Last week, I interviewed Robert Byrne, founder of Terra Technology on his demand planning benchmarking study. I enjoy creating the Straight Talk podcast series, and Rob’s findings in his benchmark study are always thought-provoking. This is his fifth year of studying demand processes. In this blog, I share the findings and why I think they should matter to supply chain professionals in consumer-facing industries.
Based on Rob’s study, the number of items in the Consumer Products Industry has proliferated by 30% over the course of the last six years. In the same time frame, shipments rose 2% while the average sales per item fell 22%. As a result, managing the supply chain is more complex. Supply chain leaders may not wake up in the morning and exclaim, “I have a longer tail!” But, in fact, over the last decade the widening of the product portfolio has resulted in fewer shipments per item and lumpier demand. Today, 50% of items only contribute 1% of sales.
What is a long tail? In Figure 1, I share a representation of a company’s “long tail.” I think that it is useful for companies to plot the length of their tail every two years to understand if they are making progress on managing products from cradle to grave. Most are not…. The tail is growing and as a result there are hidden costs in the business due to complexity. Despite all of the focused attempts to implement new forms of demand planning, the business costs and inventory spiral out of control due to unchecked complexity.
Figure 1. Calculating the Long Tail
With the lengthening of the product tail (volume of shipments per order and the frequency of orders for each product), the demand signal is lumpier and more erratic. While traditional companies are still looking at improving traditional forecasting technologies, and improving the rules for forecast consumption, I think it is time to take a different stance:
- Item Rationalization. The first step is item rationalization. Demand error and bias is higher for slower-moving products. As the product portfolio has increased, companies have more slower-moving products. One of the options is to rationalize the product line and manage products from cradle to grave. While most companies focus on only the cradle—or the birth of the product—rationalization is growing in importance. The problem has multiple dimensions. Most leaders do not know the cost of complexity with the increase of items; and instead of introducing variety to the shopper, many companies have confused the shopper resulting in the purchase of more compelling, less complex product portfolios. So, as companies manage their stage-gate processes for item introductions, it is a good time to also design and implement processes to focus on item rationalization by market. It is time to manage product lines from cradle to grave. The absence of these process has driven the supply chain out of balance resulting in increased inventory and unnecessary cost that does not drive brand differentiation.
- Test and Learn. To aid item rationalization, the use of test-and-learn concepts, in combination with cognitive reasoning, is useful to test the market basket impacts of portfolio shifts per market. While the traditional marketing and product development programs have been about the additional of item, the consumer products company that is in step with the market understands it is about having the right assortment in the right markets. This requires analytics that can learn and adapt as markets shift..
- Manage New Item Bias. In Rob’s study new item bias was 22%. With new products representing 36% of the product line, this is growing problem. In new product launch, companies are overconfident and have a positive bias. This is 4X higher than other products in the line. Traditional approaches like consensus forecasting can make this bias even worse.
- Need for Demand Sensing. As the tail of the supply chain lengthens, demand sensing matters more than ever. The longer the tail, the more important short-term forecasting; and, the less effective traditional forecasting techniques are in driving replenishment. The redefinition of inside-out to outside-in processes grows in importance. If you are scratching your head on what this means, it is the use of point-of-sale data and demand-sensing analytics to reduce the latency of the demand signal to plan and be better in-step with the market. Traditional order management to distribution requirements planning (DRP) adds latency and distorts the signal. This grows more critical with the longer tail. Demand latency for one company that I worked with last week was nine days for fast-moving products and 60 days for products in the tail. The longer the tail, the greater the need for demand sensing and the use of channel data.
- Translate Independent Demand into Production Planning. In addition to demand sensing, with the lengthening of the product tail in advanced companies, the independent item demand is translated into a segmented production plan to reduce cycle stock. This happens through integration of demand sensing data into the product rhythm wheel in production planning. Automation of demand sensing into production planning is critical to reduce cycle stock and not short-circuit production. As the tail becomes longer, the automation of production planning becomes more critical. This work into product segmented production planning is critical to manage the lumpier, less predictable demand patterns.
2015 is a year for belt-tightening. Volume has slowed, and most companies are struggling with new product launch processes. Perhaps the place to start is with product rationalization and the identification of product opportunities that stem from the consolidation of product lines. I think R&D and product development cannot just be about launch. Instead, I prefer a cradle-to-grave focus. If the product lines cannot be reduced in complexity, I think it is time to redesign supply chain processes. Unfortunately, today, most companies are not working on either. The concepts of demand latency and the impacts of the longer tail are not very well-understood.
Data sourced from the Terra Technology 2014 Benchmarking Study