Prioritisation · Frameworks, worked

What the delay costs

Put a dollar figure on a six-week launch slip from three stated inputs, and say which of the three the figure actually depends on.

estimateeasyExercise · 3 min · 16 of 59

On this page

The drafting add-on was due on 1 March and will now launch in mid-April. Somebody has to say what the six weeks cost, and the cost of delay lesson says the answer depends on which urgency profile each input has.

The situation

The drafting add-on for the support tool was due to launch on 1 March; the eval missed its pass threshold and the launch is now 12 April, six weeks later. Three inputs, all assumed, in USD. First, revenue: the add-on is priced at $5 a seat-month across 12,000 seats, $60,000 a month at full roll-out, and the launch plan bills about 600 seats (5%) in week one, 1,800 (15%) in week two, and all 12,000 from week three. Second, cost: the token bill at full usage is $54,000 a month, so the add-on's contribution at full roll-out is $6,000 a month. Third, one tenant: 1,200 agents, a $40,000-a-month contract renewing on 1 April, which sales closed last year on the promise that drafting would be live by then. The account manager puts the chance of losing the renewal at one in four if the feature is not live on the date. A prototype has been running in that tenant's sandbox since January.

The question

What does the six-week slip cost, to the nearest ten thousand dollars, and which of the three inputs does the figure depend on?

A strong answer · Rubric · Where answers go wrong

Write your answer first, then open

A strong answer

Revenue is standard profile: six weeks of add-on revenue shifts right. At full roll-out that is 1.5 months × $60,000 = $90,000 of revenue deferred. The ramp moves with the launch, so it changes when the gap opens, not how large it ends: once the late launch reaches full roll-out it is still six weeks of full-rate billing behind. Deferred, not lost: the seats are still billed, six weeks late.

Cost follows revenue, because the token bill is not paid until drafts are produced. On contribution, the deferral is 1.5 × $6,000 = $9,000. That is the whole cost of the slip if nothing has a date.

Something has a date. The renewal on 1 April is a fixed-date profile: the slip crosses it, and the exposure is the contract, not the add-on — $40,000 × 12 = $480,000 a year, at the account manager's one in four, an expected $120,000.

Total: about $130,000 ($120,000 + $9,000), and the renewal is 93% of it.

The decision this buys: spend the six weeks on that tenant, not on the launch. Offer the sandbox prototype as a dated pilot before 1 April with the FDE on site, so the renewal conversation happens with the feature in use. If the one in four falls to one in twenty with a pilot running, the expected cost falls to $24,000 and the slip costs $24,000 + $9,000, about $30,000 in all. What would change the estimate: a contract clause that makes the renewal automatic, which removes the cliff and leaves the $9,000 of deferred contribution.

Rubric

Separates deferred revenue from lost contribution and shows why the token bill nets out.
Shows that the ramp moves with the launch, so the deferral is six weeks at the full rate, not at the ramp's average.
Identifies the renewal as a fixed-date profile and computes the exposure as contract value × probability.
States the total to the nearest ten thousand dollars and says which input it depends on.
Names the action that changes the dominant input, and what the estimate becomes after it.

Where answers go wrong

"$90,000 — six weeks of $60,000 a month." Revenue is deferred, not lost, and the token bill goes with it; the contribution at risk is $9,000. The answer has costed the slip at ten times its standard-profile value and missed the one input that is actually large.
"$9,000 — six weeks of a $6,000-a-month margin." Right on the add-on, and it stops there. The renewal is on a date the slip crosses, and that single input is thirteen times the rest; the answer never asked whether anything had a date.
"$480,000 — we lose the tenant." It takes the cliff at face value and ignores the one in four, the pilot that could move it, and the difference between a contract at risk and a contract gone. An estimate without the probability is a threat, not a number.