Citation is three steps, not one
For an answer engine to cite you, three things have to be true in order. It has to be able to reach and read the page. It has to have a reason to trust what the page says. And it has to find a self-contained claim it can lift and attribute to you. A break at any step ends the chain, and most sites break at the first or the last.
Step 1 — Can it read the page at all?
Answer engines and the crawlers that feed them are not patient. If the content only appears after heavy client-side rendering, or sits behind an interstitial, or is blocked in robots rules, the engine sees an empty or partial page and moves on.
Check the page the way a crawler would: fetch the raw HTML and see whether the substance is there without running scripts. If the words a buyer needs are missing from that fetch, the engine is working from less than you think.
- The core claim and key facts are present in server-rendered HTML, not only after hydration.
- robots rules and meta tags allow the page to be indexed, not just the homepage.
- The page returns quickly and does not hide its content behind a consent wall or modal.
Step 2 — Does it have a reason to trust you?
Trust, to an engine, is not a feeling — it is corroboration and identity. A page that states who published it, links its claims to something verifiable, and carries a consistent machine-readable identity gives the engine a reason to rely on it over an anonymous page making the same claim.
This is where a machine-readable identity earns its place. Structured data that honestly states who you are — and nothing you cannot back up — lets the engine resolve "who is this" without guessing. Invented ratings or fake authorship do the opposite: they are a reason to distrust the whole page.
Step 3 — Is there a clean claim to attribute?
The last step is the one most pages fail. An engine cites a sentence it can lift whole and stand behind. If your key point is spread across three paragraphs, hedged, or buried under marketing, there is nothing clean to quote, so the engine paraphrases a competitor who said it plainly.
Write the answer to the question the page is about as one direct, self-contained statement, early, in the page's own words. That is not dumbing down — it is giving the engine something it can attribute to you instead of to someone clearer.
Turning the checklist into changes
Run a page through these three steps and the failures are usually specific: a claim that only renders client-side, an identity the engine cannot resolve, an answer with no quotable sentence. Each is a concrete change, not a vague "improve your content."
A ScanDepth audit reads a site against exactly these signals and returns them ranked by impact — what is wrong, why it matters to an engine, and the change to make. It will not promise a citation; it will remove the reasons you are not getting one.
See what an audit finds
See what an audit finds