Writing the rails
Jane Austen’s houses don’t fall apart because anyone is wicked. They wobble because no one names the method for deciding anything. A dance is more than steps; it’s tempo and who leads when. In Pride and Prejudice, “amiable” means one thing at a ball and another in a letter—meanings drift because the household has no shared playbook.
Business partners repeat this mistake at scale. When pressure rises and the method is fuzzy, every move is read as a power grab. Agreements—clear, written, testable—are the rails that keep efficiency from turning into skids.
If you listen closely, you can hear the room exhale when rails appear. It’s the quiet sound of dignity returning. Rails let the bold keep their nerve and the careful keep their standards. Rails don’t drain color from a partnership; they tune it, the way a conductor steadies an orchestra before the first note. We’re going to write those rails, and we’ll keep them humble enough to fit on a single page you can spill coffee on.
The charter that quieted the room
The brand lead felt diluted; the revenue lead felt stalled. We wrote a decision charter: the one-line question, roles by name, standard of proof, timeline, and definition of done. It took nine minutes. The volume dropped without anyone being shushed. The room recognized itself in writing, and the debate turned into design.
The pricing floor that melted at noon
Two founders led a fintech platform through a delicate quarter. Sales begged for “flex” on price. Finance begged for “discipline.” On Tuesday at standup, one founder said, “Let’s be conservative until end of month.” Heads nodded in three different ways. By noon, three separate deals had three separate discounts. “Conservative” meant “protect cash” to finance, “win logos now and make it up on expansion” to sales, and “keep the brand premium” to marketing. Everyone thought they were being good soldiers. The quarter bled anyway.
We stopped the knife fight long enough to write a one-page decision charter for “pricing exceptions.” Names, not departments: who recommends, who advises, who decides, who carries execution. We defined “conservative” for this arena—low spend, quiet visibility, narrow scope. The fights didn’t vanish; they shortened and softened because the method was visible. People stopped reading each other’s motives and started reading the page.
What agreements are (and are not)
Agreements are small, public promises about how we choose under pressure. They specify rights (who decides), standards (what “good” looks like), and rhythm (when and how we revisit). They are the difference between a room that tightens when news arrives and a room that reaches for its tools like a practiced kitchen—sharp knives, clean boards, dinner on time.
They are not bureaucracy or therapy. They do not ask you to be less yourself. They ask you to put your best self on rails when the weather turns. Rails do not slow a train. Rails make efficiency possible.
The core kit (use these in any order, but carry them all)
1) Decision charters (two pages, max)
- Question in one sentence.
- Roles by name (not department): recommender(s), advisors, decider, DRI for execution.
Clarifier — When to use which
Use DRI when one brain can own the outcome end-to-end. Use RAPID when legitimacy requires multiple voices with distinct roles. If you’re stuck, start DRI and declare explicit input from two named functions.
- Standard of proof (what data, what threshold, by when).
- Timeline and communication plan (who hears, how, and when).
- Definition of done for the first sprint of work.
Leader script: “Before we argue taste, let’s write the definition of done and the roles for this call. Then we test options against it.”
Summary
Think of a charter as a pocket stage manager’s script. When the cue hits—market turns, platform note lands, board asks for numbers—you don’t improvise your way through a lighting change. You flip to the page where it says: who calls the cue, which lever is pulled, how long the blackout lasts, what the audience should feel next. A charter gives your judgment a shorter path to the stage.
Micro-case
At a consumer brand, “influencer partnerships” kept sprawling. A two-page charter put one founder as decider, growth and brand as advisors, and a single marketing manager as DRI. Standard of proof: a 14-day holdout test showing a lift on first purchase plus guardrails on tone. The charter didn’t end debates; it ended the reenactment of the same debate every Thursday.
2) Consent beats consensus
Not every call needs unanimous enthusiasm. Many require the willingness to try together.
Consultative decision promise
“I will decide by Friday at 2 pm after hearing two clear preferences and one risk each. I’ll carry conditions forward and set a review date.”
Quick method
“I’m proposing Option B for three weeks with a review on the 15th. On a 1–5 scale, can you live with this—and what would make it safer to try?”
Record the numbers and any conditions. Read them back. Decide. Move.
Consent gradient
Inform → Consult → Consent → Decide
Say which level you’re using before debate begins; it stops the meta-arguments about “how big a deal is this.”
Summary
Consent is a humane efficiency tool. Consensus is a quilt—lovely, slow, and often lumpy. Consent is a trail marker—clear enough to move, close enough to check again. When you ask for a 1–5 and listen for what would raise a score by one, you’re not counting votes; you’re catching conditions that keep the wheels on. The respectful part is not the number—it’s reading back the conditions aloud so the room hears its own wisdom.
3) Reversibility (name it out loud)
- Reversible: easy to roll back; bias to action.
- Partially reversible: some cost to undo; stage-gate it.
- Irreversible: slow down; raise proof; document the why.
Pair consent with reversibility: “This is consultative and partially reversible; we’ll run a four-week pilot with guardrails.”
Summary
Founders love bold strokes. Reversibility lets boldness be responsible. It is the difference between a haircut and a tattoo. When you say the word aloud, people stop performing certainty and start designing safe-to-learn experiments. I’ve watched a room unclench when a leader says, “This is reversible. We’ll know by Friday if the canary coughs.”
4) Option briefs (two pages, not decks)
Each option gets:
- A short name.
- Two assumptions we’re making.
- One risk we accept and how we’ll watch it.
- The definition of done (two lines a neutral observer can verify).
- A pre-mortem paragraph: “If this fails in six months, it will be because ___.”
Why it works: Options turn ideology into design. Once there are named plans on the table, people compare instead of perform.
Summary
A good option brief feels like a well-packed carry-on: only what you’ll use, easy to grab under stress. Naming the option matters. “Harbor” vs. “Sprint” changes the room’s posture. Asking each exec to champion the option they like least is not a game; it is empathy turned into craft. The brief you don’t love gets smarter, and so do you.
5) Decision notes (write them while you’re still together)
A one-page entry in a single, living decision log:
- Decision + date
- Options considered & why we chose this one
- Consent scores & conditions by function
- Definition of done (first sprint)
- Risk owner & review date
Six months from now, you won’t remember how sure you felt. You will remember why you chose, if you write it.
Summary
Decision notes are your breadcrumb trail out of the forest. They prevent folklore from hardening into “truth.” They also act like a small apology baked into the system: we might be wrong; here’s when and how we’ll check. New leaders learn your mind by reading the last ten entries; they don’t have to borrow it in the hallway.
6) Escalation ladder (three rungs, no drama)
- Partners run FLO for 30 minutes with a chair (mirror once each; smallest true problem; three options).
Chair move — The time-boxed pause
“Heat’s at a four. Ninety-second reset. We’ll return to the question with one mirror each.” You will feel like a librarian. Then you will feel like a leader.
- Neutral internal decider (domain context, time-boxed call after two option briefs).
- External mediator/advisor (scheduled, not ad hoc; decision or recommendation).
Publish the ladder. Rehearse it quarterly on a harmless scenario so it’s there when you need it.
Summary
Ladders save friendships. Without one, stuck choices ricochet between Slack and side conversations until someone with a bigger title fires off a decree. With one, dignity stays intact: we tried our tools, we asked a neighbor, then we brought in a calm outside voice. No theatrics, just gravity and rungs.
7) Two-voice promises & lock windows
- Two-voice: any external commitment that touches revenue, legal exposure, or capacity requires two named voices before it leaves the building.
- Lock windows: before launches or public moments, declare the period where changes are off-limits by default (exceptions: safety, law)—and name who can grant them.
These two protect brand, sleep, and trust.
Summary
Two-voice promises are not red tape; they are a marriage vow for public statements. They keep a heroic salesperson from mortgaging operations, and they keep a cautious lawyer from smothering momentum. Lock windows are mercy disguised as policy. They give teams one quiet hallway before the curtain rises.
Scenes (where agreements saved the room)
Film financing under public pressure
An indie studio faced a choice: race with a streamer, wait for a European fund, or pass and back two smaller films. The option briefs forced honest math and honest fear. The streamer brief named a shadow cost: foreign rights stripped could starve the next slate. Consent scores landed at 4 and 3—good enough to move—with two conditions read into the decision note: crew protections and a financial firewall around the next two projects. The film delivered, the crew stayed, and—quietly, importantly—the partnership’s belief in itself rose a notch because the method had been fair.
Enterprise software’s two-platform gamble
A B2B team argued premium enterprise vs. self-serve. Safety was low; dissent lived in jokes. The chair named the reversibility (partially). Each exec had to champion the option they liked least. The pre-mortem surfaced a single point of failure: migration risk. A DRI took ownership, a review date was calendared, and the decision shipped. The surprising win arrived later: leaders began arriving at meetings with briefs unprompted. The culture had learned a new reflex.
Exercises you can run this month
1) Charter sprint (90 minutes)
- List five consequential decisions coming in the next quarter.
- For each, draft a two-page charter (question, roles, proof, timeline, comms, definition of done).
- Name an owner for each charter and the first review date.
Narrative cue: Run this like a writer’s room. Put the question line on the wall. Ask “What would a smart critic attack?” Let that shape your standard of proof.
2) Consent drill (20 minutes)
- Propose a time-boxed option.
- Gather 1–5 consent scores and “what would raise your score by one.”
- Read conditions back; choose; calendar the review.
Small warning: If scores sit at twos and no one speaks, you have fear, not consent. Switch to live voice, mirror once each, then ask again.
3) Pre-mortem in place (30 minutes)
- “It’s six months later and this failed. How?”
- Silent write 10 minutes. Cluster risks. Assign one mitigation per cluster with an owner and date.
Why silence first: It protects dissent from status. The quietest sentence is often the one that saves the quarter.
4) Ladder rehearsal (25 minutes)
- Walk a safe scenario up the escalation ladder.
- Verify handoffs, authority, and time limits.
- Adjust where it felt heavy or slow.
Make it human: Use a silly topic (“Should the office dog have a badge”) the first time. Laughter helps the muscle stick.
Scripts that travel
“Before we argue, I’m setting consent level and reversibility: consultative and reversible for three weeks.”
“Two voices on this: Sales and Finance both sign before anything leaves the room.”
“We’re at heat level four; let’s run a 90-second STOP/TIPP and then check consent.”
“Decision note goes in the log before we leave. Who’s the risk owner and when do we review?”
Narrative note
Scripts only sound stiff the first three times. Then they become a shared accent—your company’s way of being precise without getting prickly.
Common pitfalls (and clean repairs)
Faux consensus: endless talk, no owner.
Repair: name the decision pattern and the decider up front. “Consultative; Priya decides.”
Reversibility drift: you treat a reversible experiment like a tattoo.
Repair: shrink the blast radius, shorten the window, write the rollback in one line.
Definition theater: a beautiful wiki, ignored in the room.
Repair: put three hot words on the agenda. Define today, in the meeting, in one sentence each.
Leader exception: a solo promise made “to move fast.”
Repair: leader names the miss, restores two-voice, records the exception with a sunset date.
Decision leak: Slack reopens what the room closed.
Repair: post the decision note link and require clean commits (“two sentences, by end of day”).
Summary
Notice how each repair makes the agreement more visible. Visibility is the antidote to suspicion. People forgive mistakes faster than they forgive secrecy.
Measures that tell you it’s working
Decision velocity: how many consequential calls made this month? How many arrived late to the right forum?
Consent pattern: which functions sit at 2–3 on certain topics? That’s where the definition of done or sequence needs work.
Reversibility accuracy: how often did we label “irreversible” what was not?
Review follow-through: reviews held on the date? Thresholds honored?
Discuss for ten minutes monthly. Improve one item at a time. Write one paragraph about what changed because you measured. Numbers don’t build culture; stories about numbers do.
Governance that’s light and real
One log, one ladder, one charter template. If you add a document, retire two.
Owner: name a steward for charters and decision notes (a gardener, not a cop).
Quarterly governance day (90 minutes): sample three decision notes, test one value-as-policy against a real case, rehearse the ladder, confirm lock windows and two-voice pairs, publish a two-paragraph “what changed” note.
Summary
Good governance is hospitality. It sets the table so minds can eat. Keep it simple enough that a new VP can learn it in a week, and honest enough that the founders follow it in public.
Snapshot — The meeting that didn’t become five
We closed with a two-sentence commit and a review date on the shared screen. When the hallway tried to reopen the debate, the screen won. The next four calendar invites never got created.
Takeaways worth pinning
Clarity beats charisma when pressure rises. Write decision rights, standards, and rhythms so strong personalities can row in one direction.
Consent is willingness to try together, not unanimous joy; pair it with reversibility and a review date.
Option briefs cut performance and invite design; decision notes prevent revisionist history.
Two-voice and lock windows protect trust and sleep.
When the room heats, run STOP/TIPP, mirror once, then return to the rails.
Culture is simply small agreements kept in public. Keep them.
Definitions for Chapter Seven
Decision charter
A concise, two-page contract about how a specific decision will be made. Use it when a call touches money, people, brand, or timeline. It prevents turf wars by naming a decider and the standard of proof up front (“we move if CAC ≤ $X on two cohorts by Friday”). A good charter also states what will not be debated in this forum. Anti-patterns: vague roles (“marketing decides”), missing standard of proof, or a charter written once and forgotten. Try this: put the charter link at the top of every agenda while the decision is live.
Consent check
A fast scan for willingness, not enthusiasm. The 1–5 scale surfaces conditions (“I’m at a 3 if we keep the guardrail on churn”). Use it when time matters and total buy-in would slow you to a crawl. Signals it’s needed: looping debate, polite silence, or a “good enough to try” proposal on the table. Pitfall: mistaking silence for consent. Always read conditions back and capture them in the decision note.
Reversibility
A label for the cost of being wrong. Reversible choices invite small bets and rapid learning; partially reversible choices need stage gates; irreversible choices demand higher proof and a clearer story. Say the label aloud so the room sets the right pace. Example: “Pricing page headline = reversible; M&A term sheet = not.” Pitfall: treating every interesting decision as irreversible because it feels important.
Option brief
A two-page, comparable plan. It replaces opinion with design: name, assumptions, one accepted risk, definition of done, pre-mortem. Use it when disagreement hardens into identity (“I’m the quality person / I’m the efficiency person”). Ask each leader to improve the option they like least. Pitfall: straw-man options written to lose. Fix by assigning different authors and requiring a pre-mortem for every option.
Decision note / log
The live, single source of truth for what you chose, why, and when you’ll check. It travels better than memory and prevents “we never agreed to that.” Use one log for the whole company; too many locations create dueling realities. Minimum viable fields: decision & date, options, rationale, consent scores & conditions, definition of done, risk owner, review date. Habit: write it in the room while people watch.
Escalation ladder
A three-step path for stuck calls that protects efficiency and dignity. Step 1: partners run FLO with a chair. Step 2: neutral internal decider makes a time-boxed call after hearing briefs. Step 3: pre-scheduled external mediator/advisor. Use it to avoid “executive drive-bys” and side-channel lobbying. Pitfall: skipping rungs “to save time.” Practice the ladder on low-stakes topics so it’s familiar when heat rises.
Two-voice promise
A guardrail for external commitments that touch revenue, legal exposure, or capacity. Two named voices must sign before the promise leaves the building (e.g., Sales + Finance on pricing, Product + Legal on data language). It prevents hero moves that mortgage other teams. Write the pairs once per quarter so no one guesses in the moment. Pitfall: private exceptions by senior people. Fix with a short exception note and a sunset date.
Lock window
A time-boxed freeze before a public step (launch, earnings, keynote) so quality and credibility can hold. Declare start/end, allowed exceptions (safety, law), and the role who grants them. Post the calendar where everyone can see it. Use canaries and guardrails to monitor during the lock. Pitfall: hidden edits “because it’s small.” If it’s truly small, it can wait; if it can’t, it needs a written exception.
Bridge to literary reflection
Order is not fussy; it is merciful. When you give a room rails, intelligence arrives without tearing the carpet.
Literary reflection
Thoreau prized deliberate pace; Austen prized orderly rooms where character could show without chaos; Twain kept a pocketknife for language, trimming “almost right” into right; Dickinson kept her poems tight enough to hold a storm. Let them sit by your keyboard. Agreements are your house rules for hard weather. They don’t make you polite—they make you precise. Write the rails, run the train, and when the wind picks up, keep your hand on the review date. Your company will move—quicker, kinder, and still together.