Demand Forecasting Seasonal Horror Stories to Dodge

We enjoy watching horror movies, stock car races, and dodge ball here in the south for the same reason, the chance to watch something go wrong. Companies are watching the back to school sales for signs of our economy’s strength and leading indicators for holiday shopping; will something bad happen? This is a good time to review some epic inventory replenishment mistakes made by others in the past. I hope these horror stories help you learn better methods for managing inventory or at least make you say, hey at least we didn’t make that mistake in our demand forecasting and inventory replenishment.

A major office supply vendor was providing ‘Back to School’ VMI for one of our larger retail customers. After the back to school season had peaked, a huge shipment of school suppliers was shipped to the retailer. This new replenishment order wasn’t needed and the season was past. You can imagine the meetings and phone calls that ensued with the retailer needing the floor space for new product and the supplier stuck with product at multiple location that had low cost and heavy weight constraints.

I had read several interviews, press releases, and supply chain periodicals about the new software this huge office products supplier (vendor) was using for VMI. They touted their VMI solution as the evolution of years of research, advanced well beyond many retail replenishment systems. Since our software contained all the sales and inventory in our data warehouse, the retailer requested our review to determine why things went so badly. We found three very basic issues that together created an expensive time bomb for everyone involved.

Demand Forecasting Algorithms
Changes in Fiscal Weeks where Seasonality Peaked
Lack of Visibility into the total Supply Chain

Demand Forecasting: ‘I know what you did last summer’

Do You? Or are you only looking at the last 6 weeks. Any Demand Forecast is going to be constrained by four basic fundamentals:

Forecast Methodology Used – Demand or Sales Forecasting (See our blog: Differences in Sales and Demand Forecasting)
The data available to use in forecasting.
The skills and experiences of the people running the forecast software.
The algorithms and methods used to calculate a forecast.

Today many retailers, wholesalers and manufacturers don’t have a demand forecasting software package; many retailers still use Excel for store replenishment and planning. In our ‘Back to School’ horror story, the reasons for the later orders were simple, they were using the last 8 weeks sales as leading indicators of future sales. The forecast engine was only using a time series algorithm. The time series algorithm had several methods it could use in the calculation but it was still constrained to time series. A time series algorithm needs at least two years of data to be successful. The software needed some additional forecast algorithms and the ability to pattern after old products for new product forecast models.

Bad Demand Forecast – The VMI software was using a time series forecast algorithm that included several methods. Data modeling, other algorithms, seasonal indexes, and human input were needed to deliver better results.

Differences this Year versus Last Year in Fiscal Calendar and School Calendar

The VMI software didn’t adjust for the change in school calendar start across fiscal weeks this year compared to last year. School calendars can change yearly and coupled with company calendars where the fiscal weeks don’t line up from one year to the next is a recipe for disaster. The next major mistake by the VMI software was a double hit, seasonal indexes peaked in the wrong fiscal week this year and school calendars had changed. When a forecast software using a time series algorithm to develop a base forecast, then multiplies that poor base forecast by the wrong seasonal multiplier, the errors get really big, really fast late in the season. Because the stores were front loaded, no history exception or management review caught that there was an error in the seasonal formula until it was too late.

Supply Chain Visibility is all about Everyone on the Same Page – all the time

The supplier’s VMI software didn’t have any real collaboration tools. If the retail corporate and store locations could see the VMI demand forecast in a line graph and compare this year to last year, changes would have been obvious. If retailer could see sales and inventory projections across the entire back to school season, again, mistakes could have been avoided. In the market today, there are solutions, like ours that offer full visibility of forecast, inventory, sales and optimization models, available to retailer and supplier with dynamic updates. There are other software companies that can deliver this capability and others that have this in development. To compete in the current market, companies will need to invest in true collaboration software that offers dynamic updates and full supply chain visibility.

Demand Forecasting, Visibility and Planning allow you to Dodge the Horrors

Even with the state of art forecasting and replenishments systems and the luxury of supply chain visibility, processes like Vendor Managed Inventories require collaboration, a partnership with agreed upon goals and ongoing monitoring, particularly with two new partners. If the two companies involved had established an operating framework: reviewed historical sales, agreed upon common goals for the promotional period, an action plan with dates, activities, and roles, there would have been no surprises. They would have noticed the differences in calendars and forecast versus actual variance along the way and adjusted their expectations. At some point they also would have measured the effectiveness of the promotion mid-flight and made better buying decisions, dodging the bullets of overstock, operational pains, along with the expense of returns and write offs.

New technology and software supports collaboration, streamlines processes, and makes information sharing easier for all parties, but you still need to plan and communicate.

Are you ready to ‘Tighten the Links in Your Supply Chain?™’

Demand Forecasting, Seasonality and Real Time Supply Chain Visibility, designed from day one for you by retailers and wholesalers.Contact us for a free review of your current demand forecasting and Replenishment issues. We have the experience and tools to help you improve your business. Also, request a demo to learn how our software can reduce out-of-stocks from poor lead time performace and increase sales. We install in 30 days at a fraction of the cost of legacy systems.

About
Latest Posts

Stuart D

Stuart Dunkin is a strategic visionary who leads by calling on his broad range of experience as a former retail executive as well as a consultant for top retailers and for the E3/JDA software company. He founded Data Profits to solve a real-world problem plaguing retailers today – the disjointed relationship between supply chain data, people and business goals. He has experienced first-hand the challenges and frustrations that occur when decisions are made based on wrong data, which often leads to the wrong actions being taken with crippling results like overstock, out-of-stock, and increased working capital.
Stuart graduated from Auburn University, with bachelor of science in business administration.