Where we start
Three sessions in, Kettle has stages that are evidence, a pipeline review that cannot be steered, and a forecast contract that says who owns the number and who owns its error rate. Tonight the BD lead takes all of that into a planning meeting. The finance partner wants next year's target met by the fourth quarter; the founder wants to know whether that means two new reps, one, or none yet. The answer is three pieces of arithmetic and one gate. The arithmetic is easy and is constructed for Kettle, shown line by line so you can check it. The gate is harder, because it asks for numbers most teams have never measured: their own win rate by cohort, their own cycle time, and their own forecast error over past quarters.
★ What you walk out with tonight. A one-page hiring case for your own team: the target, what one ramped rep carries, the ramp that delays it, the readiness gate with the evidence you have and the evidence you are missing, and a decision with a written trigger for when to revisit it. It is the page you hand the finance partner instead of a headcount number with no working.
What one rep can carry turns a target into a headcount
5 min liveA target is a sum of money. A rep is a person with a limited number of deals they can take to a decision in a quarter. The bridge between the two is capacity per ramped rep: deals carried to a decision, times the share that win inside the quarter, times the median signed size. Divide the target by it and you have a headcount. Every input is one of the four numbers from Business session 1, read for one person instead of the whole team.
LiveThree numbers per rep, one division3 min▶
Deals to a decision. How many stage-2-or-later deals one rep can genuinely work to a yes or a no in a quarter: trials to set up, proposals to write, procurement to chase. Kettle's BD lead counted it from the last two quarters of the activity log: about 24 per rep (constructed). Stage-1 contacts do not count; Business session 1 moved them to the lead list for a reason.
In-quarter win share. Not the eventual win rate: the share of those deals that close inside the same quarter. Kettle's own stage table, from the quarter before, has 975 deals at stages 2 to 5 and 96 of them won in-quarter, 9.8 percent. The plan rounds it to 10 percent and says so.
Median signed size. 480,000 yen, the same constructed figure as the velocity example in Business session 1. Multiply: 24 x 0.10 x 480,000 = 1,152,000 yen per ramped rep per quarter. A 6.9m yen quarterly target divided by 1.152m is 5.99, so six ramped reps. Kettle has four.
Self-studyWhy cycle time caps what a rep can carry2 min▶
"Deals to a decision in a quarter" hides a time limit. On the real Olist rows, the median won deal closed 14 days after first contact and the 75th percentile 54.75 days. Kettle's quarter is 91 days. A deal a rep opens in the last eight weeks of the quarter has a fair chance of landing in the next one, so a rep's capacity is a rolling number, not a fresh bucket every quarter. Use the 75th percentile, not the median, when you ask "how late in the quarter can a new deal still count?", because the plan has to survive the slower half of deals.
Capacity is also not quota. Quota setting and compensation design are separate decisions with their own literature and their own politics; capacity is the input they start from, measured before anybody negotiates.
The sanity check that saved a bad plan. Kettle's first draft used Aiko's best quarter as everyone's capacity and concluded that four reps could already hit 6.9m. The analyst checked the team figure against the bench: four reps x 1.152m is 4.608m, and the replayed quarters in Analyst session 5 report "Average quarter: 4.57m yen really closed." The two agree within about 1 percent; the best-quarter version did not agree with anything. Constructed Kettle & Co example; the 4.57m is the bench's own printed line.
Hired is not productive
5 min liveA new rep spends the first quarters learning the product, building a pipeline from nothing and waiting for the first deals to cycle through. So they close a fraction of a full rep's quarter, rising over time. That curve is the ramp, and it means two lines on the plan: headcount, which jumps on the start date, and productive capacity, which climbs slowly behind it. The space between them is money the plan was counting on and will not get.
LiveTwo lines, not one3 min▶
The ramp, constructed for Kettle. A new rep closes 25 percent of a full rep's quarter in their first quarter, 50 in the second, 75 in the third and the full 1.152m from the fourth. Hire two reps at the start of Q1 and the new capacity is 2 x 0.25 x 1.152m = 0.576m in Q1, then 1.152m, 1.728m and 2.304m.
Team capacity against the 6.9m target. Add the four existing reps' 4.608m: Q1 5.184m (short 1.716m), Q2 5.760m (short 1.140m), Q3 6.336m (short 0.564m), Q4 6.912m (on target). Headcount hit six on day one; capacity hit six reps' worth in Q4.
The year in one subtraction. Counted as if both hires were productive from day one, the year holds 4 x 6.912m = 27.648m. With the ramp it holds 5.184 + 5.760 + 6.336 + 6.912 = 24.192m. The difference, 3.456m, is the cost of pretending headcount is capacity. A plan that wants Q1 on target has to hire earlier, not hire more.
Self-studyWhere your ramp numbers should come from2 min▶
Kettle's 25, 50, 75, 100 is a teaching shape, not a benchmark. Your own ramp comes from your own past hires: each rep's closed amount by quarter since their start date, divided by the full-rep figure. With three or four past hires the curve will be noisy, and that is the honest answer: write the ramp as a range ("first quarter 15 to 35 percent") and carry the range into the plan.
Ramp also has a mechanical floor that no onboarding programme removes: a new rep starts with an empty pipeline, and deals take time to close. With a real median of 14 days and a 75th percentile of 54.75 days from first contact to won, a rep who opens their first deals in week three will see a slow-tail share of them land only in the second quarter. If your own cycle is longer, your ramp is longer, whatever the training plan says.
The Q1 that was promised twice. In the version of the plan sent to finance first, Kettle's two hires started on 1 April and the Q1 forecast went up by two full lanes, 2.304m, the same day. The analyst's redraw added 0.576m for Q1 instead. Nobody had lied; the spreadsheet had a headcount column and no ramp column. Constructed Kettle & Co example using the constructed ramp.
Scale on your own evidence
5 min liveEvery number in the arithmetic above is an assumption about the future borrowed from the past. If the win share wobbles, the cycle is unknown, or nobody knows how wrong last quarter's forecast was, the headcount is a guess with decimal places. Mark Roberge, in The Sales Acceleration Formula (2015), is reported to argue for exactly this order: scale hiring only once your own numbers show you are ready, not because a benchmark or a board says the market is big. The gate below turns that stance into three numbers you can check.
LiveThe three numbers that open the gate3 min▶
1 · A win rate that holds still, by cohort. The capacity line multiplies by a win share. If that share moves by a third from one entry month to the next, so does every rep's capacity. On the real Olist rows standing in for Kettle's history, the January to April 2018 cohorts converted 13.3, 14.5, 14.2 and 13.5 percent; May converted 10.0. Analyst session 2 shows why it has to be read by cohort, not by period.
2 · A cycle time you know, with its tail. The ramp and the quarter boundary both depend on it. Real median 14 days, 75th percentile 54.75 days (Analyst session 3). Without the 75th percentile you cannot say how much of a new rep's first pipeline lands in their second quarter.
3 · A forecast error you have measured over past quarters. The plan is a forecast with names on it. If you do not know how wrong your forecasts usually are, you do not know how wrong the plan is. The bench in Analyst session 5 reports "Smallest error: history-based, 38.8%." over 20 replayed quarters: that is the kind of number to have before a hiring decision, even though it is not a comfortable one. Applied to a 6.9m target, a typical miss of that size is about 2.7m yen (6.9 x 0.388 = 2.68, constructed arithmetic).
Self-studyWhat Roberge's stance asks for, and what it does not2 min▶
As reported, Roberge's argument in The Sales Acceleration Formula is about order: prove the selling motion works on your own numbers, then scale the team. It does not hand you thresholds. No number on this page is his; the three gates and every threshold in the build-along are constructed so you can practise the reasoning, and your own thresholds should come from your own history and your own appetite for risk.
The same gate appears from the other side in Market Sizing, Business session 4 ↗, where a market size is turned into a plan and the question is whether the size estimate is good enough to hire against. That page owns the sizing; this one starts where the pipeline is. Read together, a big market and a ready pipeline are two separate conditions, and a plan needs both.
The month that broke the band. On the real Olist rows, four consecutive entry cohorts in early 2018 converted between 13.3 and 14.5 percent, and the May cohort converted 10.0 percent. A plan written in May on the January to April rate would have overestimated every rep's capacity by a third or more. Part of May's drop may be shorter follow-up, since the closed-deal log ends in November 2018 and the slowest deals take months; either way, a band that just broke is not a stable win rate yet. The same drop is why the history-based forecast in Analyst session 5 runs high on average.
A one-page hiring case
★ 22 min · everyone works it throughNo code, one page, five rows. Work it for your own team if you can, or for Kettle using the constructed numbers above. Step 1 together, steps 2 to 4 in pairs, then each pair reads its step 5 decision and trigger aloud. A case whose gate row says "missing" is a good case: it tells finance exactly what to measure before the next hire.
Step 1 · Write the target as a quarterly number. Annual targets hide the ramp. Split the year into quarters and say which quarter has to be on target. Kettle's finance partner wants 6.9m yen a quarter by Q4.
Step 2 · Capacity per ramped rep. Three numbers, one product, then one division. Use stage-2-or-later deals, the in-quarter win share from your own stage table, and the median signed size. Round the headcount up, then check it by multiplying back.
Step 3 · Put in the ramp. Write the new rep's share of a full quarter for their first four quarters, from your own past hires if you have them, as a range if you do not. Then write the team's capacity quarter by quarter, and the gap to target.
Step 4 · Run the readiness gate. For each of the three numbers, write the value you have and where it came from, or write "missing". Do not fill a gap with an industry benchmark: the gate is about your own evidence.
Step 5 · Decide, and write the trigger. State what you will do now, what that means for the target, and the observable event that reopens the case. A trigger names a number and a date, so nobody has to remember to feel ready. Read yours aloud.
Attach the forecast contract. The hiring case is a forecast with salaries attached, so it inherits the contract from Business session 3: the BD lead owns the decision and the target, data owns the capacity method, the range and the error rate. Put both names on the page.
Before your next planning meeting
30 min- Compute your capacity per ramped rep from the last two quarters, then multiply by your current headcount and compare it with what the team actually closed. If the two differ by more than a fifth, find out which of the three inputs is wrong before using it in a plan.
- Draw your own ramp. For every rep hired in the last two years, list closed amount by quarter since their start date as a share of a full rep's quarter. Write the range you see, not the best case.
- Fill the gate honestly. For each of the three numbers, write the value or "missing". For any missing number, name who would measure it and how long it would take; the Analyst track builds all three.
- Write one trigger for your next hire: a number, a threshold and a date to check it. Send it to the person who signs off headcount before the next planning meeting, not during it.
Three questions
Clear all three and Business session 4 gets stamped in your passport.
1 · Kettle hires two reps at the start of Q1, and headcount goes from four to six. Why does Q1 capacity only reach 5.184m yen, not 6.912m?
With the constructed ramp of 25 percent in the first quarter, two hires add 2 x 0.25 x 1.152m = 0.576m in Q1, not two full lanes. Hiring earlier closes that gap; hiring more does not.
2 · Mark Roberge's stance in The Sales Acceleration Formula (2015), as reported, asks a team to do what before scaling hiring?
The stance is about order: evidence from your own team first, then scale. It supplies no thresholds; the ones in the build-along are constructed for practice.
3 · Which three numbers form the readiness gate?
Each one feeds the hiring arithmetic: the win share sets capacity, the cycle sets the ramp and the quarter boundary, and the error says how far the whole plan can be off.
What this session teaches, and where it came from
Every fact on this page, and its verification tier, is recorded in the course's source map.
Business session 4 cheat sheet · pin this
That completes the Business track: stages as evidence, a review you cannot be steered in, a forecast contract, and a hiring case built on your own numbers. The Analyst track builds every number the gate asks for from the rows. It starts with Analyst session 1, the pipeline data model: why the CRM's current state forgets last month, and how to rebuild it from the event log. Bring your analyst, or become one for six sessions.