Competitive Analysis: Evidence, Framework, and Template

Direct answer: how do you do a competitive analysis?

Define the decision first, compare the alternatives a customer actually considers, collect dated primary evidence, separate facts from interpretation, and turn the findings into an action with an owner and review date.

A competitive analysis is not a spreadsheet with every feature and logo in a market. It should help answer a decision such as:

• Which segment should the company prioritize?
• Why are qualified deals lost?
• Which claim can the product prove?
• Where is a competitor structurally stronger?
• Which content or product gap matters to customers?
• How should a launch or packaging choice change?

The right scope for a pricing decision differs from the scope for SEO, sales enablement, product strategy, or acquisition.

Define competitors from the customer decision

A competitor is any alternative that can win the same customer job, including manual work, an internal build, an adjacent product, or no change.

Use four groups:

Group	Meaning	Example question
Direct	Similar product and customer	Why choose them for the same use case?
Adjacent	Different product solves part of the job	When is “good enough” enough?
Internal	Spreadsheet, script, or internal team	What makes change worth the effort?
Status quo	Delay or no purchase	What trigger creates urgency?

Do not add a company because it appears in a generic “top tools” article. Confirm that customers mention it, sales encounters it, search behavior reflects it, or the product overlaps a real buying situation.

The B2B marketing benchmarks guide can help validate outside claims, but the competitor set should come from customer and deal evidence.

Write the research question and boundary

A useful brief states the decision, audience, geography, product scope, evidence window, owner, and deadline.

Example:

Decide whether to launch the compliance package for mid-market financial software companies in the United Kingdom. Compare the current product with two named vendors, internal implementation, and no change. Use evidence current through August 2026. Deliver a recommendation, risks, and follow-up tests.

The boundary prevents the work from becoming an endless market encyclopedia.

Record exclusions. If enterprise contracts, reseller pricing, private betas, or non-English markets are outside the evidence, say so. Absence of evidence is not evidence that a capability or offer does not exist.

Build a source hierarchy

Use primary and date-stamped sources for facts, then use customer and internal evidence to interpret their importance.

Source order:

1. official product, pricing, legal, security, and documentation pages;
2. product interfaces and controlled trials;
3. regulatory filings and public records;
4. customer interviews and verified deal records;
5. recorded sales, support, and implementation evidence;
6. credible third-party research with visible method;
7. reviews and community discussions as directional evidence;
8. aggregator pages only for discovery, never as final proof.

For current pricing, link the page and record the access date. Official pages such as Salesforce pricing, ActiveCampaign pricing, and HubSpot pricing show why copied tables age quickly and why plan, billing term, contact or user tier, and onboarding terms must be captured.

Preserve the exact quote or field beside your interpretation. A reviewer should be able to reproduce the fact without trusting the analyst’s summary.

Use an evidence ledger

An evidence ledger makes source quality, freshness, and uncertainty visible.

Field	What to record
Claim ID	Stable reference
Competitor	Company or alternative
Claim	Exact factual statement
Source URL	Direct evidence
Source type	Official, record, interview, third party
Access date	When it was checked
Market	Geography, segment, plan, or product
Confidence	High, medium, or low with reason
Interpretation	Why the fact may matter
Owner	Who rechecks it
Expiry trigger	Product, price, legal, or market change

Never place a sourced fact and an inferred conclusion in the same cell. “Vendor lists SSO on Enterprise” is a fact. “Mid-market buyers will reject the lower plan” is a hypothesis that needs customer evidence.

The marketing analytics statistics guide uses source-level notes that can inform this discipline.

Compare customer jobs, not feature counts

Features matter only through the customer workflow, outcome, risk, and implementation burden.

For each important buying situation, map:

• trigger and current workaround;
• user and decision roles;
• required capabilities;
• setup and migration;
• integrations and data movement;
• governance, security, and support;
• time to a meaningful outcome;
• pricing and contract assumptions;
• proof and limitations;
• switching and exit costs.

A competitor with fewer features may win because it is easier to approve, adopt, or maintain. A longer feature list does not establish stronger product-market fit.

Ask customers to rank tradeoffs rather than rate every criterion as important. Forced choices reveal which difference can change a decision.

Analyze positioning and proof

Separate what a competitor claims, the audience it addresses, the proof it supplies, and the product experience that supports the claim.

Capture:

Layer	Question
Frame	Which category or alternative does it invoke?
Customer	Who is explicitly addressed?
Problem	What urgency or risk appears?
Promise	Which outcome is offered?
Differentiation	What is said to be unique or better?
Proof	Customer, data, demo, documentation, certification?
Limitation	What remains unclear or conditional?

Do not imitate wording because it appears frequently in the category. Repeated language may signal a convention, or it may signal that every company sounds the same.

The content marketing strategy guide provides a framework for turning differentiated evidence into a sustainable editorial system.

Analyze pricing without false precision

Pricing analysis needs the billable metric, package, term, minimum, add-ons, implementation, usage scenario, and source date.

Create three representative customer scenarios rather than one headline price:

Scenario	Users or volume	Required capabilities	Estimated components
Small team	Defined	Core workflow	Plan, users, usage
Growing team	Defined	Automation and integration	Plan, add-ons, onboarding
Complex account	Defined	Governance and scale	Quote, services, contract

Use formulas and leave unknowns visible. Do not fill an enterprise “contact sales” gap with an invented estimate.

An illustrative weighted value comparison might assign 25% to fit, 20% to implementation, 20% to governance, 15% to ongoing operation, 10% to commercial terms, and 10% to exit risk. These weights are a decision example, not a market standard.

Use competitive search data carefully

Search data can reveal language and relative interest, but it cannot prove market share, customer satisfaction, or purchase intent by itself.

Google Trends explains that search data is normalized by time and location and scaled from 0 to 100. Low-volume terms can appear as zero, and Trends is not scientific polling.

For SEO analysis, compare:

• topics connected to the customer job;
• pages that rank and the intent they satisfy;
• source quality and proprietary evidence;
• internal linking and information architecture;
• branded and non-branded demand;
• update cadence;
• gaps where the company has better evidence.

Do not copy a competitor’s keyword set without checking whether the traffic supports the same product, market, and conversion.

The best SEO tools guide can help choose research tools. The analysis still needs human review and source locking.

Turn frameworks into decisions

SWOT, Five Forces, and comparison matrices are prompts for reasoning, not conclusions.

Use SWOT to distinguish internal strengths and weaknesses from external opportunities and threats. Attach evidence and an action to each item.

Use Five Forces when industry structure matters: customer power, supplier power, substitutes, new entrants, and rivalry. Do not use it as a feature-comparison template.

Use a weighted matrix when the criteria and tradeoffs are explicit. Run sensitivity analysis. If a small weight change reverses the winner, the recommendation is fragile.

Use a narrative when sequence and causal assumptions matter. A one-page argument can be more useful than thirty colored cells.

Validate findings with primary research

Competitive conclusions should be tested against customers, lost deals, sales, support, implementation, and product behavior.

Use interviews to ask:

• What triggered evaluation?
• Which alternatives entered and why?
• Which criteria removed an option?
• Which proof was trusted?
• What created internal risk?
• Who changed the decision?
• What nearly stopped the purchase?
• What surprised the team after implementation?

Avoid asking customers to confirm your battlecard. Ask them to reconstruct their decision in sequence and name the evidence they used.

Win-loss records need consistent sampling. Reviewing only famous wins or painful losses produces a distorted picture.

Build battlecards that support judgment

A battlecard should help a seller diagnose fit and risk, not supply attack lines.

Include:

• when the alternative is a strong fit;
• when your product is a strong fit;
• discovery questions;
• verified differences;
• proof links;
• common objections;
• traps to avoid;
• escalation owner;
• last review date.

Never invent a competitor weakness. Public falsehoods create legal and trust risk. Teach sellers to reframe around the customer’s requirements and evidence.

Keep the analysis current

Assign owners and change triggers so important claims are rechecked when they are likely to change.

High-change areas include pricing, packaging, integrations, AI features, security, leadership, and terms. Stable areas can use a slower review cadence.

Set monitors or manual checks for named high-impact sources. Record the new fact, old fact, effective date, affected assets, and required action.

A quarterly meeting is not enough if nobody owns the evidence between meetings. Conversely, daily monitoring of every competitor page creates noise. Match the cadence to the decision risk.

Maintain an asset register that links each competitive claim to the places where it appears: website pages, sales decks, proposals, onboarding material, advertising, and internal guidance. When a source changes, the owner can find every affected asset instead of fixing one battlecard while stale language remains public. Archive the previous version and record why the update was made. This creates a review trail and keeps a correction from being mistaken for a new strategic claim.

Common competitive analysis mistakes

Most failures come from unclear decisions, weak evidence, and no operating owner.

Common mistakes:

• treating a category list as the customer choice set;
• copying aggregator facts;
• presenting inference as fact;
• counting features without workflow context;
• ignoring internal build and no change;
• using search interest as market share;
• hiding unknown prices;
• reviewing only selected wins and losses;
• writing hostile battlecards;
• producing a report with no decision or review trigger.

The cure is not a larger spreadsheet. It is a narrower question and stronger evidence.

A two-week analysis plan

A bounded analysis can produce a decision-ready result in ten working days when the scope and access are clear.

Days 1 and 2: define the decision, alternatives, criteria, sources, and interview sample.

Days 3 to 5: collect primary evidence and build the ledger.

Days 6 and 7: interview customers and internal teams, then compare the evidence.

Day 8: write the recommendation, uncertainty, and rejected alternatives.

Day 9: run a challenge session and sensitivity review.

Day 10: finalize actions, owners, source links, and recheck triggers.

If a critical source is missing, label the conclusion provisional rather than extending the schedule silently.

FAQ

What is competitive analysis?

It is a decision-focused comparison of the alternatives a customer or company can choose, using verifiable evidence and explicit interpretation.

Who are indirect competitors?

They are alternatives that solve the same customer job differently, including adjacent products, internal processes, manual work, and no change.

How many competitors should be analyzed?

Use the smallest set that covers the real decision. A focused analysis of three alternatives can be more useful than a shallow table of twenty.

How often should analysis be updated?

Update when an important decision or source changes. Set faster checks for volatile pricing, packaging, integrations, and terms.

Is SWOT enough?

No. SWOT can organize evidence, but it does not collect evidence, rank decisions, or assign actions automatically.

Can AI perform competitive analysis?

AI can help organize supplied evidence. A person still needs to verify sources, define the decision, separate inference, manage confidential data, and approve conclusions.

Should competitor pricing be published?

Publish only current, sourced, scoped information you can maintain. State the access date, billing assumptions, and unknown quote-based elements.

What is the final deliverable?

Deliver the recommendation, evidence ledger, comparison, uncertainty, rejected options, actions, owners, and source-recheck triggers.

Last verified: August 2026. Official Google, Salesforce, HubSpot, and ActiveCampaign pages were checked on August 4, 2026.