31 July 2010

Viewpoint : John Lamb

Improving forecast accuracy

Published:: 27 February 2009
Article Type:: Viewpoint
Byline:: John Lamb

Demand planning is vital during the recession if organisations are to avoid being left with inventory that has to be sold at marked-down prices. However, the track record of some forecasters is not good. On the macro-economic level, analysts failed to predict seven out of the last ten recessions. At the supply chain level too many suppliers, manufacturers and retailers have been muddling through on the notion that demand is predictable. In some industries demand forecasting accuracy is below 20 per cent.

So what steps can a business take to get closer to real customer demand and respond quickly? An obvious first move is to make sure the sales and operations planning process is up to speed. "Organisations that excel in this area are better able to absorb the uncertainties in today's economies, reporting lower inventories, faster cash-to-cash cycle times and better customer service," says Andrew Kinder from supply chain management software firm Infor.

Kinder says it is important to improve intelligence gathering by getting sales, marketing and development people working together with customers in predicting the trends that statistics may not pick up. Some industries such as fast moving consumer goods are better than others at compiling and analysing such information to produce more accurate forecasts.

In many organisations there is a mismatch between the people involved in forecasting and those whose job it is to manage inventory. "Those who see demand and inventory as tightly linked are poised to jump back into the market when conditions are right; says Harjot Sachdeva from forecasting software company i2.

Another benefit of this closely managed inventory is that it frees cash and improves supply chain responsiveness. Sachdeva is adamant that organisations must monitor both their cycle stock and their safety stocks.

For many manufacturers, controlling inventories involves getting a better handle on customer demand at the top of the supply chain. Whereas previously manufacturers might be only dimly aware of end-customer activity they now need to be much better informed, says Sachdeva.

"Suppliers and manufacturers need to create a more open and honest relationship - one that acknowledges demand variations, and shares risks and rewards among both businesses; says Sachdeva. "The traditional manufacturersupplier relationship has been based on the notion that end-user demand is predictable and in many respects characterised by a sort of unfounded optimism."

Opinion is divided about how best to improve forecasting techniques. Some argue that it is just a matter of tweaking existing models by giving more weight to economic factors. "Demand signals don't change their characteristics;' says Rod Daugherty. a senior director at Manhattan Associates.

Daugherty points out that his company has put its efforts into incorporating existing algorithms for forecasting into its software including Croston and Holt-Winters. These methodologies take account of intermittent demand. putting more emphasis on recent data points rather than historical ones.

Others see a need for more far reaching changes. For example, Professor Amir Sharif who heads the Operations and Supply Chain Systems research centre at Brunei University believes IT systems need to have alternative views of the world that take account of unforeseen events.

Prof Sharif believes organisations should pay more attention to managing decisions and their consequences, rather than just solving supply chain problems. 'Understanding the human component to globalised business is more important than ever before and is where the weakest link typically lies; he observes.

Among other things Prof Sharif advocates a MySpace or Facebook for supply chain management that would allow organisations to adjust their own local parts of the supply chain.

IT companies argue that investing in software tools is one way to improve forecast accuracy, but independent observers tend to counsel caution.

"Now is not the time to be investing in forecasting systems when consumers do not know what will happen tomorrow - unless there is a serious problem; says Tim Robinson, a principal at retail consultancy Kurt Salmon Associates.

Retailers should concentrate on increasing onshelf availability, reducing inventory levels and reinforcing collaborative planning, forecasting and replenishment processes, says Robinson.

Comments

The following comments have been posted in response to this article:

Its not just about Forecast Accuracy in these uncertain times, but also about agility and responsiveness to change and the ability to handle that change there is an area that many Supply Chain experts do not understand, is the connection and relationship between Forecast Variability and Demand variability. If we understand these we understand the need for demand sensing tools, inventory optimisation solutions and the need for visibility into the supply chain. If we can start to plan for variability using time phased safety stock then we stand a better chance of journeying through these turbulent times successfully. As an example we would all understand, you want significant safety stock for lemon jucie leading up to Shrove Tuesday as there is significant variability, but you want in your planning system a few of future required safety stock time phased before and after the event. The same logic then applies to promotional events, new product launches and step changes in pricing and competition. I know I have done it many times before, its a change in thinking first then you start thinking about further solutions.

Posted by: Dave Food 14 Jul 2009

Executive Jobs
Job Title	Job Location	Job Position
3PL Commercial manager ...	London	Permanent
Assistant Category Manager ...	Wiltshire	Part Time
AVP Software Relationship Management ...	London	Permanent
Category Director- Property ...	London	Permanent
Category Manager ...	Yorkshire	Permanent
Category Manager ...	London	Permanent
More Jobs