Where we start
Analyst session 4 turned Kettle & Co's market into a range: premium loose-leaf tea bought online in Japan, 19.63 billion yen a year at the top-down middle. Its year-5 plan says Kettle & Co will hold 10 percent of that. Three rivals have plans too: Entrant B 18 percent, Entrant C 22 percent, Entrant D 15 percent, while the incumbents hold 55 percent. Every share, every rival and every plan is constructed for teaching; the households and tea spend under the 19.63 are real rows from the Statistics Bureau of Japan. Tonight the question is not how big, but how much of it, and whether anyone has checked the plans against each other.
★ What you walk out with tonight. A collision line for Kettle & Co's sizing brief: what the five plans claim when added up, how much of that has no customer, and a rewritten Kettle & Co plan that says which rivals would have to lose share for it to hold. Plus two quick realism checks you can run on any share in any deck: the curve and the ceiling.
Adoption rarely arrives in a straight line
5 min liveMost plan ramps are drawn with a ruler: zero today, the year-5 share at year 5, and an even climb between. Adoption of a new product rarely behaves like that. It tends to start slow, accelerate once early buyers bring others with them, then flatten as the market fills. Held against that shape, a straight ramp books share in the early years before the customers have arrived.
LiveThe S-curve as a shape, and only as a shape3 min▶
Where the shape comes from. Frank Bass's diffusion model (Management Science, 1969, reported) describes adoption as two forces: a few innovators who buy on their own, and imitators who buy because others already have. Early on there are few adopters to imitate, so growth is slow; once enough people own the product, imitation drives acceleration; as the pool of possible buyers empties, growth flattens.
What is used here. Only the shape: slow, fast, flat. No parameters are attributed to the paper, and the curve in the figure is drawn by hand for Kettle & Co. If you ever see an S-curve on a plan slide with precise coefficients, ask where they were estimated, on whose data, and for which product.
The check. Lay the plan ramp over the shape. If the plan reaches a large share of its year-5 number in year 1 or 2, it is assuming the acceleration phase has already started. Ask what evidence says so.
LiveA realism check, not a forecast2 min▶
Forecasting how adoption will actually unfold, fitting a curve to early sales and projecting it, is a different job with its own methods and failure modes. It is owned by Learn Timeseries Forecasting with Phoebe, and nothing here re-teaches it.
In a sizing brief, the curve does one thing: it asks whether the year-5 share and the ramp to it are believable given the shape adoption usually takes. Its output is a question for the plan's owner ("why is year 2 already so close to the year-5 number?"), never a number of its own.
Self-studyWhen the shape does not fit2 min▶
Not everything follows an S. A product that enters an established category, where buyers already exist and only the brand is new, may skip most of the slow start: the question there is switching, not adoption. A product sold through a partner with an existing customer base can step up on the day the partner switches it on.
So the S-curve is a prompt, not a rule. If the plan's owner can name the mechanism that skips the slow start (a distribution deal, an existing customer list, a category that buyers already know), the straight ramp may be fair. If they cannot, the early-year gap is a risk that belongs in the brief.
The ramp drawn with a ruler. Kettle & Co's marketing lead presents a plan that climbs evenly to 10 percent of the online premium market by year 5, so year 1 already books a fifth of the year-5 share. The head of data overlays a hand-drawn S and asks what will make Japanese buyers adopt in year 1 at the rate the ramp needs. Nobody can name it. The revised plan keeps the year-5 target and moves the hiring plan back a year. Constructed Kettle & Co example.
Check a share against who got there before, and where you are sold
5 min liveTwo more checks, both cheap. First, the base rate: in a crowded category most entrants end small, so a plan should be compared with the shares that comparable entrants actually reached, not with the share the team would like. Second, the ceiling: a brand cannot win purchases made where it is not sold. For a direct-to-consumer brand, that ceiling was already in the chain from Analyst session 1.
LiveBase rates: who got there before3 min▶
The principle. In a crowded category, most entrants end small. A few win big, and those are the ones everyone remembers and puts on the slide. The honest comparison for a new plan is the shares that comparable entrants actually reached at the same age, including the ones that stalled.
What is not on this page. No base-rate study was read at source for this session, so no percentage for "how many entrants reach how much share" is quoted. Numbers like that circulate widely; treat any you meet as a claim to trace before it goes into a brief.
A constructed worked example. Kettle & Co's analyst lists five tea brands that entered Japanese e-commerce in the past decade, comparable in price tier and channel, and notes the share each reached by year 5 from public filings and panel data. The list, the brands and their shares are constructed for teaching. Four of the five sit well below Kettle & Co's 10 percent plan; one sits above it. The brief does not say the plan is wrong. It says the plan needs Kettle & Co to be the one in five, and asks what makes it so.
LiveDistribution ceilings: only where you are sold2 min▶
Analyst session 1 built the top-down chain with an online share link: of all premium loose-leaf spend, the part bought online, constructed at 0.30 in the middle. For a brand that sells only through its own site and online marketplaces, that link is not just a filter. It is the ceiling. Purchases made in supermarkets, department-store gift counters and specialist tea shops are real spend, but a direct-to-consumer brand has no route to them.
Watch for plans that quietly swap denominators: a share quoted "of the Japanese premium tea market" when the brand is sold only online. The share might be correct for the online slice and impossible for the whole. Ask, for every share on a slide: share of what, sold where?
Self-studyCeilings move, and that is a plan of its own2 min▶
A ceiling is not permanent. A retail partnership, a marketplace listing in a new channel, a corporate gifting arrangement: each raises it. But raising a ceiling is a separate decision with its own cost and timing, and it belongs in the plan as a named step with a date, not as an unstated assumption inside the share.
The test: if the share in the plan only fits under a higher ceiling, the plan must say which channel raises it, when, and who owns that deal. If it cannot, rewrite the share under today's ceiling.
The gift counter nobody had signed. Kettle & Co's BD lead quotes a year-5 target as a share of all premium leaf tea in Japan, including the department-store gift season. The head of data points out that Kettle & Co sells online only and has no department-store agreement. The target is rewritten as a share of the online slice, and the gift counters become a separate line in the plan, owned by BD, with "no agreement yet" written beside it. Constructed Kettle & Co example.
Every plan is reasonable; the sum is not
5 min liveThe checks so far look at one plan at a time. The most revealing check looks at all of them at once. When a market is big and exciting, every entrant writes a plan for a slice of it, and each slice looks modest on its own slide. Add the slices up and they can claim more than the market holds. Aswath Damodaran ran exactly this addition on real company valuations.
LiveThe 2015 addition3 min▶
The method. In a 2015 post, Damodaran took each listed online advertising company, started from its market value, and solved for the revenue in 2025 that value implied: how big would this company have to be for today's price to make sense? Then he added those implied revenues up across the group.
The result. The imputed 2025 online ad revenues came to about $523 billion. His own estimate of the 2025 online advertising market, which he described as optimistic, was about $466 billion, built up from about $545 billion of global advertising in 2014, of which about $138 billion was digital. The companies, taken together, were priced for more revenue than his generous market could supply, so the group was collectively overpriced.
The lens. He framed it as individual rationality, collective irrationality: every investor's case for one company can be sensible, and the sum of the cases still impossible. That is the collision test, run on money rather than on plan slides.
LiveWhy big markets invite it: the 2019 revisit2 min▶
In a 2019 post, Damodaran came back to the idea and named three ingredients: a big market story, overconfidence among the companies and investors chasing it, and companies that are priced on the story rather than valued on their cash flows. In his words, "young companies tend to be priced, not valued".
He does not say every company in such a market fails. Some will succeed, and some will be worth more than their price. The point is about the group: added up, it is overpriced, and the correction tends to come for the whole group at once.
Read the same three ingredients into a planning room. A big market slide, a team confident in its share, and a plan built to match the story rather than tested against the market: that is how five rival plans come to claim more than the market holds.
Self-studyThe paper behind the posts1 min▶
Brad Cornell and Aswath Damodaran also wrote the idea up as a paper, "The Big Market Delusion: Valuation and Investment Implications" (reported: the paper was not read at source for this session, so nothing here is quoted from it). The two blog posts, read at source, carry everything taught above. Cite them as: Damodaran, Musings on Markets, 2015 and 2019 posts.
Online advertising, 2015. Each listed online advertising company had a story that justified its price on its own terms. Damodaran's addition put those stories side by side: about $523 billion of implied 2025 revenue against about $466 billion of market, even on his optimistic estimate. No single valuation looked absurd; the sum did. Source: Damodaran, Musings on Markets, 2015 and 2019 posts.
Run the collision test on five plans
★ 22 min · everyone buildsThe widget below runs Damodaran's addition on Kettle & Co's market. It takes the premium loose-leaf online market at its top-down middle, computed in your browser from the real Census households and survey tea spend with the constructed shares, then applies five constructed year-5 plans to it and adds them up. Work the five steps and write down what it prints.
Step 1 · Read each constructed plan. The table lists five players and the year-5 share each plan claims of the same market. Read them one at a time, the way each board would see its own plan, and note that none of them looks outrageous alone.
Step 2 · Read the total. The verdict line adds the shares and the revenues. Copy it exactly; it is the one sentence no single plan could have produced.
Step 3 · Read the revenue per player. Each share times the 19.63 market gives the revenue the plan is counting on. Write them down; Kettle & Co's line is the smallest, and it is still part of the overflow.
Step 4 · Rewrite Kettle & Co's plan so it survives the collision. The market will not grow to fit the plans, so for Kettle & Co to keep its share, someone else must hold less. Decide who, and say why that player would lose share: a weaker channel, a narrower range, an incumbent losing online buyers. Every number you write here is constructed, and your numbers will differ from your neighbour's.
Step 5 · Write the collision line for the brief. One sentence that a leader can read without the table: what the plans add up to, how much has no customer, and what Kettle & Co's plan now assumes about its rivals.
Rivals' plans are rarely published. In real work you will not have Entrant C's board deck. Use what you can see: funding announcements, hiring, stated targets in interviews, and the market values investors have assigned, as Damodaran did. A collision test on estimated rival plans, labelled as estimates, still catches a total over 100 percent. Keep the line; Analyst session 6 puts it beside the triangulated range on the bench.
Before Analyst session 6
30 min- Overlay an S on one plan ramp you own or have seen. Sketch it by hand, same year-5 end point, and mark the early-year gap. Write one line on what would let the plan skip the slow start, or admit that nothing does.
- Write the share-of-what line for every share in that plan: share of which market, sold through which channels. Circle any share whose denominator includes purchases made where the product is not sold.
- List three to five comparable entrants and, where you can find it from public sources, the share each reached. Note where each number came from. If you find a published base rate, write down its source and method before you trust it.
- Run a collision test on your own market. List every competitor you can name with an estimated target share, label each estimate, add them up, and write the collision line in the step 5 format.
Three questions
Clear all three and Analyst session 5 gets stamped in your passport.
1 · Each of the five constructed plans looks reasonable on its own. What does the collision test add that no single plan shows?
The overflow belongs to the sum, not to one plan. The test says the plans cannot all hold; deciding whose share gives way is the analyst's next question.
2 · What did Damodaran's 2015 addition of online advertising companies show?
Individual rationality, collective irrationality: any one company may deserve its price; the group, added up, is priced for more than the market can supply.
3 · Why is the S-curve used here as a realism check rather than a forecast?
The shape asks why the plan books share before adoption usually arrives. Fitting and projecting a curve is forecasting, which has its own methods and its own home.
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.
Analyst session 5 cheat sheet · pin this
You can test a share against the adoption shape, the distribution ceiling and every rival's plan at once. Analyst session 6 is the triangulation bench: the full ladder from one method to two, the category and population toggles, the hero assumption, and the analyst-report number that leaves nothing to explain.