Most warnings about research go to the quality of the answer. Your sample was skewed, your scale re-zeroed, your denominator was wrong, your metric amplified the noise. This one is different, and it is the reason competitive pricing research deserves more care than its size suggests.
Here the answer can be perfectly correct. The sample can be clean, the method sound, the finding true and genuinely useful. And the record of having asked the question is still the problem.
Answer first
In an information-exchange case, the conduct at issue is often the asking itself, not the conclusion. In United States v. Container Corp. of America (1969) the Supreme Court found a Sherman Act violation where there was no price agreement at all. Its own words: "There was here an exchange of price information but no agreement to adhere to a price schedule."
What existed was a habit of requesting: "Here all that was present was a request by each defendant of its competitor for information as to the most recent price charged or quoted, whenever it needed such information and whenever it was not available from another source." That reciprocal practice was, the Court held, "sufficient to establish the combination or conspiracy, the initial ingredient of a violation of" Section 1. The Court called the arrangement "though somewhat casual" and condemned it anyway.
No agreement. No meeting. No conclusion acted upon. A pattern of questions, documented.
Now consider what a modern research function produces. A brief. A discussion guide. Recruitment screeners. Recordings. Transcripts. A synthesis deck. A Slack thread where someone summarises it in one line. Every one of those is a document, and a pattern of questions is exactly what they preserve.
The asymmetry nobody plans for
Research artefacts are written to be persuasive internally. That is their job: a finding has to travel from the researcher to the person who sets the price, and it travels best when it is compressed into something decisive.
Compression is where the risk enters, because the compressed form of a competitive finding almost always looks like an instruction about price. "They are at 12 percent below us on the mid tier, we should close the gap." That sentence is a reasonable summary of good research. It is also, read cold by someone reconstructing events years later, a sentence about matching a competitor's price.
The finding is not wrong. The research is not wrong. The document is simply capable of a second reading, and it will be read by people whose job is to test that second reading.
Two features of current enforcement make this sharper than it used to be.
Intent is not the test. The DOJ and FTC Antitrust Guidelines for Business Activities Affecting Workers, issued in January 2025, state that an exchange may be unlawful when it "has, or is likely to have, an anticompetitive effect, whether or not that effect was intended." A sincere research motive does not answer an effects question. We were only doing market research describes your purpose, and purpose is not the element being tested.
Indirection is not a cure. The same guidelines say an exchange may be illegal "even if companies use a third party or intermediary" to share the information, adding "including a third party using an algorithm", and that exchanges routed through third-party software "can be unlawful even if the exchange does not require businesses to strictly adhere to those recommendations." The layer of separation that feels like insulation is, in the government's framing, part of the mechanism.
That theory is being litigated now. In United States v. RealPage, Inc., Civil Action No. 1:24-cv-00710 in the Middle District of North Carolina, a proposed Final Judgment for landlord defendant Willow Bridge Property Company was filed on 6 July 2026, on the allegation that its "agreements with RealPage, Inc. and other landlords to share information and align pricing violate Section 1 of the Sherman Act." An earlier proposed Final Judgment, published in the Federal Register on 5 December 2025, restricted the software to training data "at least 12 months old and not from Active Leases", explaining that ageing "at least 16 months will exclude virtually all active leases."
Why the good-faith defence is weaker than it feels
There is a Supreme Court case directly on the point that a reasonable-sounding business purpose does not automatically save an exchange with competitors.
United States v. United States Gypsum Co. (1978) concerned "the practice of telephoning a competing manufacturer to determine the price being currently offered on gypsum board to a specific customer." The defendants had a genuinely respectable reason: they said they were verifying prices in good faith to qualify for the Robinson-Patman Act's meeting-competition defence. In other words, they were calling competitors in order to comply with another statute. The Court of Appeals had treated that purpose as a controlling circumstance precluding liability. The Supreme Court did not adopt that as a general shield.
The Court was careful to say the practice is not inherently unlawful: "The exchange of price data and other information among competitors does not invariably have anticompetitive effects; indeed such practices can in certain circumstances increase economic efficiency and render markets more, rather than less, competitive." But the analysis turns on "the structure of the industry involved and the nature of the information exchanged" - not on how good your reason was.
If verifying a price to comply with federal law is not a blanket answer, "we needed it for the pricing model" is unlikely to be one either.
Design so the record is clean, not just the conclusion
None of this argues for doing less pricing research. It argues for doing it against your own market rather than against your competitors, and for writing it in a form that reads the same way in five years.
Classify every question before fielding. There are two categories and they behave completely differently. First-party questions are about the respondent: what they paid, what they value, what they compared, what they would do at a different price. These are lawful, durable and are the substance of pricing research interviews. Third-party questions seek a competitor's nonpublic terms through someone else. The second category is where both the legal exposure and the weak evidence live, and it is almost always removable without losing the finding.
Screen competitor employees out of pricing modules. A single_choice employer screener that routes them away before the sensitive section costs one question and eliminates the worst fact pattern.
Use methods that measure your own demand curve. Van Westendorp and Gabor-Granger both produce a defensible price recommendation using only your own respondents' willingness to pay, with no competitor input at all. See the Van Westendorp price sensitivity meter and the price increase research guide. This is the substitution that matters: the same decision, sourced entirely from people who are yours to ask.
Write findings in demand language, not competitor language. "Customers in the mid tier report the current price is above what they consider fair value, and 38 percent said they would not renew at a further increase" is a finding about your customers. "Competitor X is at 12 percent below, we should match" is a sentence about a competitor's price. The two can rest on the same interviews. Only one of them reads badly out of context.
Keep the instrument, not just the insight. Retain the discussion guide and screener alongside the report. A preserved instrument showing a bounded, first-party scope is the artefact that demonstrates what was and was not asked. Ordinary anonymisation practice applies to the transcripts as usual.
Where the interviewing method itself matters
The scope problem is a moderator problem. A skilled human interviewer is trained to follow the thread, and the thread that matters here is the one you do not want followed. When a respondent volunteers something about a competitor's pricing, every instinct a good moderator has says go there. Everywhere else in research that instinct is the whole value. In competitive pricing work it is the failure mode, and it is invisible until someone reads the transcript.
A Koji AI interviewer probes deeply inside the brief you set, and does not step outside it. The same scope applies to session one and session two hundred, so what was asked is a property of the study rather than of the moderator's day. Every session is transcribed in full, thematic analysis runs automatically, and the report is generated from the transcripts, so the chain from question to finding is inspectable end to end.
That is a different offer from the rest of the market. Qualtrics, SurveyMonkey and Typeform give you a fixed instrument with no probing, so depth requires hiring moderators and accepting the drift. UserTesting and dscout put a human in every session, which is exactly the variable you are trying to control, at a per-session price that caps your sample. Dovetail and similar repositories analyse conversations after they happened, which does not help with what was asked. Koji is the only AI-native layer that gives you real conversational depth and a fixed, auditable scope - because the moderator is configuration, not a person you brief and hope. Structured questions do the rest: open_ended for the reasoning, scale for magnitude, single_choice and multiple_choice for the comparison set, ranking for trade-offs, yes_no for the decision. Bounded response spaces cannot wander somewhere expensive.
This is background, not legal advice. Antitrust analysis is fact and market specific and differs by jurisdiction. If your pricing research touches competitors, trade associations or a shared data intermediary, talk to counsel before fieldwork rather than after.
Price against your customers, not your rivals
The strongest pricing evidence you can own is what your own buyers will pay and why. It is lawful without qualification, it is specific to your product, and it explains itself in a way a competitor's number never will. The reason most teams reach for competitor prices instead is that talking to enough customers used to be slow and expensive.
It is not any more. With Koji you design the study once, set the boundary once, and run AI-moderated voice interviews across as many customers, churned accounts and lost deals as the question needs. No scheduling, no moderator bias, no research expertise required. Themes, quotes and a one-click report come out the other side, so you get from question to insight in hours instead of weeks. Pair it with pricing page research and win-loss analysis for the full picture of how you are chosen and at what price.
Start a pricing study with Koji and build a price on evidence that is unambiguously yours.
Frequently asked questions
Can researching competitor pricing itself be a legal problem?
It can be, when the information comes from or through competitors. Container Corp found a Section 1 violation on a reciprocal practice of requesting current prices, with the Court noting "an exchange of price information but no agreement to adhere to a price schedule." The conduct was the pattern of asking. Researching competitor prices from customers, public sources or your own sales team is a different matter and is ordinary lawful research.
Are our interview transcripts and research decks discoverable?
Business records generally are, in litigation and in government investigations, subject to the usual rules and any applicable privilege. The practical implication is not to keep fewer records but to make sure the instrument and the findings accurately reflect a bounded, first-party scope - so that the documents read the same way to a stranger as they do to you.
Does it help that we never acted on the competitor information?
Less than people expect. The 2025 DOJ and FTC guidelines describe exchanges as potentially unlawful "whether or not that effect was intended," and note that arrangements can be unlawful "even if the exchange does not require businesses to strictly adhere to those recommendations." Non-adherence is a fact in your favour, not a complete answer.
How should a competitive pricing finding be written up?
In terms of your own customers' demand wherever possible. A statement about what your buyers value, what they consider fair, and what they would do at a given price is a finding about your market. A statement about a competitor's price and what you should do to match it is a sentence about coordination, even when it rests on identical evidence.
Can we get a real pricing answer without any competitor data?
Yes, and usually a better one. Van Westendorp and Gabor-Granger both derive a price recommendation from your own respondents' willingness to pay. Interviews add the reasoning behind the numbers. Competitor list prices are in any case a poor input, because they omit discounting, bundling and the terms that determine what anyone actually pays.
What is the single highest-value change to make?
Separate first-party questions from third-party questions at design time, and delete the second category. In our experience almost every competitive pricing question can be re-expressed as a question about the respondent's own evaluation and choice, which is both lawful to ask and stronger evidence, because the respondent actually knows the answer.
