Use case · Demand & seasons
Demand & seasons
Forecasts that know your year: seasons, holidays, the weather where you trade, fuel prices and exchange rates.
Questions Avir answers
- 01What will orders look like over the next twelve months, and how sure are we?
- 02How much do a wet summer or an early Easter move demand for us?
- 03Which months are we most likely to be wrong about?
The situation
A retailer planned spring stock from last year's numbers.
Assumed
This spring will look like last spring, plus growth.
Simulated
Easter falls three weeks earlier this year, pulling April demand into March. Last year's plan was a month out.
Bring spring stock forward a month.
An illustrative example. Your simulations run on your own customers.
What the simulation shows
The result, by customer group.
Every scenario comes back as a month-by-month forecast with its range, and the difference for each group of customers.
Scenario
Easter three weeks earlier than last year
March +14%, April −11%
Orders by group in March, against last year's timing
What goes into it
What you need to get started.
Your data
- Orders, invoices or subscriptions, each with a customer, a date and an amount
- Two years of history or more, so the seasons can be learnt
The world outside
- Seasons
- Holidays
- Weather
- Fuel prices
- Exchange rates
- Your own monthly series
Time and cost
- A forecast in seconds, once your data is connected
- Scenarios without new behaviour run on the agents alone, in seconds
- The cost in AI credits is shown before anything runs