Sales wants the feature shipped this quarter; engineering wants it built properly; finance wants the headcount frozen. Each is right from where they sit, and as long as each defends its position, the meeting ends in a draw or a grudge. Managing competing agendas is the leadership craft of surfacing what each side actually needs underneath what they are demanding, and assembling a decision that serves enough of those needs that people will own it rather than resist it.

The quick version

  • A position is what someone says they want ("ship it in June"). An interest is the underlying need driving it ("hit our number / not get blamed for a missed quarter"). Stalemates form between positions; solutions form between interests.
  • Mary Parker Follett named three ways to handle a clash: domination (one side wins), compromise (both give something up), and integration (a new option that meets both real needs). Integration is the prize, but it is not always available.
  • Before you negotiate, map the stakes: who has the power to block this, who cares most, who is merely watching. Effort spent on the wrong stakeholder is effort wasted.
  • The trap is rushing to compromise, splitting the difference feels fair but often leaves everyone half-served and the disagreement free to return in a new form.

The idea in depth: positions versus interests

The single most useful move in any clash of agendas is to stop arguing about what each side wants and start asking why. That distinction, position versus interest, is the spine of Getting to Yes (Fisher, Ury & Patton, Harvard Negotiation Project, 1981). Their method of principled negotiation rests on four habits: separate the people from the problem, focus on interests rather than positions, invent options for mutual gain, and judge the result against objective criteria. The load-bearing one is the second: positions are finite and collide; interests are plural and often overlap more than either party expects.

Ask, in the room, "What would solving this make possible for you?", and keep asking until you reach a need rather than a demand. Sales saying "ship in June" may really need a credible date to give a key customer; engineering saying "not until Q3" may really need to avoid shipping something that pages them every weekend. Those two needs are not actually opposed. A dated beta for the one customer, with the hardened release in Q3, can serve both, an option invisible while everyone defended their date.

An honest limitation. Interest-based bargaining assumes there is enough overlap to integrate, and that both sides will negotiate in good faith. Sometimes interests are genuinely zero-sum, one promotion, two candidates, and no amount of reframing conjures a win-win. Sometimes the other party is using competition as a tactic. Treat "find the shared interest" as the first thing to try, not a guarantee; when it fails, you fall back to fair criteria and, if needed, your best alternative.

Follett's three responses, and why compromise is the booby prize

Decades before Harvard, the management thinker Mary Parker Follett set out the cleanest map of what a leader can do with a difference. In her 1925 lecture "Constructive Conflict" (collected in Dynamic Administration, eds. Metcalf & Urwick), she argued conflict is not inherently bad, it is simply the appearance of difference, and difference is where new value comes from. You can respond to it three ways: domination, where one side imposes its will and breeds quiet resentment; compromise, where each side surrenders something and the underlying tension resurfaces later; or integration, where you invent a solution in which neither side has to give up what it truly wants.

flowchart TD
  A(["Two competing agendas"]) --> B{"How do you
respond?"}
  B -->|"One side imposes"| C(["Domination
winner + a resentful loser"])
  B -->|"Each gives something up"| D(["Compromise
both half-served, tension returns"])
  B -->|"Invent a new option"| E(["Integration
both real needs met"])
  C -.->|"conflict re-emerges"| A
  D -.->|"conflict re-emerges"| A
Follett's three responses to a difference, only integration tends to settle the matter for good. Leaders Loop

Her famous illustration is mundane on purpose. Two people sit in a library room; one wants the window open for air, the other wants it shut to keep out the north wind. Compromise opens it halfway and satisfies no one. Integration opens a window in the next room, air enters, no draught crosses the desk, and neither person curtailed a single real desire. The point lands because it shows that the apparent conflict (open vs shut) was never the real one (fresh air vs no draught). Note the lineage: this library-window example and the interests-over-positions idea are widely credited as a direct antecedent of Getting to Yes, a reminder that the "new" negotiation canon is older than it advertises.

Treat splitting the difference as a fallback, not a reflex. Before you halve anyone's ask, spend five honest minutes hunting for the next-room window: a third option that dissolves the trade-off instead of dividing the pain. You won't always find one. But you'll find one far more often than a culture of reflexive compromise ever lets you look.

Compromise settles the argument; integration settles the problem.

Map the stakes before you mediate

Competing agendas are rarely a clean two-party affair. There is the person demanding the feature, the one who quietly controls the budget, the team that will live with the result, and three others copied on the email who care less than the volume of their replies suggests. Spending the same energy on each is how good mediators exhaust themselves. A simple, durable tool here is the power–interest grid, attributed to Aubrey Mendelow (early 1990s): plot each stakeholder by how much power they have to shape the outcome and how much interest they have in it. The four quadrants tell you where to put your effort, manage the high-power, high-interest players closely; keep the powerful-but-disengaged satisfied; keep the interested-but-powerless informed; and merely monitor the rest.

quadrantChart
  title Power-interest grid
  x-axis "Low interest" --> "High interest"
  y-axis "Low power" --> "High power"
  quadrant-1 "Manage closely"
  quadrant-2 "Keep satisfied"
  quadrant-3 "Monitor"
  quadrant-4 "Keep informed"
Mendelow's power–interest grid: match the depth of your engagement to where a stakeholder sits. Leaders Loop

Draw the grid before the meeting, not after the deadlock. Two minutes naming who genuinely has the power to block this, and who is loud but, on this decision, merely interested, tells you whose interests you must integrate and whose you only need to acknowledge. It also heads off a common failure: negotiating hard with the person in front of you while the real decision-maker, parked in the "keep satisfied" corner, never gets brought along.

Where the grid stops helping. It is a snapshot, not a law of physics. Power and interest shift, a "monitor" stakeholder becomes a blocker the moment the decision touches their budget, and the model says nothing about relationships between stakeholders or how to actually move someone. Use it to allocate attention, then do the human work the matrix can't: ask people what they need, and listen for the interest under the position.

A worked example

Take a product org, call it Meridian, where the platform team and the growth team are locked over the next quarter's roadmap. (Illustrative throughout; not a real company.) Growth's position: "We need the new onboarding flow live by end of August or we miss the back-to-school surge." Platform's position: "We're mid-migration; we can't take new feature work until October without risking an outage." Two dates, flatly opposed. The instinct in the room is to compromise on mid-September, which, predictably, satisfies no one: growth misses the surge anyway, and platform ships under pressure it warned against.

The leader running it does three things instead. First, she maps the stakes: the VP of Marketing (high power, high interest, manage closely) actually owns the surge target; the platform director (high power) cares about uptime, not the calendar; a vocal PM is loud but, on this call, low-power. Second, she digs for interests. Growth's real need is not "August", it is hitting the back-to-school number. Platform's real need is not "October", it is not being the cause of an outage during migration. Third, she hunts for the next-room window: ship a lightweight onboarding variant on the existing stack for the surge (no platform dependency), and build the full flow on the migrated platform for October.

flowchart LR
  A(["Growth: 'Live by August'
interest: hit the surge"]) --> C{"Find the
shared option"}
  B(["Platform: 'Not until October'
interest: no outage"]) --> C
  C --> D(["Lightweight flow on old stack
for the surge"])
  C --> E(["Full flow on migrated platform
for October"])
  D --> F(["Both real needs met
integration, not split"])
  E --> F
Naming the interest under each date turned a flat either/or into a sequence both teams could own. Leaders Loop

Neither team gave up what it actually needed. That is the difference between a decision people execute and one they quietly undermine, and it came not from authority or charm but from a discipline: separate position from interest, map who really matters, and look for the option that dissolves the trade-off before you reach for the one that divides it.

Frequently asked questions

Isn't this just "win-win," which doesn't always exist?

Partly, and the honest answer is that integration isn't always available. Some conflicts are genuinely zero-sum, and pretending otherwise wastes everyone's time. The discipline is to try for the integrative option first, because it's available far more often than positional arguing reveals, and to fall back to fair criteria or compromise only once you've genuinely looked and found no shared option. "Always look, don't always expect to find" is the realistic stance.

What if the other side won't move off their position?

Don't counter-push, which just hardens them. William Ury's follow-up work, Getting Past No, calls the first move "going to the balcony", stepping back from your own reaction, and then reframing their attack as a problem to solve together: "Help me understand what you'd need for this to work." Open, problem-solving questions pull a fixed position back toward the interest underneath it more reliably than argument does.

How is this different from just compromising?

Compromise divides a fixed pie, each side gives something up, and the unmet bit of each agenda tends to resurface later as the same fight in new clothes. Integration looks for a different option entirely, one where neither side surrenders a real need. Follett's window is the test: if your solution still requires both people to want less than they did, it's a compromise; if it gives both what they actually came for, it's integration.

Where do I start when there are six stakeholders, not two?

Map them on power and interest before you engage anyone. Concentrate your integrative effort on the people who both care about the outcome and can shape it; keep the powerful-but-disengaged informed enough that they don't block you late; and don't burn energy negotiating with people who are loud but, on this decision, hold neither power nor a real stake. Multi-party clashes are usually two-party clashes plus an audience, find the two.

Doesn't all this listening just slow decisions down?

It front-loads time and saves it later. A decision reached by domination or a grudging split tends to leak, re-litigated in hallways, slow-walked in execution, reopened at the next planning cycle. Spending twenty minutes surfacing interests so people own the outcome is usually cheaper than the months of friction a half-served compromise generates. Speed at the decision is a false economy if it costs you speed at the doing.

Related in the Toolkit

Managing competing agendas leans hardest on two neighbours: the ability to move people when you can't simply order them (influence without authority), and the structured craft of bargaining itself (negotiation), where interests, BATNA and ZOPA get their full treatment.

Where to go next