How Forum turns a $40M gamble
into a prepared decision.

A $40-million bet that can't be unwound.

Cadence makes premium connected-fitness hardware — $2,000+ machines plus a recurring subscription. Every season they tell their contract manufacturer how much to build. Once the production cells are reserved, the commitment is irreversible — and the practical point of no return is July 1.

Sales wants to commit maximum capacity so the company never sells out. Operations owns the inventory that commitment creates. Leadership has one cross-functional meeting to decide how much to build — the kind of call that, made wrong, shows up as a writedown two quarters later.

“We're selling everything we make.”

The story in the building was growth. Sales was forecasting 20% quarter-over-quarter, backed by a waitlist, and pushing to commit the full plan.

“20% quarter-over-quarter growth, and it's built on order intake plus a pretty substantial waitlist.”

It's a confident, reasonable-sounding case. A single AI assistant — or an exec in a hurry — would take it at face value and commit. Forum didn't. It sent an analyst to interview each team and pressed on the specifics.

The confident story started to soften.

Talking to Sales' own people, the growth number got shakier the more it was pressed:

“10 to 13% over the last couple of quarters is orders we pulled forward — channel incentives, early-renewal offers.”

“Conversion has slipped from around 70% to 55% the last couple quarters.”

So the growth was thinner than it looked. Part of it was borrowed from future quarters, and fewer shoppers were actually closing. The waitlist looked big; the demand under it was softening.

Meanwhile, the Operations analyst was hearing a very different story about the same business:

“About 9 weeks of finished goods right now, up from around 4 a quarter ago, drifting higher for over a month.”

“Roughly forty million in working capital tied up — about two quarters of unsold finished goods — if we commit full and demand normalizes.”

And inventory was quietly piling up. It had more than doubled and was still climbing — commit max capacity into softening demand and ~$40M freezes in unsold machines.

And Marketing had already spent against the optimistic plan — with one risk nobody had taken on:

“We still don't have a real read on what happens to resale pricing if we end up overbuilding. That's a gap I keep flagging and nobody's really owned yet.”

Read the full evidence — Operations →

Same company. Opposite readings. Never reconciled.

Both readings were true. Both came from the same business. And no one had put them side by side — Sales doesn't carry the inventory P&L, Operations doesn't own the forecast, and the two numbers lived in different rooms. Once Forum surfaced it, even Sales' own VP saw the tension:

“Operations is showing supply up around 9 weeks from 4 a quarter ago, which doesn't square with 'selling everything we make'… real tension that needs reconciling.”

This is the thing Forum exists to catch — the disagreement that decides the outcome but never reaches the room, because no single person can see both sides at once.

From “commit max” to a staged bet with an exit.

Forum brought the Sales and Operations agents together to work the tension through directly. They didn't split the difference — they reasoned to a structure that protects the upside without betting the balance sheet on a forecast that's softening:

Just as important, the agents escalated what they couldn't verify instead of inventing it — three items went to the humans: the manufacturer-stamped confirmation date, Finance's sign-off on the ~$1–2M option premium, and the exact trigger thresholds (proposals, not record-backed).

Would 70/30 alone have saved them? Not necessarily. A glut is already forming, and even a staged commit can be too much if demand keeps sliding. That's the point — Forum doesn't hand you a magic number. It hands you the structure, the trade-off, and the open triggers, so leadership picks 70/30, or 50/50, or pause, with both signals finally on the table.

See the agents work it through →

One page, instead of a week of back-and-forth.

Forum's Coordinator pulled everything into a single pre-meeting brief: what the teams agree on, where they disagree (both sides plus the crux), the open gaps, and the specific decisions to make — each one traceable to a real interview.

• Aligned July 1 is the real point of no return; the 70/30 structure works for both teams.
• Tension ~$30M stockout exposure (Sales) vs ~$40M overproduction tie-up (Ops) — reconciled, not averaged.
• Gap Resale-price erosion if Cadence overbuilds — flagged by Marketing, owned by no one.
• Decide Ratify 70/30 · stamp the Q+2 date · authorize the premium · validate the triggers.

The meeting starts at the decision — not at “wait, what does Sales actually mean by growth?” The full report is embedded below — click into any part of it.

Forum didn't know how the story ends.

A fair question: did we reverse-engineer a demo because we knew the famous ending? No — and the way the run works is why:

• The people are grounded in the period's public record — what was knowable at the time: the bullish sales talk, the pulled-forward orders, the inventory climbing, the port backlog. None of them hold the outcome.
• The analysts answer only from those records. Every figure traces to one (that's the Evidence screen). They can't leak the ending because the ending isn't in their data.
• The contradiction was already in the inputs. One team's signal said up, another's said down. Put independent, grounded findings side by side and the tension appears on its own — no one scripted it.
• Forum isn't predicting the crash. It's surfacing an unreconciled contradiction before the decision. Even if demand had held, the brief still reads “these two signals disagree — reconcile before you commit.” The value is putting it on the table, not calling the winner.

And it generalizes: any decision where teams hold contradictory, grounded signals, Forum surfaces it. Cadence just happens to have a famous ending.

See every figure traced to its record →

One misstep of several — and the kind Forum is for.

Cadence is fictional — but its situation is taken straight from Peloton. In 2021, Peloton kept committing manufacturing capacity on a permanent-growth story while warehouses filled and ports backed up. The two signals — bullish demand and swelling inventory — were never reconciled before the spend locked. When demand normalized: a glut, deep writedowns, and roughly $12.5B in market value erased.Public record: City of Hialeah Emps.' Ret. Sys. v. Peloton Interactive (2d Cir. No. 24-2803, 2025) — confidential-witness accounts of ballooning inventory and ports backing up; incoming CEO Barry McCarthy called it “drowning in inventory.”

To be clear, the capacity overcommit wasn't the only thing that hurt Peloton — the Precor acquisition, an over-built supply chain, and a broad misread of post-lockdown demand all weighed on it. This was one of several compounding calls, and Forum wouldn't have “fixed” the company.

What Forum does is narrower, and more honest: it surfaces this class of failure — the cross-functional contradiction that decides an outcome but never reaches the room — so leadership makes the call with both signals on the table. It won't pick the right number for you. It makes sure the disagreement is seen before the decision, not in the post-mortem — the difference between a judgment and a guess.