How forecastable is your business, and how does that compare to best-in-class companies? Can real-time data and artificial intelligence create better forecasts, and just how much better are they? How do industry leaders make inventory work harder, and what does that even mean for you?
These are all common questions that leaders in the field are asking with one goal in mind: how to significantly improve demand planning and forecasting performance to achieve the most agile, digitally-enabled supply chain.
So what does best-in-class forecast accuracy look like?
As mentioned in our recent webinar, State of the Nation for Planning – Key Findings from E2open’s 2019 Forecasting and Inventory Benchmark Study, several factors go into understanding best-in-class planning performance and forecast accuracy. The first step is to understand just how easy your business is to forecast by using a naïve forecast, which requires no additional investments, technologies, model tunings or planning systems. But business context matters and the perception of forecast accuracy depends on your role in the company.
According to the 2019 E2open Forecasting and Inventory Benchmark Study, “For the CFO making quarterly financial projections for the company as a whole, accuracy is typically greater than 90%. At this level, there may be little perceived benefit from investments to improve forecast excellence. In the supply chain, however, forecast accuracy matters—especially for fulfillment activities that rely on demand predictions at the weekly-item-location level. At this level, accuracy drops to 52%, resulting in excess or insufficient inventory, impacting return on capital, eroding margins and putting service at risk.”
Making sense of demand sensing
So you have your forecastability grounded, but now comes the hard part—improving your performance and reducing errors to attain a competitive advantage in the industry. That’s where demand sensing comes into the picture.
First things first: what even is demand sensing? Instead of relying on historical shipments as the sole input for planning, a true demand sensing solution comprises all the following criteria:
- Multiple, real-time signals to create daily forecasts reflecting current market realities
- Artificial intelligence (AI) and machine learning (ML) pattern recognition technology to process masses of big data and extract meaningful information
- A fully automated system with self-tuning algorithms that learn from data without human interaction and publish daily forecasts directly to the supply system for execution
- Encompasses the entire portfolio for all channels, not just certain categories or brands for specific customers
One key finding from the Forecasting and Inventory Benchmark Study is that, within the first two weeks a product is on sale, demand sensing creates a step-change in performance that continues to deliver over the life of the product, cutting error by 30%. In essence, demand sensing is an essential part of building a more agile and efficient supply chain to improve forecasting accuracy. It just makes sense!
“It’s fun to stay at the M-E-I-O”
Ok, maybe it’s not as fun as singing along to the YMCA, but there are still ‘many ways to have a good time’ with multi-echelon inventory optimization (MEIO). MEIO creates mathematically optimal stocking levels for each item, raw material or component at each node of the end-to-end supply chain to determine the lowest possible system-wide inventory. It’s the final answer to the question, ‘How do industry leaders make inventory work harder?’, resulting in efficiency improvements, reduced working capital and lower carrying costs. And the best part? It works extremely well and is complementary to demand sensing techniques.
MEIO using demand planning forecasts yielded a 17% reduction in safety stock compared with traditional single-echelon inventory management. However, the combination of MEIO and the more accurate demand sensing forecast almost doubled performance, cutting safety stock by 28%.
“While MEIO alone is effective, the combination of MEIO with demand sensing doubles the reduction in safety stock. The takeaway—if you are serious about improving inventory productivity, do both.”
We’re only scratching the surface
After all, this is only a blog post, right? While we could write endlessly on the various findings and methodologies available to help improve your planning and forecasting, you still have 32 pages of jam-packed data and statistics awaiting you! But just in case you’re not fully sold yet, here is what makes the 2019 Forecasting and Inventory Benchmark Study so unique:
- It is the largest study of its kind, encompassing more than $250 billion in annual sales
- All of the data is reported and aggregated the same way—a true apples-to-apples comparison year-over-year
- It utilizes operational data, rather than survey responses or anecdotes
- A wide variety of industries are included such as consumer packaged goods, oil and gas, food and beverage, chemicals and animal care
- It includes customer-facing inventory measures and supply-side insights on manufacturing and distribution performance
Everyone likes options though, and that’s why we’re providing two other resources to help you on your journey to planning excellence. Feel free to click on any of the links below or contact us directly to learn how E2open can help improve your planning performance with holistic, end-to-end capabilities.