What It Is · Key Takeaways · Do's and Don'ts · Pair With · Things to Explore
On 29 June 1972, a Soviet buying team took a suite at the Madison Hotel in Washington and started working the phones: Cargill in Minneapolis, Continental Grain in Manhattan, Cook Industries in Memphis. The Soviet team knew more than their prospective suppliers about the imminent poor USSR grain harvests that year. Unlike each of their counterparts, they also knew how much they were buying in aggregate. By August, the team had bought 440 million bushels, roughly a quarter of the American wheat crop, for $700 million. The U.S. government subsidized the Russian and other export sales with $300 million. In the year to come, American wheat prices tripled, from $1.50 per bushel to $4.45.
What It Is
Ground Truth is the factual foundation of the negotiation: claims about reality that are—in principle—testable, falsifiable and constrained by something outside the room. It is the first pattern of Rationality, where Logos meets Substance. Its neighboring patterns are Relative Advantage, which governs which option is preferable over its alternatives, and Structural Integrity, whether the deal’s logic is robust against unforeseen conditions. The key phrase is in principle. A supplier’s production cost is a fact about the world even if the buyer can’t see it. “Checkable” does not necessarily mean “checked”.
Most of what gets treated as fact in a negotiation is not. A stated budget isn’t a law of physics but an allocation decision: it moves under pressure, changes with authorization, and expands when the deal is attractive enough. A forecast is a claim about things that haven’t happened yet, and never may. The underlying model adds a second constructed layer: there is only one way to calculate Net Present Value, but a range of choices for which discount rate to use to calculate it. Preferences, standards and causal stories are built, not discovered. What is left, once stripped of beliefs and preferences, is a narrower set than most negotiators assume.
The pattern’s role in a negotiation is determined by the side to whom the fact is presented. A claim does not automatically join the shared record solely because somebody made it. A counterparty can do four things with a claim. Accept it and reason from it. Dispute the number. Dispute the source or the method behind it. Or refuse to accept an undisputed fact as relevant. It’s a response you rarely find in textbooks but every practitioner has used. I don’t dispute your benchmark. But if we do this, we have to shut a production line and lay off two hundred people. Countering logic not with logic, but with emotion.
Key Takeaways
I. Arguing a fact, you concede it is relevant
Disputing a number challenges specific meaning, but not that it matters. So does going after the source, and so does going after the methodology; all three are contested on counterparty terrain. A fourth response works differently. Leaving the fact alone but sidestepping its relevance moves the argument to where it doesn’t prove anything.
II. The usual failure is not a lie but a fact out of context
Negotiators are trained to look for the inflated figure and the suppressed uncertainty. A new entrant accepts industry production yield benchmarks as the basis for its own investment. The benchmarks are representative of mature operators, but not for a startup climbing the learning curve. No claim is false or misrepresented, but the arrangement fails nonetheless. The question is not only whether a claim is true, but also under which conditions.
III. Facts are virtuous, but not sacred
Standard advice is to agree the facts first and then bargain hard over the common ground. But building common ground out of facts is not free. It shapes which inputs count, foreclosing positions the other side has not yet thought of. It also discloses information about your position, creating ammunition that can be used against you. Disclosure without an offensive and defensive reason is naïveté, not integrity. Too much settled fact can also over-constrain the range of outcomes. Who is to say that the most rational deals are always the best ones? Better ones, as decided by the negotiators alone, may consist more of character and feeling than ratio. Extinguish those with facts and it may cost you the one you wanted, in your heart of hearts.
IV. Their optimism comes back at you as operating cost
When a counterparty’s assumptions are unexpectedly beneficial to you, the temptation is to take the win and leave the consequences to them. Still, a party who has committed more than they can deliver will put the cost somewhere: interrupted supply, management attention, the supervision nobody budgeted for. Whether or not it’s a term in the contract, water will find its way. An architect who cut $140,000 out of the contractor’s price, leveraging her awareness his schedule needed filling, nailed it in Kolb and Williams’s Everyday Negotiation: “That extra supervision time was the trade-off I made when I pushed only on price.” You have an interest in the accuracy of their numbers, including if they favor you. Ignoring reality is naïveté too.
A Few Do’s and Don’ts
Don’t Read precision as evidence that something is a fact
$47.3 million coming out of a model doesn’t make it any more factual than one that says forty to fifty-five. This is false precision, suppressed uncertainty masquerading as a measurement. Don’t fall into the trap of clean mechanisms laundering dirty numbers: even if you agree to split the difference—not recommended—it will not automatically produce a fair outcome if suspect numbers go in.
Do Test whether a fact is meaningful before you let it drive decisions
For a fact to drive a term of the agreement, being true is not enough. It has to track the thing you want to have a mechanism for. An American parts maker wanted to tie wages to its profit margins; the union pointed out margins fluctuate for many reasons (including accounting choices). Good drivers are representative, timely, verifiable by all parties, and not controlled by a party.
Don’t Forget to scrutinize your own numbers as much as theirs
Just because they’re yours, doesn’t make your walk-away, forecasts and analyses any more factual than theirs. Critical fact-checking is an agnostic discipline. Your own side is not immune to biases, and they don’t show up in the quality of the reasoning. You may be able to structure for it, by employing estimates and models from sources who don’t have an interest in the outcome, or don’t know upfront which side they’ll end up with.
Do Price a claim you cannot check
When the other side talks up a number you cannot verify, an alternative to contesting it is building terms around it. If they want low pricing because they expect high volumes, create a volume-dependent price schedule. Both outcomes work for you: if they accept, they are held accountable for their own forecast; if they decline, their “fact” loses credibility and their evasive argumentation will reveal their position.
Pair With
Relative Advantage (2.2). Comparison is only as good as what it compares. An argument about which option is preferable can be impeccable logic on worthless grounding. The party establishing the facts has already won half the argument that follows.
Track Record (1.1). Not a parallel consideration but the successor pattern. When decision criteria and thresholds are known, assumptions and scenarios tend to converge on them. This makes the submissions indistinguishable, shifting the decision to reading the people who submitted them.
Reversibility (6.3). Facts harden because reopening them gets expensive, not because they get truer. The size of changes actors are still willing to raise grows bigger as the process goes on. A number can be an estimate on Monday and a fact on Friday.
Things to Explore
Book
David McCullough, The Path Between the Seas (1977), the chapters on Washington in 1901–02. The Isthmian Canal Commission’s November 1901 report recommended Nicaragua as the location for the canal. However it did so conditionally: on engineering grounds, Panama was the better route—but only at $40 million to the French company that owned the half-built works; not the $109 million wanted. The French lowered to $40 million on 4 January 1902; and yet, on 9 January the House of Representatives voted 308 to 2 for Nicaragua on a wrong reading of the report. Search handle: Walker Commission supplementary report.
Case study
The 1973 Major League Baseball Basic Agreement, criteria for salary arbitrators. Two paragraphs listing what an arbitrator may take into account and then, more usefully, what is excluded: the financial position of either party, press comment and testimonials, and any offer either side made before the arbitration began. That last exclusion exists to ensure how you negotiated cannot be used against you once negotiation failed. Parties can agree the rules of evidence long before they have anything to argue about.
Article
James Sebenius, “The Computer as Mediator,” Journal of Policy Analysis and Management 1:1 (1981). A model of deep-seabed mining economics, built at MIT for an American agency and another purpose entirely, ended up shaping the United Nations Convention on the Law of the Sea (UNCLOS). Tommy Koh, the chairman of the negotiating group, introduced it; delegates then used the model as an evasion tactic, hiding behind an “objective” calculation. Paul Engo of Cameroon, chair of the committee, complained delegates had been reduced to “spectators in the inconclusive tournament among experts”. An objective instrument is not necessarily neutral or innocent.
And one you would not expect
The Large Area Crop Inventory Experiment, 1974–78. When satellite imagery of wheat fields became available, somebody had to establish whether they accurately represented reality. Field crews with tape measures established ground truth, a feature of remote sensing since the late 1960s. Field measurements were gathered independently of the classification they were checking, and gathered afterwards, so that the check could not be contaminated by what it tested. That is tighter verification discipline than most negotiation books describe.
Almost simultaneous to “The Great Grain Robbery” but too late to make a difference, on 23 July 1972, the Landsat-1 observation satellite was launched from Vandenberg Air Force Base in California. Fourteen months after the event, Congress required every exporter to report its foreign grain sales.
Ground Truth is pattern 2.1 of twenty-seven. The two axes, the nine categories and the full set are laid out in The Pattern Language.