Competitors with Claude: compare offers on equivalent terms
Comparing competitors with Claude: structure the table without ever filling an unverified cell.
By Educasium
Compare with Market Competitors →
You want Claude to compare your offer against two or three identified competitors, with a clear table of prices, features and positioning. Once the request is made, the temptation is to let the model fill in every cell of the table even when the exact information was not found — an approximate price, a guessed feature, a market share that "sounds plausible." That is precisely the risk this skill must neutralize, more than any other in the marketing catalog.
Market-Competitors relies by default on a script that gathers accessible information from identified competitors’ public pages; if that script is unavailable in your environment, its instructions explicitly provide for gathering the same information by hand, through a plain web search. Both paths produce a real result, provided every observation is always tied to a page and a date — and a price, feature or market share that has not been verified is never presented as an established fact.
This article details how to structure this comparison without ever crossing the line between a verified observation and a guess that sounds right.
Summary
- The risk specific to comparing competitors: plausibility standing in for verification
- Choosing the right companies before comparing them
- What you get depending on your working environment
- Comparing offers on equivalent terms, step by step
- What Market-Competitors never guarantees
- Training to tell a verified observation apart from an estimate
- Frequently asked questions
The risk specific to comparing competitors: plausibility standing in for verification
A comparison table with an empty cell draws the eye and invites filling; a language model, like a rushed person, can fill that gap with an estimate that resembles a fact without ever being one. The risk is not that a price is obviously wrong — it is that a plausible, never-verified price ends up treated as solid data in a positioning decision.
Why this risk runs higher here than for other skills
Comparing competitors involves sensitive business data — prices, discounts, market shares — that the companies involved do not always publish in full, and for which a wrong estimate can steer a pricing or positioning decision the wrong way. A skill that describes your own brand voice from your own texts works on material you control; a skill that describes a competitor works on material you do not control and must verify.
The simple rule to keep in mind before any comparison
Every cell in the table should be traceable to a page consulted and a date, or explicitly marked "not available" or "on request" when the source confirms it. A cell filled in without that traceability should be treated as unreliable, however plausible it looks.
Choosing the right companies before comparing them
A direct competitor answers the same need for the same audience, an indirect alternative answers the same need through a different approach, and an inspiring example comes from a different sector but illustrates a practice worth considering. These three categories play different roles in a comparison, and mixing them without saying so muddies the final table.
Choosing three to four relevant companies, with a brief reason for including each, produces a more useful comparison than a list of ten names gathered without clear logic. The number of competitors cited is not a mark of rigor; the quality of verification for each one is.
What you get depending on your working environment
| Working environment | What you get | Practical difference |
|---|---|---|
| Claude Code, with the runnable collection script | Structured collection of accessible pages from several competitors in one pass | Time saved on collection, not on verification, which stays manual |
| Claude.ai in conversation, without the script | Collection of the same information page by page, via a plain web search | Every page must be opened and cited individually — a bit slower but just as real |
| ChatGPT or Gemini, in conversation | Collection rebuildable on the same principles, depending on the tool’s actual web access | Check that the tool actually cites the source page for each observation, or the comparison is not verifiable |
| Any environment, with no web access | No collection possible; only data you have already supplied can be compared | Supply the pages or figures to compare yourself rather than letting the model guess |
What neither the script nor a web search ever gives you
Neither the bundled script nor a manual web search gives access to a competitor’s internal figures — its margin, its customer acquisition cost, discounts actually negotiated with specific clients. That information stays out of reach of any method based on public pages, and a report that states it as an established fact goes beyond what the method can genuinely know.
Comparing offers on equivalent terms, step by step
Step 1: Document your own offer with the same rigor as the competitors’. Record your price, service scope and terms exactly as you would for a competitor, with the same demand for precision, rather than treating your own offer as self-evident and not needing the same care in writing it down.
Step 2: Open and cite every page consulted for every competitor. Note the exact URL and consultation date for each piece of information recorded — price, feature, contractual commitment — so a later disagreement can be checked directly against the source rather than the conversation’s memory.
Step 3: Convert units before comparing two prices. A monthly rate and an annual rate that includes support do not compare line for line without adjustment; bring both to the same period and keep the scope difference as a distinct piece of information rather than letting it vanish in the conversion.
Step 4: Explicitly flag every piece of information not found. A feature absent from a page you consulted is not necessarily absent from the product; write "not mentioned on the page consulted" rather than "absent" — a nuance that avoids wrongly attributing a disadvantage to a competitor.
Step 5: Separate the finding from the proposed positioning. Once the table is filled in and checked, the part proposing a positioning for your offer must be clearly set apart from the factual observations preceding it, so a reader can challenge the proposal without having to dispute the facts recorded.
What Market-Competitors never guarantees
The skill never guarantees the accuracy of a price publicly displayed at a given moment: rates change, unpublished discounts exist, and a price recorded today may no longer hold in a few weeks. Every price cited in the report must carry its collection date to stay usable over time.
It also provides no market share, no revenue figure and no audience data that is not explicitly published by the competitor itself or by a cited third-party source. An "estimated" market share with no source behind it remains an invention dressed up with an appearance of precision, and has no place in a serious comparison report.
In the comparisons we prepare for self-employed professionals and small businesses, the costliest mistake is almost never an obviously wrong figure — it is a plausible, never-verified figure that ends up steering a pricing decision without anyone ever having checked it against its source.
Training to tell a verified observation apart from an estimate
Learning to structure a comparison without letting a gap fill itself in is a skill that takes practice, particularly for profiles discovering AI use in marketing. Educasium’s Master Claude training devotes modules to verifying sources in AI-produced analyses. For a business owner or a self-employed professional under the relevant scheme, the FIFPL covers part of the cost of this Qualiopi-certified training under the 2026 criteria set at €300/day and €900/year, with e-learning capped at 50% of the daily rate.
Frequently asked questions
Can I trust a competitor price found by this skill without verifying it myself?
No: the skill records a price displayed on a page, but does not guarantee it is current or that it matches what a client would actually pay after a negotiated discount. Always open the cited page and note the date of your own check before using that price in a decision. If the price changes frequently or depends on a custom quote, note that instability in the report rather than locking in a figure that will quickly go stale.
What should I do if the collection script is unavailable in my environment?
Explicitly ask for manual collection through a web search: the skill’s instructions cover this case and produce an equally real result, simply page by page rather than in one automated pass. The quality of verification matters more than the collection method used. Either way, require a URL and a date to accompany every recorded piece of information, or it should not appear in the final report.
How should I handle a feature I cannot find mentioned anywhere for a competitor?
Record it as not mentioned on the pages consulted rather than absent from the product, unless an explicit source actually confirms its absence. A feature can exist without being highlighted on the public pages examined, and stating it as absent without certainty would mislead the reader. If this uncertainty matters for your decision, contact the competitor directly or check a recent customer review rather than deciding based solely on the absence of a mention.
Can the report give a market share to situate my offer?
Only if a genuinely public, citable source exists for that figure; absent such a source, the report should state that market share remains unknown rather than offering an estimate that would resemble verified data without being one. This section remains one of the most frequently fabricated when verification is skipped. Market share remains an especially sensitive figure to verify, since a wrong estimate can directly influence a pricing or positioning decision.
Comparing competitors on equivalent terms, with a source and a date for every cell in the table, protects a positioning decision far better than a more complete but partly invented table. That is the role of the Market-Competitors comparison pack, usable with or without its collection script depending on your environment.
Once competitors are compared on verified grounds, formalizing what your own brand actually says calls for a complementary method: our guide to defining brand voice with approved examples details how to approach it without inventing history.
Training 100% fundable via OPCO/FIFPL. Qualiopi-certified programme. To learn how to verify your sources in a competitive analysis with AI, as part of our Master Claude training, contact Educasium and specify your status (employee, self-employed, business owner) and your goal.