Three valuation methods, applied to the same twenty-year archive, produced numbers a factor of four apart. Only one of them survived the buyer's due diligence, and it was not the one the publisher opened with.
The request arrived in the ordinary way. A model developer wanted a licence covering the full text archive, offered a number, and gave six weeks to respond. The publisher's first instinct was to price the archive by what it cost to produce: two decades of salaries, wire fees, and photography, discounted for age. That figure was large, defensible in a spreadsheet, and irrelevant to the buyer, who was not proposing to reproduce the archive. They already had most of it.
Three ways to price the same thing
Cost to replace answers a question nobody asked. Comparable deals are better, but the comparables are private, partial, and structured differently — a per-article rate in one contract is a lump sum with a training carve-out in another. Forgone traffic is the only method anchored in the publisher's own numbers: what the archive earns today, and what it stops earning if the licence goes ahead.
Once the third method was on the table, the conversation changed shape. The publisher stopped defending a production cost the buyer had no reason to accept, and started pricing a specific, measurable loss. That number was smaller than the opening ask and considerably firmer, because every line of it came from the publisher's own analytics rather than from an argument about effort.
Term length is where the money moves
Publishers argue about price and concede on duration, which is the wrong way round. Price is settled once. Duration compounds: it fixes today's assumptions about referral traffic, model capability, and market structure for as long as it runs. In this engagement the board approved a walk-away number and, more usefully, a maximum term. The deal closed at three years instead of five, at a price close to the buyer's original offer.
That looks, on the day, like a concession on price. Measured over the life of the contract it was the larger win, and it is the part of the negotiation nobody writes a press release about.