> ## Documentation Index
> Fetch the complete documentation index at: https://docs.synscribe.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Search Intent and Keyword Cannibalization

> One page targets exactly one intent with one keyword, and your own pages must never compete for the same intent. A 2-D framework decides what survives.

Every page targets exactly one search intent with exactly one keyword. The corollary is a
cannibalization discipline: two of your own pages must never compete for the same intent, and a
keyword doesn't automatically earn its own page unless the intent is genuinely distinct. We
triage keywords through a **2-dimension framework (intent × relevance)** and only the "clean
accept" ones survive.

> **"Does it target ONE search intent with ONE keyword?"**

## Why one intent per page, and no self-competition

If two of your pages chase the same intent, they split authority and confuse the engine about
which one to rank. That's **self-cannibalization**, and it caps a whole pillar without anyone
noticing. Go the other way too: near-duplicate keywords that share an intent should collapse to
one page rather than spawn two.

The 2-D framework separates two things people tend to conflate. Intent is how commercial or
high-purchase the query is (the BOFU-first axis from [1.3](/theory/bofu-first) is this axis in
action). Relevance is how well the query actually matches your product.

You keep only the high-intent, high-relevance terms and tag them for production. Leading with
relevance is also what guards against chasing volume that never converts, because a high-volume
term you're not genuinely relevant for is a trap dressed up as a prize.

## The discipline, in the gates and the build

The **Landing Page Quality Gate** asks *"Does it target ONE search intent with ONE keyword?"*
Production tagging is an explicit accept/reject gate before anything gets built: *"load all
keywords into synscribe, annotate them and tag all the clean accept keywords as 'Core'."* The
framework itself lives at *"see 2d keyword labeling framework."* And **Wonderchat's** pillar list
is deliberately non-overlapping by facet (industry vs tech vs data-source vs objective vs
replacement vs compliance), so cannibalization is avoided by construction, before a single page
exists.

For the concrete "why remove / why add / why merge" tables and the keyword survival gates, see the
[Organise-keywords SOP](/sops/organise-keywords#reference-—-why-pi-recommends-removing-a-keyword).

## How the two axes score

Intent is a set of seven tests a keyword can pass, and it's multi-select: urgency, purchase,
location, specification, product, problem-solution, and alternative. The more tests a keyword
passes, the stronger its intent, so a term like "crm with call forwarding for agencies" stacks
several at once. Relevance is single-select and ranked, most to least relevant:

| Relevance (ranked) | The keyword…                                                    |
| ------------------ | --------------------------------------------------------------- |
| OFFERING           | describes a core product or feature we provide                  |
| ALTERNATIVE        | names a direct competitor we're a viable switch from            |
| COMPLEMENTARY      | a related product we integrate with or sit alongside            |
| INCUMBENT          | a market default or free utility that's impractical to displace |

ALTERNATIVE shows up on both axes and that's deliberate: on the intent axis it's the "ready to
switch" signal, on the relevance axis it's "we're a real alternative to a named competitor."

The whole point of the grid is the top-right corner, high intent and high relevance, meaning
OFFERING or ALTERNATIVE terms with strong buying signals. Anything drifting toward informational
intent or INCUMBENT relevance gets deprioritised or skipped. We keep a large pre-labelled keyword
database on this scheme so the strategy conversation always runs in the same vocabulary. One thing
the framework does not decide is winnability: whether we can actually rank is the SERP-cliff read
from [1.4](/theory/serp-cliff), not a separate gate here. Labelling picks the keywords worth
winning; the cliff picks the ones we can win.

## Operationalize it

* Dedupe, tag, and de-conflict a batch: [2.2.3 — Organise, tag & file keywords](/platform/organise-tag-file-keywords).
* One-intent-per-page when you design pillars: [2.3.1 — Design a landing-page pillar](/platform/design-landing-page-pillar).
* The full cycle procedure: [3.4 — Organise keywords for a cycle](/sops/organise-keywords).