From size to plan

It is the furthest from anything measured. Year 1 can lean on the pilot: real visitors, real first orders. Year 5 leans on four years of assumptions stacked on top of each other: how fast new customers arrive, how many come back, what rivals do, whether the market itself grows or shrinks. Each layer adds error, and none of the layers has been observed yet.

It is the number the whole plan is reverse-engineered from. In most plans the year-5 share is written first ("we want 10 percent") and the ramp is drawn backwards to reach it. That makes it the most wished-for cell, not the most evidenced one.

It is also the number the budget is quietly built on. Hiring, media and stock are often sized for the destination. That is why the gate in Part 3 sits between year 1 and year 2: the budget follows the measured years, not the hoped-for one.

Frank Bass's 1969 paper in Management Science (reported here, not read at source) gave planners a familiar shape for how a new product spreads: a slow start among the few who try new things first, an acceleration as others copy them, then a flattening as the people left to convert get scarce. Kettle's ramp borrows only that shape: small steps early, bigger steps in the middle years.

The shape is a realism check, not a forecast. No parameters from the paper are used here, and any growth rates you see on a ramp are planning assumptions. Analyst session 5 uses the same curve to ask whether a plan's middle years climb faster than adoption plausibly allows. Fitting curves to real sales and forecasting them belongs to Timeseries Forecasting ↗.

The figure keeps the ceiling at 19.63 for all five years so the ramp is the only thing moving. A real plan would ask whether premium loose-leaf bought online is itself growing in Japan. That is a forecasting question, and it is a separate line on the plan with its own range, never folded silently into the share.

One question: of the companies that entered a market like this one, what share did they actually reach, and how long did it take? Not the winners you remember, but the comparable set: similar product, similar channel, similar size at launch, including the ones that closed.

Then compare. If Kettle's 10 percent sits inside what comparable entrants reached, the plan is ordinary, which is a compliment. If it sits above, the plan has to say why Kettle is different: a channel nobody else has, a cost advantage, a customer base it already owns. "We are better" is not a reason; every entrant believed that.

Where the evidence comes from. Company filings, trade association reports, retailer rankings, the data team's own scan of rival storefronts. No single published base rate covers every category, and the honest answer for a niche one is often "we found five comparables, not fifty". Write the count next to the band; a band built on five companies is a weaker check than one built on fifty, and the board should know which it is looking at.

A base rate checks one plan against history. The collision test checks every plan in the market against each other. In the constructed Kettle market, Kettle plans 10 percent of SAM, Entrant B 18, Entrant C 22, Entrant D 15, and the incumbents expect to hold 55. Add them up and the plans claim 120 percent of a 19.63 billion yen market: 23.55 billion yen of planned revenue, 3.93 billion of it with no customer to buy it.

No single plan looks absurd on its own. Together they cannot all come true, and someone's year-5 cell is going to miss. Analyst session 5 builds the test live in the browser. The plans are constructed; the arithmetic is the engine's.

Mark Roberge, who wrote The Sales Acceleration Formula (2015) and has written since about scaling go-to-market teams, is reported to argue a simple order of operations: do not pour money into sales and marketing because the market is big; pour it in when your own numbers show the engine works. Paraphrased here from secondary accounts, not quoted, and no figures are his.

The four numbers that usually stand for "the engine works": win or conversion rate (do the people you reach actually buy), cycle (how long from first contact to first purchase), payback (how long until the margin on a customer repays what it cost to win them), and repeat or retention (do they come back). A market size touches none of them. That is exactly why it cannot be the reason to scale.

For a direct-to-consumer brand like Kettle, "win rate" becomes conversion and "cycle" becomes days to a first order. For Kettle's office tea idea, run through a BD team, they would be the classic sales versions: deals won out of deals pitched, weeks from first meeting to signature.

Before the pilot starts, not after it ends. A bar written after the results are in tends to land just below them. Kettle's bars are agreed at the board meeting that approves the pilot.

The bet owner proposes, the data team checks it can be measured. The same split as the brief in Business session 3: the BD or marketing lead says what level of conversion would justify scaling; the head of data says whether the pilot will produce enough orders to tell. A bar of 3 percent conversion means little if the pilot only reaches a few hundred visitors.

What the gate protects against. Spending as if the market share were already won. The commonest failure in a plan is not a wrong size; it is a right size spent against too early.