JANUARY 2020

Delivering on time and in full, in unprecedented times

Demand planning to navigate covid’s extreme volatility

January 2020 saw us kick off a demand planning programme with Portugal-based, The Navigator Company, one of Portugal’s largest exporters of tissue and paper. Little did anyone know that a global pandemic was round the corner – and was going to put our tools and implementation approach to the ultimate test.

The Navigator Company is the country’s 3rd largest exporter, accounting for 1% of the nation’s total GDP, producing 1.6 million tons of paper, 1.5 million tons of pulp each year. Their products are shipped to approximately 130 countries in 5 continents. So when the pandemic hit, our implementation approach withstood a double whammy of volatility. Suddenly, global supply chain disruption was in the evening news, and social media feeds were full of toilet paper memes as demand for it hugely increased. This was to be a challenge unlike any anyone had ever faced.

Despite all this, I-Plan was able to deliver on time, and in full, all while implementing our tools and processes remotely. Here’s how.

Using traditional and non-traditional methods to forecast demand

When the huge spikes in demand for toilet paper began impacting The Navigator Company, we had to reverse our usual approach of data first, forecasting methods, second. The nature of the extreme demand and the pandemic circumstances meant we started with statistical forecasting. Statistical forecasting often provides the basis for the forecast accuracy improvements realised by our customers.

For Navigator, we used 5 years of sales history to generate a forecast for 2020 by overlaying the historical data. The measure we use is called the Mean Absolute Percentage Error (MAPE) – and it calculates how the accuracy of a forecast system. The industry average is a difference between the history and the forecast of 20%. Our MAPE on this project achieved a world class standard of less than 10%.

In addition to the statistical forecast gleaned from historic sales, we leaned heavily on I-Plan’s progressive forecasting techniques:

Event forecasting functionality – to recognise anomalies in the data
Dynamic regression – to synthesise future external market trends and forecasts with the historical picture
Expert selection algorithm – to apply the best for forecasting methods for each SKU

These banks of different forecasting methods for different data sets, combined with our special algorithms and machine learning allowed us to also include external trends such as the country’s GDP.

In this way, we were able to account for the extreme blip that saw the spike in demand for toilet paper, and meet it. But we could also safely forecast 2021 by ring fencing 2020’s pandemic data as an anomaly – and not let this extreme spike impact future forecasts.

Conclusion

In times of such extreme volatility it’s tempting to say it’s pointless to make a forecast. I-Plan, however, believes that a forecast will always be wrong precisely because it is only a forecast. The real task at hand is to use the correct tools to minimise error. And when statistical methods prove less effective, I-Plan comes with a toolkit that can adapt to volatility. The proof is in the pudding, of course.

Despite the many challenges, and the uncertainty, I-Plan was able to deliver significant improvements on their KPIs:

20% forecast accuracy improvement at SKU level
50% reduction in the sales team’s manual effort to prepare and distribute the forecast across the business.

The Navigator demand planning programme was concluded on time in June 2021, having realised the expected benefits. This was the first I-Plan implementation to be entirely remotely due to the Covid-19 travel restrictions.

Give I-Plan a go

Get started with a free demo

Request a demo