Demand Forecasting

With Artificial Intelligence (AI) based demand forecasting, one aims to forecast future customer demand over a defined period using historical data and other information such as seasonality, trends, macroeconomic factors, etc.

How does Demand forecasting add value?

By knowing the demand for products one can decrease inventory costs by reducing or eliminating redundant and obsolete inventory thus reducing the inventory carrying costs. Without demand forecasting one is at risk of making poor decisions that may increase inventory holding costs, poor supply chain management, and not having enough inventory all negatively affecting customer satisfaction and profitability.

Reduce inventory carrying cost
With better predictions of demand, one can better plan the stockouts to minimize overstocking.
Supplier relationship management
By having concrete numbers of future customer demand, it’s possible to calculate how many products to order. Thus making it possible to decide whether you need new supply chains or to reduce the number of suppliers.
Increase Customer Satisfaction
Customers want the product available immediately when they plan on buying something. If it is not, one does not only risk creating a bad customer experience but also losing the sale itself. By better anticipating the demand you can determine which categories of products need to be purchased in the next period from a specific store location.
Inventory optimization
Avoid over- and under-stock situations with demand forecasting as one can optimize the supply chains ahead of time. Making sure inventory is in place to meet demand while also not overstocking such that unsold goods won’t occupy retail space.
Marketing campaigns
By understanding how marketing will drive the demand, one can either use marketing to increase the demand or refrain from marketing during certain periods if one already knows the demand will be close to the limit of what the supply chains can handle.
Enterprise resource planning
Using the forecasted demand one understands how one needs to plan productions to meet it.
Providing Insights
With a model of the demand forecast, one can use it to gain better customer insights to understand how various drivers affect sales, such as media, price, etc.

Why Use Artificial Intelingce for demand forecasting?

We live in uncertain times, with high inflation and unique challenges such as the COVID-19 pandemic. It is crucial that the forecasts can handle these rapidly changing situations. This comes down to having a sane model that expresses real-world dynamics and also having proper uncertainty estimation in the forecasts. Thus one can feel confident in being able to plan for the best and worst-case scenarios.

Some of the way Artificial Intelligence and Machine learning improves demand forecasting are

Dynamic
Enable to run the forecasts more frequently as new data arrives
01
Accurate
Better models provide better forecasting
02
Automation
Warnings can be triggered as soon as new data shows warning signs of increased/decreased demand that requires updates to the supply chain.
03
Efficient
With machines doing the work, it is possible to analyze a lot more data than a human possibly could.
04
Uncertainty Estimation
Propper uncertainty quantification of the forecasts
05

When using Artificial Intelligence expect an improvement of 10 % - 35 % over the baseline set by exciting demand planers. But the goal is not to replace the demand planners, but rather to make the demand planning team more efficient and accurate. As the demand planner can always adjust the forecasts with information that has not been provided to the mode ex. Information from communications with the client.

Customer Case: Forecast fuel sales

How we help Uno-X with demand forecasting of fuel sales on an SKU level per gas station

Read Now!

What is needed?

The better understanding the model has of the dynamics of the business and the causal drivers the better demand forecasts are possible, normally the model gets better as more demand drivers are provided such as

macroeconomic factors,
covid restrictions,
promotions,
Marketing,
pricing etc.

When one has the data for all the drivers one needs to consider, one also needs to decide the granularity of the data one wants to use. In general, the lower the time scale the more stochastic sales and demand tend to be, so depending on the specific application one must find correct data granularity, should one predict hourly, daily, or weekly demand?

Choosing the correct data to use when creating a model for demand forecasting is vital,
What drives demand will be different from business to business, but some high-level questions one should think about it:

How big an impact does pricing have and id it change recently?
Does media advertisement impact my demand?
Did we recently sell out thus creating a build-up of demand?

But beyond that, there are a vast number of potential drivers. Internal drivers such as pricing can be controlled by the business and and external drives such as weather and macroeconomics will affect the demand, but the external drivers can not be controlled.

What can we do?

There is a wide range of areas where our models add value, here is an example of some of the models we support.

Demand forecasting

In demand forecasting, one aims to forecast future customer demand over a defined period using historical data and other information such as seasonality, trends, macroeconomic factors, etc.

By knowing the demand for products one can decrease inventory costs by reducing or eliminating redundant and obsolete inventory and at the same time reduce the inventory carrying costs. Without demand forecasting one is at risk of making poor decisions that may increase inventory holding costs, poor supply chain management, and not having enough inventory all negatively affecting customer satisfaction and profitability.

Logistics Demand Forecasting

In logistics demand forecasting the goal is to accurately predict the demand for products, services, shipments etc. throat the supply chain. Depending on the use case,

It can be predicting demand for the next 48 hours in order to enable better planning
Dynamically predict when a product will run out and trigger alerts to get new deliveries done on time.

Logistics forecasting is a complex process and in many cases influenced by things such as weather patterns, fuel costs, seasonality, and macroeconomic factors. By integrating demand forecasting one can get reduced operational costs, enable dynamic pricing, increase efficiency in terms of employees and/or vehicles, and better scalability.

Inventory Optimization (IO)

With inventory, the goal is to minimize it as much as possible and thus freeing up working capital while at the same time not running out of inventory resulting in a los of sales. By forecasting the demand and combining that with the current level of inventory one can easly plan when more inventory is needed. Even trigger alerts when one is at risk of running out of inventory given the expected demand.

Multi echelon inventory optimization (MEIO)

Going beyond simply inventory optimization to understand the drivers of excess inventory and understand how to balance cost and customer service. And how they will be impacted by pricing changes, media advertisements etc.

MEIO enables you to simulate various scenarios to analyze trade-offs between costs and service levels.

Time Series - Introduction

Introduction to how time series models can be used to add business value

Logistics

Use Artificial Intelligence to optimize the supply-chain, logistics operations, ...

Business Intelligence

Use Artificial Intelligence to boost your Business Intelligence