Member Article

Forecasting: new promise after so many false starts

William Hesketh Lever, the founding father of Unilever and one of the pioneers of modern marketing, once said that “I know half my advertising isn’t working, I just don’t know which half.”

Perhaps we shouldn’t be surprised; assessing advertising effectiveness is notoriously difficult, and Lever was talking a century ago. What is surprising is that the same thing could be said of forecasting in businesses today, particularly given its central role in both the supply chain and business planning more generally.

Recent research suggests that 50 percent of forecasts are less accurate than a simple ‘naïve’ forecast (i.e. using the prior period’s actual as the forecast for the current period), and the reason for this is inappropriate forecasting interventions. On average, across the entire portfolio, forecasting performance often fails to improve on this crude procedure. If this is right, it means that the millions of pounds that companies like the ones surveyed have spent on forecasting software and on sophisticated processes have yielded zero return, or worse still, destroyed value. They could have achieved the same results using the old fashioned ‘replenish what you have just sold’ philosophy that Lever himself might have used.

‘What continues to surprise us’ says Crispin Mair of Crimson & Co, the UK’s leading independent Supply Chain consultancy, ‘is that most companies are unaware of the level of waste in their forecast processes because the metrics they are using to measure performance are too simplistic, time consuming and uninformative.’ Another recent piece of research identifies poor forecasting as the main reason why 1000 top US companies surveyed had a staggering $1.1trillion of excess working capital. ‘Corporate America is littered with companies that have invested heavily in demand software but have little to show for it’ according to CIO magazine.

But is forecasting software to blame? Not according to Mair. ‘Most forecasting packages will do a good job if used appropriately. The corollary is that, used inappropriately, most packages will also fail to beat a simple replenishment strategy. There is no single, silver bullet - and this is where most companies’ expectations become frustrated,“ he says.

It is like buying a super-car; the additional horsepower offers no guarantee that you will get from A to B more quickly, without crashing. You need to learn how to drive it. In the same way you need a Forecasting Performance Management system (FPM – see below) to help you learn when to keep things simple, when to overlay judgement on system-generated forecasts and when to open the throttle and let the forecasting engine do what it was designed to do’.

Many companies hope that investing in a sophisticated statistical package will solve all their forecasting problems, but these hopes often turn to frustration when the expected benefits are not immediately forthcoming. You need to learn when, where and how to use it to extract value, which starts with analysing and tracking forecast performance in a way that is relevant to the business; ‘It shouldn’t be about accuracy, it should be about adding value’.

To help companies striving to drive down inventory and improve service Crimson & Co have been working with CatchBull, a leading provider of Forecasting Performance Management software, to bring their toolkit called ForecastQT, to companies that stand to benefit most from its approach.

“Whether your company has invested in a super-premium forecasting system costing millions or just home-grown Excel solutions, a FPM system that measures and re-calibrates performance each time a forecast or actual is updated, is likely to be the quickest route to delivering the benefits you thought you were buying with your forecasting software investment.”

What is a Forecasting Performance Management system?

A FPM is a new type of application which helps managers maximise the value of their forecast process in the same way that Engine Management Systems optimise the performance of car engines. It helps managers set up their process in the best way to maximise forecast value added and provides practitioners with the continuous tailored feedback they need to make sure the system stays in tune as well as the forensic tools to diagnose and fix systemic problems.

CatchBull is a pioneering British start-up company helping to define and develop this category.

* MORLIDGE, S. (2014) Forecastability and Forecast Quality in The Supply Chain. Foresight, 33, 26-31.) http://forecasters.org/foresight/

By: Dr Steve Morlidge, Product Director at CatchBull

About CatchBull:

CatchBull is an innovative software company with a mission “to make forecasting pay” – particularly in the supply chain. After much research, CatchBull identified why most forecasting processes fail to live up to expectations. In response, CatchBull is first-to-market with a Forecast Performance Management system that sits on top of any forecasting system and process to constantly measure (also in $/£/€) and recalibrate forecasting performance as it happens.

The company was founded in 2010 and, by 2013 was named as one of Gartner’s Cool Vendors in Supply Chain Management. Since then it has been helping clients in N. America and Europe to improve their forecasting effectiveness. For further information, please see: www.CatchBull.com

[1] Dr S Morlidge ‘Forecastability and Forecast Quality in The Supply Chain.’ Foresight, Spring 2014, 26-31.

This was posted in Bdaily's Members' News section by Steve Morlidge .