Unconstrained demand forecast: uncover demand beyond sales

An unconstrained demand forecast estimates what customers would request if supply limits did not restrict purchasing. Sales recorded during a stockout show fulfilled demand, while missing requests remain partly unobserved. Label any estimated adjustment as an assumption rather than presenting it as another sales transaction.

What the forecast measures

An unconstrained forecast answers how much demand may exist at a stated price, offer, location and period with sufficient availability. A constrained plan answers how much the business expects to fulfill with the supply it has. The gap helps purchasing and operations assess whether more availability could serve additional customers.

Past demand reconstruction is an input to that forecast, not the forecast itself. Estimating what customers might have bought last month does not establish next month's seasonality or promotion response. Use the corrected history alongside your separate baseline demand forecast.

Keep three categories visible: observed fulfilled units, recorded unfilled requests and estimated requests that were never captured. An abandoned shopping trip belongs to the last category unless there is evidence of the request. A backorder already included in the order register does not need another lost-demand adjustment.

Gather sales and availability on the same basis

Start with a daily item-and-location export. Include sales units, transaction date, usable stock status, selling hours, promotion status and price. Add rejected-order or backorder records when available, using order IDs to prevent double counting.

Availability must mean the customer could actually buy. Stock in inspection, in another warehouse or unavailable for the relevant channel may not qualify. A positive closing balance also does not establish that the item was available throughout the day.

The planner should record why the chosen comparison periods represent the missing demand. Forecasting: Principles and Practice discusses judgment when data are incomplete and recommends documenting and justifying adjustments. That discipline matters when the missing observations are precisely the purchases you want to estimate.

Work an explicit stockout estimate

Consider a US office-supplies retailer selling the same item at two locations in November 2026. This illustrative example assumes unchanged prices, no promotions, comparable trading days and no backorders. Both stores sell only when stock is available.

Each location covers 30 days: 20 + 10 at East and 30 + 0 at West. For this simple calculation, assume the average selling rate during available days also applies during unavailable days.

In-stock rate = Recorded sales units / Fully available days

Assumed lost demand = In-stock rate × Unavailable days

Estimated unconstrained demand = Recorded sales + Assumed lost demand

The reconciliation is observed sales of 600 + 900 = 1,500 units, plus 300 assumed lost units, giving 1,800 estimated demand units. The adjustment is 300 / 1,500 = 20.0% of recorded sales. Units are whole numbers; percentages are rounded to one decimal place.

At the stated unit price, the valuation is:

The assumed $7,500 is potential sales value under these assumptions. It is not recorded revenue or a profit estimate.

Show how sensitive the answer is

The most fragile input is the rate assigned to East's missing days. Keep all other inputs unchanged and supply alternative daily-rate assumptions:

These are scenarios, not statistical confidence limits. They show the consequence of changing an assumption. A planner should explain which comparable days support each rate before increasing a purchase commitment.

Validate and assign the decision

After stock returns, compare the estimate with sales on comparable fully available days. Match weekday patterns, prices and promotions. Do not treat a burst of delayed purchases as proof that all earlier lost demand was permanent.

The forecasting text's accuracy guidance emphasizes evaluation on new observations. Preserve the estimate before reviewing the recovery period; repeated revisions made after seeing sales cannot validate the original method. A monthly demand forecast provides the version history for that review.

Demand planning owns the estimate, operations verifies availability and the commercial owner reviews substitution or delayed purchases. Keep demand-estimation error separate from ordinary forecast bias by SKU and site.

Where it goes wrong

• Calling assumed demand observed demand. Preserve a separate adjustment column and its approval history.
• Using promotion days as normal days. A promoted selling rate may overstate missing demand at the regular price.
• Adding backorders twice. Captured unfilled requests may already belong to the demand total.
• Ignoring substitutions. Demand moving to another item or store may reduce the company-level opportunity.

Examine your demand assumptions in Covirage

Bring a sample of your own sales and availability exports with your supplied lost-demand assumptions. Covirage maps the columns, its tools compute and check the comparisons, and the external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring the assumptions you want to evaluate.

See Covirage for supply chain teams for the supporting workflow.