> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lanternhq.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Review demand forecasts

> Understand why expected demand can differ from a historical average, inspect the customers behind it, and relate the forecast to your stock target.

Open a forecast when you want to understand why Lantern expects an item to
sell more or less. Sales history tells you what happened. The forecast estimates
what is ahead, so it can differ from a simple average of past usage.

Your review connects that estimate to the customers and events you know. A
seasonal pattern, a completed project, and a customer moving business to another
supplier can call for different decisions even when their recent sales totals
look similar.

## Why not just use average sales?

An average spreads historical sales across a period. That can hide when the
demand occurs and whether it is changing. For a seasonal item, compare the
forecast with the relevant season as well as the latest months. A forecast can
be lower than summer demand and still be higher than last winter.

Use the displayed history to challenge the forecast. An unusual order is not
automatically a mistake, and a model cannot be assumed to know whether a
customer's project will repeat.

## Inspect an item's demand

1. On the recommendations page, find the item and location, then select **Forecast**.
2. In **Demand Forecast Details**, confirm the item and the location being shown.
3. Compare historical demand with the forecast over the periods displayed.
4. Use the **Location** and **Company** views to compare the selected location
   with the wider demand you are authorized to see.

Start with the location where you are making the stocking decision. Company-wide
demand adds context, but a strong company-wide trend may be concentrated at
other locations.

### Is the forecast the amount I should keep on the shelf?

No. Read the series labels shown in your chart:

| Series | What it explains |
| - | - |
| **Past demand** | The recorded demand for the displayed periods. |
| **Monthly forecast** or **Forecast** | The baseline estimate of future demand. |
| **Service-target forecast**, where shown | The demand estimate at the item's service target, showing protection against demand uncertainty. |

The service-target estimate is not the complete inventory target. Lead-time
buffers, replenishment coverage, boundaries, and overrides can also affect
Min/Max. Select an underlined **New Min** or **New Max** to see how those factors
shape the target.

For example, quick replenishment can support a target below a month's demand
because the stock is replaced during that month. Conversely, a low baseline
forecast does not mean Lantern has planned no buffer. Compare both the demand
and the coverage before deciding whether the target is too high or too low.

## Look for an explanation

| Pattern | Questions to ask |
| - | - |
| A recent spike | Was this a one-off project, a promotion, or the start of recurring demand? |
| A decline | Has a customer stopped buying, moved locations, or delayed an order? |
| A repeating seasonal pattern | Does the forecast reflect the timing and size of the next season you expect? |
| A sudden change at one location | Is it local demand, or has the way the item is supplied changed? |

## Was the demand one customer or a broad pattern?

Select historical sales activity in the chart to open **Customer Analysis**
where that detail is available. Some points open **Committed orders** first;
use **View customer sales** to continue to customer activity.

Confirm the period and location in the analysis header. **All Locations** is
broader than the one branch you may be deciding for.

| What to inspect | What it helps you understand |
| - | - |
| **Total Units** and **Total Customers** | The size of the activity and how widely it is spread. |
| **Top 5 Share**, when shown | Whether a few customers account for much of the period's demand. |
| **QTY** and **Qty %** | Each listed customer's quantity and share. |
| **Total POs** | The order activity shown for that customer. |

Use **All Customers** when you need the full list. Check whether demand is
concentrated in one project or spread across regular customers before deciding
on an override.

## Plan a new or replacement item

If the item has little history because it is new or replaces another product,
use **[Item Links](https://docs.lanternhq.com/features/item-links.md)** to bring in relevant demand. Choose
**Replacement** when demand is moving to the new product, or **Reference** when
an existing equivalent is a useful starting point and will continue selling.

## Bring in business context

If customer information changes the inventory decision, record a temporary
target [override](https://docs.lanternhq.com/guides/overrides-and-boundaries.md) with a reason and review date.
If the concern affects many items, investigate the shared input with Lantern.
The adjustment records your chosen target; the forecast remains evidence for
the review.

## Check the full recommendation

Demand is one input. A target can also change because replenishment takes
longer, your ordering interval changed, or a different buffer is needed. Review
[lead times](https://docs.lanternhq.com/guides/lead-times.md), [order frequencies](https://docs.lanternhq.com/guides/order-frequencies.md),
and [safety stock](https://docs.lanternhq.com/guides/safety-stock.md) when demand alone does not explain the
result.

Before returning to your shortlist, be able to explain the demand pattern you
observed and whether you need an adjustment or further investigation.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.