Negotiators look for issues the other side values more than they do and trade on those, so both leave with more than a split of the original sum would give.
An example
Year-end was the sticking point between a bank's head of treasury operations and its chief information officer over a core payments system migration. IT wanted to cut over in the last week of December, when volumes were low. Treasury refused, because year-end is when the bank's largest corporate clients move the most money. Each meeting was a fight about one date. The treasury head finally asked the chief information officer what else mattered to his team, and learned that IT valued a long weekend window and two treasury analysts assigned to testing for six weeks. Treasury valued protection for month-end and quarter-end dates and a rehearsed rollback plan. They traded: cutover on a long weekend in mid-February, two analysts seconded for testing, a frozen schedule around month-end, and a rollback rehearsal treasury could attend.
Why it works
Trading across issues makes the deal look better to both sides at once. When a negotiation is about a single point, every gain for one side is a loss for the other. Once more issues are on the table and the sides value them differently, each can give what costs it little and get what it values most. It depends on discovering what the other side cares about, which means asking and listening, and on having more than one issue in play.
The thinking behind it
David A. Lax David Lax and James Sebenius's The Manager as Negotiator: Bargaining for Cooperation and Competitive Gain (1986) treats negotiation as both creating value, enlarging what is available through joint gains, and claiming it. Sharing information helps create value but exposes a party to having value claimed from it, the negotiator's dilemma. The treasury head's question about what else mattered to IT created value, and it is why the page warns that revealing priorities can be exploited.
Max H. Bazerman Bazerman and Neale's Negotiating Rationally (1992) names the mythical fixed pie as a common mistake: negotiators assume their interests are directly opposed, so whatever the other side gains they lose, and they miss trades that would leave both better off. A fight over one cutover date is the fixed pie in miniature. Adding testing staff, weekend windows, and rollback rehearsals showed that the two sides' priorities differed.
When it fits
Any negotiation that has stalled on one issue is a candidate, especially between internal units who must keep working together. You need curiosity about the other side's interests, enough authority to trade on several issues, and time to explore them. One-off deals with a counterpart who will not share what it values leave little to work with.
How to use it well
Stop arguing the stuck issue for a while and ask what else matters to the other side, and why.
List every issue in play, from timing and people to money, risk, scope, and credit. Rank them for yourself, and estimate how the other side would rank them. Propose trades where you give on what is cheap for you and dear to them, and put several issues into one package so trades are possible. Share some of your own priorities to draw out theirs, since holding everything back makes the other side do the same.
What it costs and where it backfires
Revealing what you value can be exploited by a counterpart who takes the information and gives nothing back. Exploring issues takes time, and a bigger package can be harder to get approved. State your priorities truthfully: pretending to care about something so you can trade it away later, or misstating what an issue costs you, is deception, and it poisons the next negotiation.
When it is used on you
Share some of what you value when the other side asks, and ask the same of them. Compare what they offer against your own ranking, and make sure what they give truly matters to you. If they reveal nothing, slow the trading until they do.