You have landed on our UK Site.
US/Global

Demand Forecasting Best Practice

Head of Product Evangelism, Sammy Walton delivers his best practice guide to demand forecasting

Demand forecasting can be the foundation of all data-driven decisions that occur in the business workforce decisions. Having an accurate forecast allows you to provide better customer service, lower costs, and even improve employee happiness. 

Understand what drives your business

Naturally, the first step in generating a forecast is understanding what you’re forecasting, and the answer to what you should forecast is understanding what drives your workforce. It could be the number of sales, tickets, deliveries, items received, orders, appointments or anything else, but whatever the KPI, the one(s) you settle on should be tied directly to the reason for scheduling an amount of labour.

Once you’ve settled on the KPI’s you want to use, they can be stored in, integrated to, and viewed inside Quinyx right alongside the schedule. It’s these KPI’s that Quinyx’ A.I is going to generate a forecast, and these KPI’s that will help drive your scheduling requirements, be that automated or manual.

Think about the granularity

The next thing you’ll want to decide on is what time granularity you want to generate and view this data at. Quinyx can operate in 5, 15, 30 or 60 minute increments, but again, the decision is dependent on your business and how reactive you want the schedule to be in relation to changes in your business.

You may even want to consider seasonal changes and the effect they have on your schedules. Do you hire more flexi or temp staff during peak seasons? Or do you simply shift your workforce plan to account for this? When Quinyx A.I generates a demand forecast for you, seasonal, monthly, weekly or more granular changes and trends will all be taken into consideration.

Keep it simple to start 

You may have an existing labour model you want to use, and that’s great. But there’s something to be said for starting simple, especially when migrating to a new WFM system. Forecast models can be highly complex if you let them get that way, incorporating tens and sometimes hundreds of measurements that need to be taken into account, but complexity doesn’t always equate to more accurate forecasts and schedules. 

Starting with a simple model will also allow you to assess the outcomes, make changes and see impacts with a better understanding of how the outcome is affected, allowing you to revert or tweak additional changes more easily. 

Using Human intelligence and Artificial intelligence together

Quinyx has industry leading A.I that helps our customers forecast and schedule in more optimised ways, but we believe that human intelligence and artificial intelligence should work together to keep the business on track. 

Using artificial intelligence to generate a demand forecast is a great place to start. Our algorithms can perform huge numbers of calculations and comparisons and select the right forecast for the right KPI at lightning speed - a huge time saving. 

Combined with the human element, of marking any one off or new repeating events that affect the business and its labour pool, will help A.I become more accurate. 

What constitutes a good forecast?

Being able to generate an accurate forecast is a great start to optimise your workforce, but it's only part of the story. A good forecast will take into consideration both long term and short term trends, internal and external events and give you a realistic result to work with. It's also one that can be explained (how we got to this result).

A good forecast is also one that is accurate, yes, but we shouldn't be chasing 100% accuracy in forecasts. That time can be better spent fine tuning the schedule, so we want a forecast that we are able to generate a quality schedule from, one that works both for the employee and the employer, keeps the business moving in the direction of its goals and engages the workforce. We call this next step workforce optimisation. 

Find out more about demand forecasting, or alternatively get in touch so we can start a conversation