# How Does Demand Forecasting Work for Consumer Brands?

{% hint style="info" %}
**Quick Answer:** Demand forecasting analyzes historical sales data, seasonal patterns, and external signals to predict future customer demand. For consumer brands, this typically means generating SKU-level unit predictions across a rolling planning horizon — usually 3 to 12 months — to inform purchasing, production, and inventory decisions.
{% endhint %}

**Demand forecasting is** the process of estimating how much of each product customers will buy over a future time period. Consumer brands use demand forecasts to determine what to produce or order, when to place purchase orders, and how much safety stock to hold to protect against uncertainty.

Accurate demand forecasting reduces two of the most costly inventory problems: **stockouts** (lost revenue when customers can't buy) and **overstock** (cash tied up in inventory that doesn't move).

## How the Forecasting Process Works

{% stepper %}
{% step %}

#### Collect Historical Sales Data

Forecasting starts with your actual sales history — typically 12 to 24 months of data at the SKU or variant level. More history improves pattern recognition for seasonality and trends.
{% endstep %}

{% step %}

#### Identify Patterns and Signals

Forecasting models analyze:

* **Trend** — Is this product growing, declining, or flat?
* **Seasonality** — Does demand spike in certain months or weeks?
* **Promotions** — Did past campaigns create demand spikes that shouldn't be extrapolated?
* **New product behavior** — How do products without sales history ramp up?
  {% endstep %}

{% step %}

#### Generate the Forecast

The model produces a unit-level prediction for each future period. The output is typically a weekly or monthly demand estimate by SKU, variant, or channel.
{% endstep %}

{% step %}

#### Review and Adjust

Planning teams review the statistical baseline and apply overrides for known events — upcoming promotions, new product launches, market changes, or supply constraints. This is the "human-in-the-loop" step.
{% endstep %}

{% step %}

#### Feed into Planning Decisions

The approved forecast drives replenishment planning, purchase orders, production schedules, and inventory allocation.
{% endstep %}
{% endstepper %}

## Types of Demand Forecasting Approaches

{% tabs %}
{% tab title="Top-Down" %}
**Top-down forecasting** starts with a total revenue or unit target and breaks it down by product category, then SKU.

Best for: Brand-level planning, financial budgeting, annual targets

{% content-ref url="../../planning-and-execution/forecasting/how-to-create-a-forecast/generate-a-top-down-forecast" %}
[generate-a-top-down-forecast](https://learn.moselle.io/planning-and-execution/forecasting/how-to-create-a-forecast/generate-a-top-down-forecast)
{% endcontent-ref %}
{% endtab %}

{% tab title="Bottom-Up" %}
**Bottom-up forecasting** starts at the SKU level and rolls up to totals.

Best for: Operational purchasing decisions, replenishment planning, SKU-level accuracy

{% content-ref url="../../planning-and-execution/forecasting/how-to-create-a-forecast/generate-a-bottom-up-forecast" %}
[generate-a-bottom-up-forecast](https://learn.moselle.io/planning-and-execution/forecasting/how-to-create-a-forecast/generate-a-bottom-up-forecast)
{% endcontent-ref %}
{% endtab %}

{% tab title="Blended" %}
**Blended approaches** combine both — using top-down targets as guardrails while bottom-up SKU forecasts drive operational decisions.

Best for: Mid-size and growth-stage brands balancing finance and operations

{% content-ref url="../../planning-and-execution/forecasting/how-to-create-a-forecast" %}
[how-to-create-a-forecast](https://learn.moselle.io/planning-and-execution/forecasting/how-to-create-a-forecast)
{% endcontent-ref %}
{% endtab %}
{% endtabs %}

## What Affects Forecast Accuracy for Consumer Brands

* **Data quality** — Missing or inconsistent sales history degrades accuracy
* **SKU complexity** — High SKU counts with irregular demand patterns are harder to forecast
* **Seasonality depth** — Deep seasonal peaks (e.g., holiday, back-to-school) require sufficient history to detect
* **New products** — No history means the model must use comparable product proxies
* **Promotions** — Promotional demand must be separated from baseline demand

## How Moselle Handles Demand Forecasting

Moselle generates forecasts at the SKU level using your connected sales channels and historical order data. Forecasts can be refined using Mo, Moselle's AI assistant, and adjusted manually before being used to drive replenishment plans and purchase orders.

{% content-ref url="../../planning-and-execution/forecasting/how-to-create-a-forecast" %}
[how-to-create-a-forecast](https://learn.moselle.io/planning-and-execution/forecasting/how-to-create-a-forecast)
{% endcontent-ref %}

{% content-ref url="../../planning-and-execution/forecasting/refine-your-forecast" %}
[refine-your-forecast](https://learn.moselle.io/planning-and-execution/forecasting/refine-your-forecast)
{% endcontent-ref %}

## Frequently Asked Questions

### How much historical data do I need for demand forecasting?

**Answer:** A minimum of 6 months is needed to detect basic patterns. 12–24 months of history is recommended to capture seasonality accurately.

### Can demand forecasting account for promotional events?

**Answer:** Yes. Moselle lets you upload a marketing calendar so promotional periods can be isolated and modeled separately from baseline demand.

### What happens if a product has no sales history?

**Answer:** Moselle can use comparable product data or allow you to manually seed initial demand estimates for new product introductions.