Most contradictions resolve as both experts being right about different segments, many as one holding stale information, and only a small minority as genuine disagreement about the same facts. The instinct is to decide who is more credible; the more useful question is what each of them was actually describing.
Why do two experts disagree?
Most often because they are not describing the same market. One sells into national accounts and one into regional independents; one covers the segment where the product is discretionary and one where it is specified. Both give accurate accounts and the accounts do not reconcile.
The tell is usually in what they take for granted rather than in what they assert. Two people describing the same market share assumptions about how it works; two people describing different segments do not, and the mismatch surfaces in the unexamined parts of each answer.
Is one of them simply wrong?
Occasionally, and less often than the framing suggests. The more common version is that one is describing a market that has changed since they left it — accurate as of eighteen months ago, and stated in the present tense because nobody has told them otherwise.
That is why time since operational exposure is worth capturing on every expert record. It is a better predictor of whose account holds than seniority, and it is the variable most screens do not track.
How do you resolve a contradiction?
Establish the segment boundary first, which is a question the discussion guide should already carry. A single clarifying question to each — which customers, which channel, which size band — dissolves a large share of apparent contradictions before anyone needs to adjudicate, which is work a briefed moderator does on the call rather than after it.
Where that does not resolve it, put the two in the same room. Each then has to defend a position in front of someone who knows the market, which surfaces the basis of the disagreement far faster than two sequential follow-up calls do.
Segment boundaries are also worth recording once established, because the same boundary usually applies to the next question in that market. A program that establishes where national accounts end and regional independents begin has produced a piece of structure that outlives the contradiction it was drawn to settle.
When does a panel settle it?
When the disagreement is about interpretation rather than about facts. If both experts agree on what is happening and disagree about what it means, hearing them argue is the fastest route to understanding which reading is better supported — the difference between convening a panel and booking two calls.
It settles less well where the disagreement is about a specific number neither can source, or where the two are competitors and will not speak candidly in front of each other. Both are worth checking before convening.
What if it stays unresolved?
Report it as a contradiction, with both positions, their basis, and which decision it affects. A finding reported as contested is more useful than a finding averaged into something neither expert actually said.
Averaging is the failure mode worth naming. It produces a number that looks like consensus, carries none of the underlying disagreement, and cannot be interrogated by whoever relies on it later.
Reporting a contradiction well also requires saying what would resolve it. A finding that names the two positions, the basis for each, and the specific piece of evidence that would settle the question is actionable; one that simply records disagreement leaves the reader with less than they started with. The resolution path is usually a third conversation with someone positioned to see both segments.