Run a Competitor Analysis

Build a competitive view that helps the user make a real decision.

The goal is not to collect company facts for their own sake. It is to understand how relevant competitors differ, what those differences mean, and which conclusions are supported by current evidence.

1. Start from the decision

Establish:

• business decision;
• audience;
• deadline;
• market boundary;
• timeframe;
• useful output;
• expected depth.

Clarify why the analysis is being run.

For example, researching a company may support:

• partnership evaluation;
• positioning;
• product strategy;
• pricing;
• campaign planning;
• market entry;
• preparation for a conversation;
• tracking strategic movement.

The purpose determines which evidence matters.

Use approved company material, the user's website, and prior accepted research when useful for understanding:

• product;
• positioning;
• customer;
• current assumptions.

For client work, keep each client's:

• sources;
• conclusions;
• accounts;
• artifacts;
• confidential context

strictly separated.

2. Define the comparison set

Identify:

• direct competitors;
• indirect alternatives;
• emerging players;
• substitutes that could materially affect the decision.

If the comparison set is uncertain, propose a reviewable first set.

Explain:

• why each company is included;
• why obvious alternatives may be excluded;
• which players are borderline.

Do this before spending heavily on deep research.

3. Choose the comparison dimensions

Use dimensions that matter to the decision.

Possible areas include:

• product;
• pricing;
• packaging;
• positioning;
• audience;
• channels;
• traction;
• capabilities;
• geography;
• partnerships;
• strategy;
• recent changes.

Reuse an accepted comparison framework when it still fits.

Do not force a generic competitor template when the user's decision needs a narrower frame.

4. Choose evidence depth deliberately

Start broad when the market is unclear.

Go deeper on:

• competitors;
• questions;
• claims

most likely to change the decision.

For small competitor sets, independent research passes may improve depth, but they require more work and should be used when the added confidence matters.

Prefer current primary sources.

Use reputable secondary sources to corroborate or fill material gaps.

Treat differently:

• company claims;
• customer reviews;
• market estimates;
• third-party analysis;
• public records.

Only infer what each source can reasonably support.

5. Research the competitors

Use available:

• live web;
• visible browser;
• approved logged-in research tools;
• prior research;
• internal materials.

Useful evidence may include:

Product and offer

• product features;
• services;
• packaging;
• pricing;
• offers.

Positioning and marketing

• homepage messaging;
• audience;
• campaigns;
• ads;
• content strategy;
• recent messaging changes.

Strategic signals

• launches;
• partnerships;
• funding;
• financial disclosures;
• hiring;
• geographic expansion.

Organizational signals

• job postings;
• leadership posts;
• talks;
• interviews;
• articles.

Customer evidence

• reviews;
• community discussion;
• switching behavior;
• recurring strengths;
• recurring complaints.

Follow links to the underlying source whenever possible instead of relying on search summaries.

Record:

• source;
• source date;
• what was checked;
• what could not be verified;
• potentially stale evidence.

6. Normalize before comparing

Make the comparison sufficiently like-for-like.

Identify differences in:

• currency;
• billing period;
• package;
• scope;
• geography;
• audience;
• source definition;
• metric methodology.

Do not flatten unlike evidence into a misleading comparison.

Keep distinct:

• observed facts;
• competitor claims;
• external interpretation;
• your analytical inference.

Example:

• hiring several enterprise salespeople in a region = observed signal;
• expansion into that region = hypothesis unless supported by stronger evidence.

7. Interpret the competitive picture

Explain what the evidence means for the user's decision.

Identify:

• meaningful gaps;
• crowded positioning;
• credible opportunities;
• strategic risks;
• changing market patterns;
• important unanswered questions.

Do not treat difference as advantage automatically.

Explain why a difference matters to the target audience or business decision.

8. Deliver a decision-ready comparison

Produce the agreed format:

• brief;
• comparison table;
• report;
• presentation.

Lead with:

• decision-relevant conclusions;
• implications;
• recommended areas to investigate or act on.

Then show:

• evidence;
• dates;
• source links;
• uncertainty;
• missing information.

Allow the user to correct:

• comparison set;
• evidence bar;
• dimensions;
• research depth;
• interpretation

before expanding the analysis.

9. Route adjacent work appropriately

When the request becomes primarily organic search analysis, route to the SEO audit workflow.

When the work becomes a broader market taxonomy or entity landscape, route to the market-mapping workflow.

Do not duplicate those specialist workflows unnecessarily.

10. Keep execution permissions separate

Treat these as separate actions:

• public research;
• logged-in intelligence research;
• internal artifact creation;
• sharing;
• campaign changes;
• website changes;
• product changes;
• pricing changes.

Follow active scoped permission for every:

• source;
• client;
• account;
• destination;
• action.

Stop or ask when:

• identity changes;
• scope changes;
• impact changes;
• sensitive data handling changes.

Verify completed external actions when possible.

11. Preserve the accepted competitor method

After the workflow proves useful, preserve:

• scope;
• competitor set;
• source hierarchy;
• comparison dimensions;
• research depth;
• output structure.

Add recurring competitor monitoring only when there is:

• a real cadence or trigger;
• defined competitors;
• defined sources;
• meaningful-change rules;
• approved output;
• destination;
• review behavior;
• stop conditions.

Do not assume a weekly or monthly schedule by default.