Competitor scan for positioning decisions

Use this workflow before approving a positioning direction, campaign platform or landing-page message. It gives product marketers and brand leads one decision document: what competitors say, where their proof is weak, and which message territory you should test or own.

Run it against a defined competitor set and a named audience. Do not ask the model for a generic market summary. Give it the pages and campaign material you want compared.

1. Set the decision and comparison boundary

Write a one-sentence decision question before collecting material. For example: Which message territory should lead the enterprise security campaign for IT operations leaders?

Then create a scan brief with these fields:

• Your product: product name, category and the capability being positioned.
• Audience: job title, company type, current problem and buying context.
• Decision: the message or positioning choice that this work must inform.
• Competitors: three to six named companies. Include direct alternatives and, where relevant, the manual process or incumbent tool.
• Pages in scope: home page, product page, relevant solution page, pricing or offer page, campaign landing page and comparison page.
• Geography and date: the market being reviewed and the date you captured each page.
• Your constraints: approved product facts, prohibited claims, current campaign promise and available customer evidence.

Keep the scan narrow. If you are deciding a campaign message for one audience, do not mix in every product line or regional site. A broad scan produces bland gaps such as “be more customer-centric”.

2. Capture the source pack

For each competitor, save the page URL, page title, capture date and the relevant text in a working document. Include screenshots where layout or visual hierarchy carries the message. Record only what is visible on the page, not what you assume the company can deliver.

Use a source table with one row per page:

Competitor	Page and URL	Captured	Evidence to extract
Named competitor	Product page	Date	Headline, subhead, CTA, proof
Named competitor	Solution page	Date	Audience, pain, use case, offer
Named competitor	Landing page	Date	Campaign promise, form, objections

Extract these items verbatim where possible:

• The headline and supporting line.
• The audience named or implied.
• The problem, outcome and mechanism claimed.
• The offer: trial, demo, assessment, guide, consultation or pricing cue.
• Proof: customer names, quantified outcomes, certifications, analyst references, product evidence or testimonials.
• Calls to action and friction: number of form fields, gated asset, calendar booking or self-serve route.
• Repeated phrases, category labels and comparison language.

If you use the model to organise page material or find patterns, paste the source text and labels with it. Availability of features and input limits can vary, so check the current guidance in the xAI documentation overview before loading a large source pack.

3. Build the comparison matrix

Give the model the scan brief and source pack. Ask it first for a structured comparison, not recommendations. Require a source reference for every cell, using your page label such as C2-solution-page.

Use this prompt:

You are supporting a positioning decision. Use only the supplied source pack.

For each named competitor, create a comparison matrix with these fields:
- primary claim, quoted where possible
- named audience and buying context
- problem framed
- promised outcome
- mechanism or product explanation
- offer and call to action
- proof points, classified by type
- landing-page pattern and conversion friction
- repeated language or category term
- missing evidence or ambiguity

For every conclusion, cite the source label. Mark a field as “not evidenced” if the source pack does not support it. Do not infer product capabilities.

Review the matrix before asking for analysis. This is where most errors enter the workflow. A compelling summary built on a misread headline will push the decision in the wrong direction.

4. Separate patterns from gaps

Ask the model to group the matrix into three lists: market conventions, contested claims and credible whitespace. Define whitespace as an audience problem or buying concern that competitors either ignore, state without proof, or address with an unsuitable offer.

Use this second prompt:

Using the cited comparison matrix, identify:
1. Claims used by two or more competitors.
2. Audience needs that are acknowledged but weakly evidenced.
3. Needs, objections or outcomes not clearly addressed.
4. Landing-page patterns that create trust or friction.
5. Risks of making a similar claim without stronger proof.

For each item, state the supporting competitors and source labels. Separate observed evidence from interpretation. Do not call an area whitespace if the source pack is too small to support that conclusion.

Treat an absence carefully. A missing claim on a page may mean the competitor does not lead with it, not that it lacks the capability. Phrase findings as “not evidenced in reviewed pages”, rather than “does not offer”.

5. Draft the positioning decision document

Create a two-page working document with the following sections:

1. Decision required: the exact positioning or campaign choice, owner and decision date.
2. Scope: audience, competitors, pages reviewed and capture dates.
3. Market message map: the five to seven most important competitor patterns, each with citations.
4. Gaps and risks: gaps worth testing, crowded territories to avoid, and claims requiring stronger evidence.
5. Proposed message territory: one lead territory, one alternative and one rejected option.
6. Message architecture: audience tension, promised outcome, reason to believe, proof required and campaign CTA.
7. Validation plan: the next asset, audience or internal review needed before launch.

Write the proposed territory as a testable statement, not a slogan. For example: For [audience] facing [specific job], we lead with [outcome] because we can show [proof]. Then state what makes it distinct from the scanned claims.

6. Check the output before the decision meeting

Read every recommendation against the source table. Flag it if it has no citation, turns a page omission into a capability claim, or combines evidence from different audiences. Check that the proposed territory is specific to the named audience and does not rely on vague words such as “better”, “simple” or “leading”.

If you see this	It usually means	Do this
Every competitor sounds similar	The categories are too broad	Compare the headline, audience and proof separately
A gap has no cited evidence	The model inferred it	Rewrite as a question or remove it
Your territory matches a competitor headline	The position is crowded	Change the outcome, proof or audience focus
Proof is weaker than the promise	The claim may not hold up	Reduce the claim or collect stronger evidence

When the workflow does not work, do not ask for a more creative answer. Return to the source pack. Add the missing solution or campaign pages, correct the audience boundary, and rerun the matrix. If the evidence remains thin, record the decision as a hypothesis and assign a specific validation task before using the message in campaign copy.