Engine log: the experiment that beat every post we published this year

Key takeaways

• Adding a single 50-word definition block to the top of 12 existing posts lifted average citation rate from 18% to 41% of relevant test queries — the biggest single-experiment result of the year.
• The mechanism: definition blocks give engines a pre-extracted, self-contained answer to the 'what is X' query that opens most topic investigations. Owning that extraction earns top-of-funnel citations for every downstream query on the topic.
• The 12 posts were chosen by a simple filter: topics where we ranked on page one but had no definition block. The lift was consistent across all 12 — no outliers above or below 10 percentage points of the mean.
• Next: we are running the same experiment on 30 more posts and will report results in the September engine log.

We added a 50-word definition block to the top of 12 existing posts and measured citation rate over 14 days. Definition-query citation rate went from 18% to 41%. Related-query citation rate went from 14% to 27.

Hypothesis

AI engines answering definition queries ('what is X') prefer pre-extracted, self-contained answers. A post without a definition block forces the engine to synthesise a definition from body prose — a lossy process. The engine has to infer the definition, pick a sentence, and hope it's clean enough to cite. A definition block removes that inference step entirely.

The secondary hypothesis: citation on a definition query establishes the post as the authoritative source on the topic. That authority carries over to related queries, increasing citation probability downstream. One block, two citation surfaces.

Method

Selection filter: posts where Avakata ranked on page one for the topic keyword but had no definition block. 12 posts met the filter across five topic clusters.

Definition blocks were written to a strict 50-word constraint — single sentence, defined term as grammatical subject, key differentiator included. No jargon unless the jargon was defined in a preceding sentence.

Measurement window: 14 days post-publish. Metrics tracked:

• Citation rate on definition queries (queries matching 'what is [topic]' or equivalent)
• Citation rate on related downstream queries for the same topic

Baseline was the 14-day window immediately before the edit.

Results

Definition-query citation rate 18% 41% +23 pp Related-query citation rate 14% 27% +13 pp

Consistency was high. No individual post was more than 10 percentage points above or below the mean lift on either metric. The effect was not driven by one or two outliers.

Mechanism

Definition queries: The definition block is a ready-made extraction unit. The engine lifts it verbatim or near-verbatim. There is no synthesis step, no ambiguity about which sentence represents the definition, and no risk of the engine pulling a weaker sentence from deeper in the post. The block does the engine's job for it.

Downstream queries: Once the engine has cited a post for a definition, it has established that post as the authoritative source on the topic. Subsequent queries on adjacent aspects of the same topic — use cases, comparisons, how-to questions — draw on the same authority signal. The post gets cited again, not because it was re-optimised, but because the definition citation created a prior.

This is a compounding effect. The definition block earns the first citation. The first citation earns the second. The cost is one 50-word sentence.

What makes a good definition block

Four rules:

1. The defined term is the grammatical subject. The sentence opens with the term, not with 'A [term] is…' buried after a clause.
2. 25–50 words, single sentence. Long enough to include the differentiator, short enough to be lifted verbatim.
3. Includes the key differentiator. What separates this term from the adjacent or competing term a reader might confuse it with.
4. No unexplained jargon. If the definition requires a technical term, define that term first in a preceding sentence.

Weak: GEO is a new approach to SEO that focuses on generative AI systems and how they surface content in response to user queries, which is different from traditional search engine optimisation in several important ways.

This is 38 words but buries the differentiator, uses 'new approach' as a hedge, and does not name what it differs from precisely.

Strong: Generative Engine Optimization (GEO) is the practice of structuring content so that large language models cite it in generated answers — distinct from traditional SEO, which targets ranked links in a results page.

The term is the subject. The differentiator is named. It fits in one sentence. An engine can lift it without editing.

What this experiment did not test

Whether definition blocks affect traditional organic rankings. We tracked ranking positions across the 12 posts during the measurement window and observed no negative movement. But this was not a controlled test — other variables were not held constant, and 14 days is too short to draw ranking conclusions.

We are not claiming SEO neutrality. We are reporting that we did not observe harm. A longer controlled study is needed before making that claim.

Next 30 posts

The selection filter for the next batch is expanded:

• Page-one ranking or top-3 ranking on a related query
• No existing definition block
• Topic has measurable definition-query volume (verified via query data, not assumed)

Posts are prioritised by current citation rate, lowest first. The logic: posts already performing well on citation have less headroom; posts with low citation rates have the most to gain from a single structural change.

30 posts meet the expanded filter. Definition blocks will be added over the next four weeks. Results will be reported in the September engine log.

Frequently asked questions

What is a definition block and how does it affect AI citations?

A definition block is a 25–50 word single-sentence definition placed at the top of a post, with the defined term as the grammatical subject. AI engines prefer pre-extracted, self-contained answers for definition queries — they pull the cleanest, most complete unit they find. Owning that extraction earns you citations not just for the definition query itself, but for downstream queries on the same topic where your definition establishes the frame.

How do you write a definition block for a blog post?

Four rules: the defined term is the grammatical subject, the definition is a single sentence of 25–50 words, it includes the key differentiator from adjacent terms, and it uses no jargon. If the term requires jargon to define accurately, define the jargon in its own definition block first. A definition that requires prior knowledge to parse will not be extracted cleanly.

Why does adding definitions to existing content lift citation rates faster than publishing new posts?

Existing posts already carry ranking signals and topic authority — the engine trusts them. Adding a definition block gives it an extraction unit it was previously missing, and the lift is immediate because the post is already indexed. New posts must first earn ranking signals before any definition block on them can be cited. Retrofitting definitions is the highest-leverage edit you can make to a mature content library.