Why citations are the new backlinks

For twenty years, the SEO loop ran on one signal above all: links. Get more reputable sites to point at your URL, and Google’s algorithm rewarded you with rank. The whole industry — content, outreach, PR — bent around that loop.

Answer engines run on a different loop.

What models actually do

When ChatGPT or Perplexity answers a question, it doesn’t rank a list of pages. It synthesizes a response. To do that, it pulls from a retrieval layer — a set of documents the model considers credible enough to use as evidence. The documents it pulls and chooses to name are citations.

That second part matters: a model can read a page and silently absorb its claims without ever crediting the source. Citations are the subset of inputs that get attributed.

Why citations behave differently than backlinks

Three differences worth internalizing:

• Citations are per-query, not per-domain. A site can be cited heavily for one topic and invisible on another. Authority is no longer a single number.
• Recency matters more. Models are increasingly grounded in fresh retrieval. A six-month-old post can outrank a five-year-old one that has more backlinks.
• Structure beats volume. Models prefer pages with clean claims, dates, and verifiable specifics. A 400-word post that says exactly what it means often gets cited over a 3,000-word post that buries the answer.

What to do about it

You don’t need to throw out your SEO work. Most of it still helps — crawlability, structured data, internal linking. But the target of optimization shifts. You’re no longer writing for the algorithm to rank you. You’re writing for a model to quote you.

That means:

1. State your claims explicitly, near the top.
2. Cite your own sources — models trust pages that show their work.
3. Update dates honestly. Stale content gets demoted in retrieval.

We’ll go deeper on each of these in future posts.