IGNITE EXTENDED / LESSON 38
How can differences in priorities help both parties?
A package comparison and a communication-channel rationale.
A multi-issue negotiation can contain issues on which interests align, issues with directly opposed preferences and issues each party values differently. Treating every issue as distributive can leave mutually beneficial trades undiscovered.
The New Recruit payoff tables make the distinctions explicit. Job assignment can have an outcome both sides prefer. Other issues provide opportunities to trade something relatively inexpensive to one side for something more valuable to the other. Understanding the priorities requires information about the package rather than only the headline salary.
The source’s bonus schedule gives the candidate 4,000 points for 10% and the recruiter 0. At 2%, the candidate receives 0 and the recruiter 1,600. Vacation has a different pattern: five days give the recruiter 4,000 and the candidate 0; twenty-five days give the recruiter 0 and the candidate 1,600.
Holding other issues fixed, high bonus with low vacation produces 8,000 joint points. Low bonus with high vacation produces only 3,200. Trading across the issues can create more joint value because the parties’ relative priorities differ.
Claiming value concerns distribution. Creating value finds combinations that improve the joint outcome because the parties value issues differently.
Improving your share of the available value.
Finding arrangements that produce more joint value through differences in priorities, timing, risk or capabilities.
Illustrative application: a client values early delivery more than extended support, while a supplier finds basic support less costly. An early/basic package can outperform a compromise made separately on each issue. Price can then allocate some of the additional value between them.
Explore priorities and packages before assuming every issue has equal and opposite importance. A useful joint result must still be acceptable to each party.
The full payoff table for eight issues with recruiter and candidate points. Job assignment and location are highlighted as compatible: both sides score highest on the same option (Division A; San Francisco). The maximum joint outcome is 13,200 points.
Compatible issues need no trade, only discovery. Pairs who assume opposition can still settle on a worse division or city, a loss for both that no bargaining skill on other issues can recover.
The same table with bonus and moving expenses marked as mattering more to the candidate, and vacation and insurance marked as mattering more to the recruiter. Salary and starting date are distributive.
The value comes from relative priorities, not absolute preferences. The candidate gives up vacation, which costs her little, to get bonus, which is worth a lot to her, and the recruiter does the reverse.
The payoff table with the conclusion: the maximum joint outcome is 13,200 points, while splitting the difference on every issue yields a joint outcome of only 4,800 points.
Compromise issue by issue destroys 8,400 points (64% of the available value) even though it looks fair. In the table, the split takes Division C on job assignment, costing both sides 1,200, and the middle options on every integrative issue.
A plot of efficient agreements in thousands of points, running from (-4.8, 13.2) through (2.4, 10.8), (6.6, 6.6) and (10.8, 2.4) to (13.2, -4.8) for recruiter and candidate.
Efficiency does not settle distribution: the points range from a large win for one side to a large win for the other, and only the central part reaches the 13,200 joint maximum. Find the frontier first, then argue about where on it to land.
Two frontiers on the same axes: a straight line for distributive bargaining, where one party's gain is exactly the other's loss, and a curve that bulges outward for integrative bargaining, because over some ranges one party gives up little to create a lot for the other.
The space between the curve and the line is the value created by trading on differences in priorities. Pure distributive negotiators can only move along the straight line.
An agreement is Pareto efficient within the modeled alternatives when no available change can improve one party’s outcome without worsening the other’s. The source frontier plots combinations of recruiter and candidate outcomes.
Efficiency does not identify a unique fair allocation. Several points can lie on the frontier and divide value very differently. A point inside the frontier suggests an unexploited improvement. Search for a better package before focusing exclusively on how to divide the current package.
Parties may disagree about a future outcome while both believe their own forecast. A contingent term specifies what happens if an observable condition occurs. This can bridge a disagreement without forcing the parties to agree on one probability estimate.
The condition must be measurable, verifiable and difficult to manipulate. Consider who controls the outcome and whether the payment creates a harmful incentive. The communication-channel material adds another dimension: the medium can affect cues, reflection and the risk of misinterpreting the counterpart.
A contingent term links part of the agreement to an observable future event. It can address different forecasts when measurement, incentives and enforcement are workable.
One party expects a high demand level; the other expects a lower level.
Condition part of the agreement on an observable future outcome, if it can be measured and enforced sensibly.
Illustrative application: a buyer expects low demand and a supplier expects high demand. A base quantity plus a defined additional payment or quantity if verified demand crosses a threshold may let both act on their beliefs. Specify the measurement period, data source and who can influence the measure.
A contingency can also create disputes or distorted incentives. Evaluate the rule and observability before treating it as a solution.
Parties often disagree about the probability of future events or differ in risk and time preferences. They can avoid impasse and improve outcomes by adding contingent clauses that condition future actions on agreed criteria: if X, then Y.
A contingency turns a disagreement about forecasts into a tradable term: each side bets on its own forecast. It works only with a measurable, verifiable condition that neither side can easily manipulate.
A chart from Dannals, Zlatev, Halevy and Neale (2021, Journal of Applied Psychology; 2,552 negotiators including executives, working adults and students). With weak alternatives, women and men extracted similar points (about 4,870 and 4,790); with strong alternatives, men extracted more (about 5,750 versus 5,370).
The gender gap appeared only when alternatives were strong: having power did not translate into value equally. Organizations can reduce gaps by making alternatives and ranges explicit, for example through pricing guidance for every account executive.
Channels placed on two axes: synchronous versus asynchronous, and visual and vocal cues versus text only. Face-to-face and video calls are rich and synchronous, voice mail is asynchronous with vocal cues, and email is lean and asynchronous.
Channel choice is two separate decisions: how many cues (sight, sound) and whether replies are immediate. Each can be chosen for a stage of the negotiation.
Effect sizes (Cohen's d) from Swaab, Galinsky, Medvec and Diermeier's meta-analysis (Personality and Social Psychology Review, 2012; the slide footer says 2013). With a neutral orientation, visual, vocal and synchronous channels improved outcome quality (d = 0.51, 0.53, 0.58); with a cooperative orientation, effects were near zero (0.14, 0.08, 0.04); with a noncooperative orientation, they hurt (-1.05, -0.38, -0.57).
Visual cues hurt most when parties are competitive (d = -1.05). Video is the right default for a neutral first meeting and the wrong one when a dispute has turned adversarial.
Lean channels can be beneficial when you are the weaker or lower-status party, when you need time to think through trade-offs, when the relationship is competitive and distance allows de-escalation, when both sides are committed to the process, and when stepping back helps the parties focus on the forest rather than the trees.
For a multi-issue package, a written proposal gives both sides time to evaluate trades they would reject if pressed live. Use rich channels to learn priorities and lean ones to compare packages.
The archived negotiation lecture distinguishes distributive, compatible and integrative issues in a multi-issue negotiation. Its central lesson is to exchange and interpret information about priorities rather than automatically split every issue. The exercise here uses a new self-contained packet.
Which trade creates more value with relatively little sacrifice?
In a three-month 2021 pilot in Canada, Walmart used Pactum's chatbot to negotiate with tail-end suppliers of goods not for resale, such as carts and store equipment. The bot traded issues (for example, longer payment terms in exchange for concessions on Walmart's right-to-terminate clause) and closed deals with 64% of suppliers, saving 1.5% on average. Expanded programs later reported 68% and 3%. The bot was multi-issue by design.
Intercom prices its Fin AI agent at $0.99 per outcome; a resolution counts when Fin answers and the customer confirms it helped or leaves without asking for more help. Intercom chose this over seat pricing because Fin's value grows as it reduces support headcount and because early customers doubted it would work, the same difference in forecasts that contingent contracts are designed to bridge.
Relay could offer Northpoint a per-filled-slot price alongside the flat fee. At 40% of about 42 open slots a week, a location sees about 72.8 fills a month, so $339 is equivalent to about $4.66 per fill against roughly $110 of visit revenue (illustrative). A precise definition of a 'fill' is what makes the term enforceable.
Sources: Harvard PON: Chatbot negotiations, what can AI do for you? · Talking Logistics: Negotiating with a chatbot, a Walmart procurement case study · Stripe: Intercom innovates outcome-based pricing for its Fin AI agent · Intercom pricing (learning center)
Use the four options and point values on these slides. The client values early delivery most; the supplier values basic support most. Start with no price issue.
Compare all four packages. Then add a fictional transferable price adjustment that could redistribute 20 points between the sides without changing joint value, and discuss acceptance.
Lesson 37 set Relay's limit at about $209/location/month and its target at $339; Northpoint asked for $250. Discovery calls revealed six issues beyond price. Illustrative 3-year values in $k, relative to a $250 contract on Relay's standard terms (Relay / Northpoint): data rights to de-identified data for model training +90/-15; case study and reference rights +45/-5; 99.9% uptime SLA with service credits -20/+50; termination for convenience after year 1 -60/+40; fill-rate guarantee (price steps down to $250 if fill rate is below 30% in months 4-9) -10/+35; two-wave rollout (12 locations, then 30; all 42 billed from month 1) +20/+60. Each $1/location/month moves $1.512k between the parties.
Classify each issue, compute Relay's and Northpoint's values for two packages (A: a $295 compromise with SLA and termination for convenience; B: a value-creating package of your design), find the price at which Northpoint is indifferent between them, and choose a channel plan for presenting the packages.
Agreeing price first and then 'the details' makes every later issue a separate tug-of-war and removes the currency for trades. Keep all issues open and exchange packages.
To create value, tell Northpoint that data rights matter a lot to Relay and SLA terms matter less. Do not tell them $209. Priorities enable trades; limits only invite the counterpart to claim value.
'Price steps down if Relay underperforms' invites a dispute. Define the metric (fill rate = refilled slots / late-cancellation slots), the period, the data source and the remedy, and check that neither side can game it.
A Northpoint package comparison: issue classification, package A (+$10k joint) versus package B (+$250k joint), Northpoint's indifference price of about $318, a contingent fill-rate term, and a channel plan.
Next lesson: The business narrative & evidence portfolio
Check your understanding
Question 1
In New Recruit, the maximum joint outcome is 13,200 points and splitting the difference on every issue yields 4,800. What share of the available joint value does the issue-by-issue compromise give up?
Lost value = 13,200 - 4,800 = 8,400 points; 8,400/13,200 = 63.6%. The 36% option is the share retained. Compromise feels fair on each issue but ignores differences in priorities and includes a mutually worse job assignment.
Question 2
Northpoint values termination for convenience at +$40k, and it costs Relay $60k. A fill-rate guarantee is worth +$35k to Northpoint and costs Relay $10k. What does swapping the first for the second do to joint value?
Old term: -60 + 40 = -20 joint. New term: -10 + 35 = +25 joint. Change = +45: Relay +50, Northpoint -5. Relay can pay Northpoint back the 5 (for example, about $3.31/location/month on price) and still keep about $45k.
Question 3
Relay believes its fill rate will be 40%; Northpoint's CFO thinks 25% is more likely. Price talks have stalled at $295 versus $340. Which term best uses the disagreement?
Each side expects to win the bet, so the contingent price is worth more to both than a fixed midpoint. The term must be measurable and hard to manipulate, hence a defined metric and auditable data source. Splitting the difference ignores the difference in forecasts.
Question 4
A team reaches an agreement that is on the Pareto frontier. Which statement is correct?
Efficiency is about waste, not fairness. The New Recruit frontier runs from (-4.8, 13.2) to (13.2, -4.8): efficient points can be lopsided and even below one party's alternative. Distribution and BATNAs remain separate questions.
Question 5
The Northpoint talks became adversarial after a failed pilot demo. Maya plans a two-hour video call to 'clear the air and close.' Based on the channel meta-analysis, what is the better plan?
Swaab and colleagues found channel richness helps neutral parties and hurts competitive ones. A written proposal lowers the temperature and gives time to evaluate trades; a call can follow once the orientation is less competitive. 'Always' is the misconception.
Question 6
In Relay's package B, Northpoint's non-price terms are worth +$125k and package A gives Northpoint +$21.96k. With each $1/location/month worth $1.512k, at what price is Northpoint indifferent between B and A?
-1.512 x (P - 250) + 125 = 21.96, so P - 250 = 103.04/1.512 = 68.15 and P = $318.15. $295 is package A's price; $333 ignores the 21.96 Northpoint already gets under A.
Workbook
Module output: A negotiation brief, package comparison and individual decision portfolio.
Sources
Original pages and full documents are in the Library edition.