernhill Coffee brewed 12 batches at four ratios and four steeping times, and published which one their tasters liked best. That page has something an AI answer can quote that no other page offers.
What doesn't
Rewriting the top results in your own words adds very little, however well it's written. Neither does making a page longer. A short page with one new fact can carry more than a long one with none.