# Know what the market could buy. Test what could change it.

> SIMCEL builds a statistical baseline from sales history, then layers in commercial and market assumptions. Every scenario carries its own revenue and margin consequence.

Canonical: https://www.simcel.io/demand
Language: English
Other languages: Vietnamese https://www.simcel.io/vi/demand/index.md, French https://www.simcel.io/fr/demand/index.md, Japanese https://www.simcel.io/ja/demand/index.md
Index: https://www.simcel.io/llms.txt

SIMCEL cleans your history into true demand, builds the forecast method by method, and turns what your sales teams know into dated events. Every scenario shows volume, revenue and margin.

One forecast can’t explain every future.

A statistical forecast knows the history. It does not know that a competitor leaves the market next quarter, or that a key account has changed how it orders. That knowledge sits with your commercial teams, in calls and spreadsheets, outside the plan. And the financial impact shows up after the commercial discussion is over.

- **Several plans. Several assumptions.** Budget, IBP/S&OP and short-term planning each start from a different version of the future.
- **Market knowledge sits outside the baseline.** Prices, promotions, competitor moves and launches are added by hand, if at all.
- **Money shows up late.** Revenue and margin appear once the volume is already agreed.

Start with a baseline the data supports.

Fifteen forecasting methods compete on every product group. Each one is scored on error and on bias together, at the lags where your supply decisions are taken. The best score wins, and the planner sees the candidates side by side. Nothing is a black box.

- **Compare methods.** No single method fits every product.
- **Keep the baseline explainable.** The statistical forecast is a reference point, not the final answer.
- **Plan at useful granularity.** Product, customer, channel or market: the level the decision needs.
- **Put the effort where it pays.** The portfolio is segmented by value and by forecastability. Stable items run on their own. The planner's hour goes to the few that need it.

The baseline is only the start.

Your commercial teams know what history does not. In SIMCEL, each piece of that knowledge becomes an event: dated, sized, owned, and with its source. Create an event by hand. Or give Lana the spreadsheet of initiatives, or the transcript of a market call. Lana lists the events it found. The planner checks them and confirms.

More volume can still lose money.

- **Price changes behavior.** Model the customer's response, not a fixed volume.
- **Gross sales are not net sales.** Trade terms and discounts change with each scenario.
- **Read the margin consequence.** Each scenario goes down to net sales, cost of goods and margin.

Consumers buy at the shelf. You sell to the distributor.

Between what the consumer buys and what you invoice, there is the distributor's stock. SIMCEL models sell-out, distributor inventory and sell-in together, so your revenue follows the channel and not a flat assumption.

- **Start with consumer demand.** What really leaves the shelf.
- **Pass through distributor stock.** Inventory and days of cover create the timing gap.
- **Resolve to sell-in and revenue.** The company plan follows the channel.

Bring us the assumption your team is debating.

A promotion, a price move, a launch date. We turn it into an event and show what it does to demand, then to supply, revenue and margin.

Forecast demand, not sales

A stockout is not a drop in demand.

A forecast built on sales history learns past supply, not past demand. When a product was out of stock, sales were zero, and the next forecast carries that zero forward. SIMCEL cleans the history before it forecasts. It flags the outliers, applies the corrections it already knows, and asks the planner to validate the rest. An unexplained peak gets a name: a promotion, a tender, a one-off order.

Stockouts flagged, not learned.

Promotions separated from the baseline.

Every correction validated by a planner.

Already running a forecasting tool? Keep its baseline, and start at the next step.

Lana reviews the assumption

A promotion is planned at +25%. The last three delivered +10% to +14%. Lana proposes +12%, and the planner decides.

Review assumptions, not numbers.

The demand plan is shown in volume and in value, with the budget and the last cycle's plan as reference lines. Build an optimistic case and a pessimistic case from the same events. Each one carries through to net sales, gross profit and EBIT.

Against the budget and the last plan

An optimistic and a pessimistic case

A resilience plan for a disruption

Score the plan, not only the forecast.

Two numbers matter. Forecast accuracy scores the statistical forecast. Demand plan accuracy scores the plan the business committed to, and that is the number that cost money. SIMCEL keeps the two apart and tracks both, with the bias beside every error figure. Because the forecast and the plan are scored separately, the value the planning process adds can be tracked: Forecast Value Added.

- **Bias and error, always together.** One says which way the plan leans. The other says by how much.
- **Every figure with its lag, its level and its time bucket.** A number without them cannot be acted on.
- **Forecast and plan kept apart.** That is what makes Forecast Value Added measurable.

One demand plan. Supply runs on it. Finance reads it.

In SIMCEL, the demand plan is the first step of the IBP/S&OP cycle, not a file sent to supply. Supply tests it against capacity and stock. Finance reads its P&L. The budget stays beside it as the reference.

Explore the Supply module

Explore the Finance module

Two months of zero sales and a promotion peak. A forecast built on this history learns both.

The stockout is filled with true demand. The promotion leaves the baseline and stays as an event.

Optimistic and pessimistic cases
