Case 24

Assumptions Log for a Forecast

Scenario

You own a three-year demand forecast for a new service. The steering group keeps asking which assumptions matter most.

Why this framework

Uses PIN: what it's for, the context and source of truth, and a narrow ask. Enough for most everyday work.

Example context

Forecast model notes: Quarterly demand forecast for a regional meal-kit delivery service, next 8 quarters, base 41,000 weekly boxes. Growth assumed at 6% a quarter for the first four quarters, 3% after that. Churn assumed 4.5% a month, taken from last year's average. Marketing spend assumed flat at £180,000 a quarter. New postcode areas (three) assumed to open in quarter 3. Average basket assumed £38. Supplier price inflation assumed at 3% a year. Seasonality uplift of 8% in November and December, from two years of data. Business case: Sets a target of 55,000 weekly boxes by the end of year two to justify a second distribution hub costing £1.2m. Cites a customer survey of 600 people for growth intent. Does not state a source for churn or for the postcode launch dates. Owners are not named in either document.

Prompt

Purpose: Build an assumptions log for my demand forecast so the steering group can see what it rests on.

Context: Use the attached forecast model notes and the business case.

Narrow the ask: A table with each assumption, its source or "none", how sensitive the forecast is to it (high, medium, low with a reason), and who should own it. Highlight the three that would change the answer most.

Materials (pasted here in place of the attachments):
Forecast model notes: Quarterly demand forecast for a regional meal-kit delivery service, next 8 quarters, base 41,000 weekly boxes. Growth assumed at 6% a quarter for the first four quarters, 3% after that. Churn assumed 4.5% a month, taken from last year's average. Marketing spend assumed flat at £180,000 a quarter. New postcode areas (three) assumed to open in quarter 3. Average basket assumed £38. Supplier price inflation assumed at 3% a year. Seasonality uplift of 8% in November and December, from two years of data.
Business case: Sets a target of 55,000 weekly boxes by the end of year two to justify a second distribution hub costing £1.2m. Cites a customer survey of 600 people for growth intent. Does not state a source for churn or for the postcode launch dates. Owners are not named in either document.

Example output from ChatGPT

Demand Forecast - Assumptions Log

Regional meal-kit delivery service | 8-quarter forecast | Steering group review

Overall assessment: The forecast relies heavily on sustained growth, manageable customer churn and the opening of three new postcode areas. These assumptions need validation before the steering group can rely on the forecast to support the £1.2m distribution hub investment.

  1. Assumptions register

Sensitivity ratings are qualitative assessments, not results of model sensitivity testing. Proposed owners are recommendations because neither source names owners.

AssumptionSourceSensitivity and reasonProposed owner
1. Quarterly demand growth: 6% for Q1–Q4, 3% for Q5–Q8Model notes; business case cites survey of 600 for growth intent, but no quantified linkHigh - Top 3. Compounds across eight quarters; directly determines whether the 55k target is achieved.Commercial Director
2. Monthly customer churn: 4.5%Model notes: last year's averageHigh - Top 3. Small sustained changes in retention can materially affect the active customer base and replacement acquisition needs.Customer Retention Lead
3. Three new postcode areas open in Q3Model notes; no supporting launch dates in business caseHigh - Top 3. Delays or weak demand could reduce forecast volume and change distribution capacity requirements.Operations Director
4. Starting demand: 41,000 weekly boxesModel notes; underlying actuals not suppliedHigh. All future volume projections depend on this baseline being accurate and representative.Finance / BI Lead
5. Marketing spend remains £180,000 quarterlyModel notes; no supporting budget approvalHigh. Growth may require additional acquisition spend, especially if churn remains elevated.Marketing Director
6. Average basket value stays £38Model notes; no historical validation suppliedMedium. Directly affects revenue, although not box-volume demand.Commercial Finance Lead
7. Supplier price inflation averages 3% annuallyModel notes; no supplier contracts or price outlook suppliedMedium. Affects margins, pricing decisions and investment returns; indirect impact on demand.Procurement Director
8. November–December seasonal demand uplift: 8%Model notes: two years of historical dataMedium. Changes quarterly peaks and capacity needs, but is unlikely alone to determine the eight-quarter outcome.Demand Planning Lead
9. Customer growth intent translates into purchasesBusiness case: customer survey of 600 peopleHigh. Stated intent may not convert into recurring orders; survey methodology and conversion assumptions are unknown.Head of Customer Insights
  1. Three assumptions most likely to change the decision 1

Growth trajectory

The 6% and 3% quarterly growth rates are the central drivers of the volume forecast. Validate them against recent actual growth, acquisition capacity and the survey evidence.

2

Customer churn

The 4.5% monthly assumption comes from a historical average. Test whether it remains valid across customer cohorts, seasons and newly acquired customers.

3

New postcode launches

Q3 expansion is not supported by confirmed dates. Obtain launch readiness, expected weekly demand by area and ramp-up profiles.

  1. Initial target reconciliation

Target by end of year 2

55,000 Weekly boxes

Illustrative Q8 forecast

58,125 Weekly boxes, rounded

[Inference] Assuming quarterly growth compounds from the baseline without additional separate churn deductions:

41,000 ( 1.06 ) 4 (1.03) 4 ≈58,125

That is approximately 3,125 weekly boxes (5.7%) above the target.

Illustrative weekly box demand

Growth-only scenario versus the business-case target; churn treatment and expansion effects not independently modelled.

Projected weekly boxes Target 40,000 45,000 50,000 55,000 60,000 Base Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Base: Projected weekly boxes 41,000, Target 55,000 Q1: Projected weekly boxes 43,460, Target 55,000 Q2: Projected weekly boxes 46,068, Target 55,000 Q3: Projected weekly boxes 48,832, Target 55,000 Q4: Projected weekly boxes 51,762, Target 55,000 Q5: Projected weekly boxes 53,315, Target 55,000 Q6: Projected weekly boxes 54,914, Target 55,000 Q7: Projected weekly boxes 56,561, Target 55,000 Q8: Projected weekly boxes 58,125, Target 55,000

Important limitation: The model notes do not explain whether growth is net of churn or whether postcode expansion and seasonality are already included in the growth rates. Therefore, 58,125 is an illustrative calculation, not a verified model output.

  1. Steering group decisions needed

Confirm modelling definitions: Are growth rates net of churn, and do they already include postcode expansion and seasonality? This is essential to avoid double counting.

Commission sensitivity testing: Test lower growth, higher churn and delayed postcode launches, both individually and together.

Require evidence and named owners: Each assumption owner should provide supporting data, confidence level and a validation date before the hub investment decision.

Recommendation: Treat the 55,000-box threshold as a business-case hurdle, not proof that the £1.2m investment is justified. The steering group also needs a capacity and financial return analysis to establish whether the second hub is necessary and economically viable.