An engine has no reason to cite your page if it can get everything on it somewhere else. The reason it cites you is the part that’s only on your page.
Information gain is the amount of new, unique value a page adds beyond what is already available. Answer engines favor content with high information gain because they gain nothing by citing a new source that merely repeats what established sources already say. If your page contributes something genuinely new, the engine has a reason to reach for it; if it restates the known, the engine stays with what it already trusts.
This concept quietly governs which content gets cited, and it explains a frustrating pattern: well-written, comprehensive pages that still never get quoted. They’re good — but they cover ground already well covered, so they add no gain. The engine can already answer the question from a dozen sources; one more careful restatement changes nothing.
An answer engine assembling a response is trying to give the best, most complete answer it can. Sources that add new information make its answers better; sources that repeat existing information don’t. So the engine has a built-in preference for content that contributes — a unique data point, a distinct perspective, a deeper level of specificity, a synthesis no one else has made. That contribution is what earns a citation, because it’s the part the engine couldn’t have produced from its existing sources.
This is closely tied to original research, the purest form of information gain — but gain doesn’t require formal research. A genuinely different framing, a more precise answer, or first-hand specificity all add it. The common thread is that your page makes the answer better in a way no other page does.
Before publishing anything, ask the gain question: if this page vanished, would any answer be poorer for it? If the honest answer is no — everything on it is available elsewhere — it needs more before it’s worth publishing for AEO. Add what only you have: your data, your experience, your specific examples, your distinct take. Go a level deeper than the existing coverage stops at. Replace generic statements with particular ones. The goal isn’t to write about the topic; it’s to add to it.
This is also why content volume alone doesn’t produce citations. A hundred pages that each restate the known add no gain a hundred times over. A smaller library where every page contributes something new outperforms it, because each page gives an engine a real reason to cite. Build for gain per page, not pages for their own sake.
Information gain is the amount of new, unique value a page adds beyond what's already out there. Answer engines favor content that contributes something not already well-covered, because repeating known information gives them no reason to cite a new source over an established one.
Add what others don't have: original data, a genuinely different angle, deeper specificity, first-hand experience, or synthesis nobody else has done. The test is whether your page would change an engine's answer — if removing it loses nothing, it has no gain.
No. Paraphrasing what's already widely available adds no new value, however well-written. Engines can already get that information from established sources. Gain comes from contributing something new, not restating something known.
We assess whether your content adds information an engine can't get elsewhere — the difference between pages that get cited and pages that get passed over.