Operational Forecasting Workshop for Finance folks Dr Steve - PowerPoint PPT Presentation

Operational Forecasting Workshop …for Finance folks

Dr Steve Morlidge Unilever 1978 – 2006 roles include: • Controller Unilever Foods UK ($1 billion turnover) • 2002 – 2006 Leader Dynamic Performance Management Change Project (part of Unilever’s Finance Academy) Outside Unilever • Chairman of BBRT 2001 – 2006 • BBRT Associate/ Non Executive Board Member 2007 - • 2006 - Founder Director Satori Partners Ltd • 2005 – PhD Hull University (Management Cybernetics) • 2007 – Visiting Fellow Cranfield University • 2009 - Publish book ‘ Future Ready: How to Master Business Forecasting ’ • 2010 Editorial Board of Foresight Magazine • 2011 Founder CatchBull (Forecasting Performance Management Software) • 2017 Publish ‘The Little Book of Beyond Budgeting’ • 2018 Publish ‘The Little Book of Operational Forecasting’ • 2019 Publish ‘Present Sense: the Art and Science of Performance Reporting, with the Brain in Mind’

Agenda 1. Mutual Introductions 2. Hopes and fears 3. Forecasting Fundamentals 4. The Challenges 5. Forecasting Game 6. Better Forecasting in practice

Forecasting Fundamentals

Operational Forecasting Simulation Demand = two die • Three decisions • What should we hold in stock to start (X)? • What do we think we will sell next period (F)? • What should we order from our suppliers at the end of every period (Y)? • Period 1 Period 2 Period 3 Opening Stock X (Remaining + Order) Calculated Demand Dice Dice Dice Remaining (X-Dice) (Opening – Dice) Calculated Next period forecast F F F Order Y Y Y

Forecasting Exercise Debrief • How did you decide? • What are the commercial consequences of less than perfect forecasting? • Too much stock • Too little stock • How well did you do?

Two types of stock Start End Start End Start End Cycle Cycle Cycle Stock Cycle Stock Stock Stock Safety Safety Safety Safety Safety Stock Stock Stock Stock Stock Safety Stock Under Over Perfect Forecast Forecast Forecast

Safety Stocks and Service Levels

How is real life different? • Demand changes • We can’t predict demand perfectly • Failure to predict changes • Over forecast • Under forecast • We don’t know the probability distribution in advance • Many products • Different demand patterns • Different characteristics • Shelf life • Cost • Margin • Strategic significance • Sensitivity to failure to supply • Replenishment lead times

The Challenges

Forecasting Exercise 1. Pair up 2. Person 1 think of a number (n.n) • between 5 and 10 (include 1 decimal point) – don’t disclose it With eyes closed try to stop • the clock at n.n seconds Reset and repeat 10 times • 3. Person 2 Record results •

Stopwatch Game 1. Record actuals (A) 2. Guess the targets 3. Calculate the average A and compare to guess 4. Add hidden target (B) and calculate absolute difference (C) and average

Debrief • How far adrift were your guesses (percentage) ? • Why? • What caused the ‘actuals’ to vary? • Could this be eliminated? • Could this be predicted/forecast?

Economics of ERROR Forecast Error VARIATION BIAS Immediate benefit UNDER- OVER- UN- AVOIDABLE Delayed benefit FORECAST FORECAST AVOIDABLE Cash Benefit STOCK LOST SALES EXPEDITING COSTS Profit Benefit STORAGE COSTS OBSOLESENCE COSTS FINANCING COSTS

Forecasting Game

Forecasting Exercise 1. Working in pairs – given 10 months of history 1. Produce a forecast for the next month – statistics/judgement 2. Get the actuals for the month 3. Calculate the forecast error 4. Repeat for 14 periods 5. Calculate the average error 2. After 1. The winner gets a prize 2. Compare vs ‘perfect’ forecast 3. Calculate impact of avoidable forecast error 4. Calculate value added/destroyed 3. Debrief – what have you learned?

Forecasting Exercise - Template Plot actuals and forecasts Forecast – Calculate Provided one period in advance

Alternative forecasts Best Forecast Naive Forecast 120 120 100 100 80 80 60 60 40 40 20 20 0 Jan- Feb- Mar- Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec- Jan- Feb- Mar- Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec- 0 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 -20 Jan- Feb- Mar- Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec- Jan- Feb- Mar- Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec- 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 -20 -40 -60 -40 Non Winner -80 Winner -60 Actual Forecast Actual Forecast 120 Actual Forecast 120 120 100 100 100 80 80 80 60 60 60 40 40 40 20 20 0 20 Jan- Feb- Mar- Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec- Jan- Feb- Mar- Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec- 0 Jan- 19 Feb- 19 Mar- 19 Apr- 19 May- 19 Jun- 19 Jul- 19 Aug- 19 Sep- 19 Oct- 19 Nov- 19 Dec- 19 Jan- 20 Feb- 20 Mar- 20 Apr- 20 May- 20 Jun- 20 Jul- 20 Aug- 20 Sep- 20 Oct- 20 Nov- 20 Dec- 20 -20 0 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 -20 Jan- Feb- Mar- Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec- Jan- Feb- Mar- Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec- 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 -40 -20 -40 -60 -40 -60 -80 -60 -80

Forecasting Exercise Debrief 1. Analysis of results 2. Discuss in tables 1. How well did you feel that you were doing? 2. Were you surprised/disappointed by the results? 3. How easy did you find the exercise? 4. How could you do things better? 5. Has it changed your view about operational forecasting?

What could this be worth? Cost of Sales Per €1bn revenue 1 Total Cost of Error 4%-6% €20m-€30m Value Added by Forecasting 0%-2% €0m-€10m Avoidable Error 1%-3% €5m-€15m Avoidable Inventory 2 €1m-€5m 1 Assuming 50% Gross Margin 2 Assuming 10 weeks stock cover

Better Forecasting in Practice

How VALUE is ADDED Top 200 Packs by Size 100 Simple error statistics are 90 Destroying MISLEADING 80 Value 70 60 Absolute Error % Because volatile 50 demand patterns are less 40 FORECASTABLE 30 20 Adding Value also ‘same as last time’* 10 is the ultimate 0 0 10 20 30 40 50 60 70 80 90 100 ALTERNATIVE TO Demand Volatility % FORECASTING * naïve forecast

Forecastability trap: case study Because of differences in forecastability, the region with the lowest errors… Actuals Forecast Error Country A Country B Country C …often doesn’t have the best forecast Relative Absolute Error Mean Absolute Error Demand Volatility (MAE/DV) (Forecast error vs naïve error*) 1 Country A 7% 1 Country C 88% 1 Country C 0.53 2 Country B 11% 2 Country B 16% 2 Country B 0.70 3 Country C 47% 3 Country A 6% 3 Country A 1.02 * Naïve forecast = sell this month what we sold last month

Any other questions?