Executive Briefings

A Different Approach to Fixing Your Long Tail

Long tails can get, well, rather long. Some see this as a mandate to hunt down and kill off the items furthest out on the tail.

A Different Approach to Fixing Your Long Tail

Gartner analyst Stan Aronow speaking at his firm's Supply Chain Executive Conference in London presented something along these lines last September. He advocated regularly segmenting SKUs into four quadrants on a matrix depending on inventory turns and gross margin, then either "keeping", "curing" or "killing" items, depending on where they sit.

But this approach needs to evolve for today’s economy.  For many companies, if not most, long tails are good. Yes, good. That’s because many companies often don’t make their highest gross margins off their fastest movers. High volume products are often the most commodity-like and competition drives down margins.  Think of the “loss leader” at a retailer. Or how much a parts vendor gets for a part they sell to the original equipment manufacturer (OEM) versus how much they sell that same part for in the aftermarket. The television program “60 Minutes” once calculated that at aftermarket prices it would cost $125,000 to build a $20,000 car out of spare parts.

New products, niche products, ancillary products, seasonal products, specialty products (think “green beer”) all have the same story: Much lower volumes, but usually higher margins.  And they may also be just the thing you need to keep your customers happy. Whether it’s an urgently needed spare part or just the right specialty pasta, the breadth of your product line may be an important part of what keeps customers coming back.  Expanding the points of sale that your customers can choose from and the long tail (measured in SKU-Locations) can grow by multiples.

The problem is that these long-tail items can be difficult to plan and stock. And that’s where a different approach comes in.  New, highly automated, and smarter inventory optimization techniques are making the long tail a lot easier and a lot more affordable to manage. So vendors and retailers can both have a long tail and lower cost-to-serve.

Case study: How Dart Container cured its long-tail

Dart Container Corporation, the well-known manufacturer of beverage cups and containers, offers a great example of how to achieve these twin aims. Dart achieved 99.6 percent service levels with 20 to 30 percent less inventory without killing any long-tail items.  Here is what happened. In 2012 Dart Container acquired Solo Cup, famous for its disposable red party cup. Solo also sells a broad assortment of other products through retail and food service channels, B2C channels and B2B channels like Staples.

Those iconic red cups hardly conjure up visions of a long-tail scenario –  quite the opposite. However, the Solo acquisition also expanded Dart’s range of channels and SKUs. Now Dart had both high-volume items across few SKUs, as well as a long tail comprising about 11,000 made-to-forecast SKUs. Some 30 percent of these had less than one case per day of demand.

Customization added to the SKU proliferation. Solo’s customers can choose products made from a range of materials and with different finishes to meet various price and environmental concerns. They may also order small quantities of more expensive items or vice versa. Customers can even order cups with their own branding (such as a retail chain with its own design), custom packaging for better shelf fit, and so forth.

Applying new approaches

Dart wanted to provide high service levels for these items in order to grow the product profile and market penetration. But a focus on top-down forecast accuracy wasn’t as effective with low-volume items and high complexity chains – S&OP teams are less interested in those.

Rather than investing time trying to improve forecast accuracy, Dart’s approach “threw more science at the problem” and implemented a multi-echelon inventory optimization tool to address the long tail and deal with the integration challenges of combining the two firms. Dart had been applying traditional ABC analysis (Pareto’s close cousin) to estimate inventory importance, assigning safety stock values based on volume.  Instead it used a statistical inventory planning approach that takes into account demand and supply variability, as well as replenishment parameters of the items, to set safety stock targets.

Cured long tail – healthier outcomes

The new approach worked. The new safety stock combined with replenishment cycle stock gave enough inventory to fill the long tail orders. Many of those items would have been killed off if Dart had been operating off the old targets. With the new system in place, Dart was able to optimize across all levels within the network to come up with the most efficient inventory mix to hit the overall corporate 99.6 percent fill rate.

Dart was also able to use their buffer distribution denters (DCs) more strategically than in the past. Previously buffer DCs simply held excess production overflow. Under the new system, Dart strategically set buffer DCs to reduce both the downstream inventory requirements as well as the total inventory requirements within the network. So instead of just using its DCs to store excess inventory, Dart is now using them for “risk pooling”, a statistical approach where inventory held in an upstream warehouse services demand aggregated across multiple channels. Using this model it becomes more likely that high demand from one channel will be offset by low demand from another. The resulting reduction in demand variability allows for a decrease in safety stock and average inventory.

When Dart applied the new statistical safety stock targets to the legacy Solo network, it hit an estimated 99.6 percent service level with 20 to 30 percent less inventory.  Dart is now raising the bar again by upping distribution points for its full product range from 9 to 21. Until recently the Dart and Solo distribution chains were not completely consolidated. The original Dart and Solo DCs carried mostly their own products – often requiring customers to place Dart and Solo orders independently. Now Dart and Solo products are being carried at the same locations, enabling single point ordering. This almost doubled the SKU-Locations to about 20,000, so the tail grew even longer.

Long-tail vs ‘traditional’ planning

Let’s face it. Consumer demand for variety, shortened product life spans and omnichannel sales all add up to long-tail demand patterns. The 80/20 rule, where 80 percent of the revenue is generated by 20 percent of the products, no longer holds. Pareto is dead. One of our internet retail customer’s says its long tail accounts for 79 percent of its revenue. Supply chain analyst Lora Cecere reports that fast-moving consumer goods manufacturers are now earning about 50 percent of revenue from the tail. Even food and beverage companies can have 36 percent of their revenue from the tail. 

As the long tail grows and fragments in response to demand volatility and complexity, planners see more periods of zero or intermittent demand and spiking “lumpy” demand. Traditional models that worked fine for high product volumes and predictable demand don’t do well in this environment. Many actually deem long-tail items with very low demand rates to be “unforecastable”. This is why lots of companies today are either augmenting or replacing them with more sophisticated tools that come with demand sensing, machine learning and other advanced capabilities.

Mastering the long tail calls for new approaches that can manage the complexity inherent in variable demand and high service levels.  These include:

Demand modeling:  Understanding unique demand distributions of each SKU-Location at the order line level. Traditional “single number” forecasts and forecast accuracy metrics aren’t very meaningful in the long-tail environment.

• Service level planning:  Inventory optimization must incorporate the understanding of demand distributions and a model of the supply chain configuration and its variability.

• Execution to plan:  Replenishment plans must incorporate an understanding of high demand volatility to provide degrees of freedom for replenishment, along with intelligent and exception-driven alerts.  Traditional single number DRP approaches won’t work in this environment.

Avoiding the path of least resistance

Future supply chains will grow more complex. Long tails will grow longer. If a company with a seemingly simple product assortment like Dart has a long tail problem, chances are yours does, too. You can tackle this problem. The tools are available, the processes are tried and tested and there is a growing army out there gaining practical experience curing long tail problems.

Don’t take the path of least resistance and cut off your long tail. Optimize it.

Source: ToolsGroup

Gartner analyst Stan Aronow speaking at his firm's Supply Chain Executive Conference in London presented something along these lines last September. He advocated regularly segmenting SKUs into four quadrants on a matrix depending on inventory turns and gross margin, then either "keeping", "curing" or "killing" items, depending on where they sit.

But this approach needs to evolve for today’s economy.  For many companies, if not most, long tails are good. Yes, good. That’s because many companies often don’t make their highest gross margins off their fastest movers. High volume products are often the most commodity-like and competition drives down margins.  Think of the “loss leader” at a retailer. Or how much a parts vendor gets for a part they sell to the original equipment manufacturer (OEM) versus how much they sell that same part for in the aftermarket. The television program “60 Minutes” once calculated that at aftermarket prices it would cost $125,000 to build a $20,000 car out of spare parts.

New products, niche products, ancillary products, seasonal products, specialty products (think “green beer”) all have the same story: Much lower volumes, but usually higher margins.  And they may also be just the thing you need to keep your customers happy. Whether it’s an urgently needed spare part or just the right specialty pasta, the breadth of your product line may be an important part of what keeps customers coming back.  Expanding the points of sale that your customers can choose from and the long tail (measured in SKU-Locations) can grow by multiples.

The problem is that these long-tail items can be difficult to plan and stock. And that’s where a different approach comes in.  New, highly automated, and smarter inventory optimization techniques are making the long tail a lot easier and a lot more affordable to manage. So vendors and retailers can both have a long tail and lower cost-to-serve.

Case study: How Dart Container cured its long-tail

Dart Container Corporation, the well-known manufacturer of beverage cups and containers, offers a great example of how to achieve these twin aims. Dart achieved 99.6 percent service levels with 20 to 30 percent less inventory without killing any long-tail items.  Here is what happened. In 2012 Dart Container acquired Solo Cup, famous for its disposable red party cup. Solo also sells a broad assortment of other products through retail and food service channels, B2C channels and B2B channels like Staples.

Those iconic red cups hardly conjure up visions of a long-tail scenario –  quite the opposite. However, the Solo acquisition also expanded Dart’s range of channels and SKUs. Now Dart had both high-volume items across few SKUs, as well as a long tail comprising about 11,000 made-to-forecast SKUs. Some 30 percent of these had less than one case per day of demand.

Customization added to the SKU proliferation. Solo’s customers can choose products made from a range of materials and with different finishes to meet various price and environmental concerns. They may also order small quantities of more expensive items or vice versa. Customers can even order cups with their own branding (such as a retail chain with its own design), custom packaging for better shelf fit, and so forth.

Applying new approaches

Dart wanted to provide high service levels for these items in order to grow the product profile and market penetration. But a focus on top-down forecast accuracy wasn’t as effective with low-volume items and high complexity chains – S&OP teams are less interested in those.

Rather than investing time trying to improve forecast accuracy, Dart’s approach “threw more science at the problem” and implemented a multi-echelon inventory optimization tool to address the long tail and deal with the integration challenges of combining the two firms. Dart had been applying traditional ABC analysis (Pareto’s close cousin) to estimate inventory importance, assigning safety stock values based on volume.  Instead it used a statistical inventory planning approach that takes into account demand and supply variability, as well as replenishment parameters of the items, to set safety stock targets.

Cured long tail – healthier outcomes

The new approach worked. The new safety stock combined with replenishment cycle stock gave enough inventory to fill the long tail orders. Many of those items would have been killed off if Dart had been operating off the old targets. With the new system in place, Dart was able to optimize across all levels within the network to come up with the most efficient inventory mix to hit the overall corporate 99.6 percent fill rate.

Dart was also able to use their buffer distribution denters (DCs) more strategically than in the past. Previously buffer DCs simply held excess production overflow. Under the new system, Dart strategically set buffer DCs to reduce both the downstream inventory requirements as well as the total inventory requirements within the network. So instead of just using its DCs to store excess inventory, Dart is now using them for “risk pooling”, a statistical approach where inventory held in an upstream warehouse services demand aggregated across multiple channels. Using this model it becomes more likely that high demand from one channel will be offset by low demand from another. The resulting reduction in demand variability allows for a decrease in safety stock and average inventory.

When Dart applied the new statistical safety stock targets to the legacy Solo network, it hit an estimated 99.6 percent service level with 20 to 30 percent less inventory.  Dart is now raising the bar again by upping distribution points for its full product range from 9 to 21. Until recently the Dart and Solo distribution chains were not completely consolidated. The original Dart and Solo DCs carried mostly their own products – often requiring customers to place Dart and Solo orders independently. Now Dart and Solo products are being carried at the same locations, enabling single point ordering. This almost doubled the SKU-Locations to about 20,000, so the tail grew even longer.

Long-tail vs ‘traditional’ planning

Let’s face it. Consumer demand for variety, shortened product life spans and omnichannel sales all add up to long-tail demand patterns. The 80/20 rule, where 80 percent of the revenue is generated by 20 percent of the products, no longer holds. Pareto is dead. One of our internet retail customer’s says its long tail accounts for 79 percent of its revenue. Supply chain analyst Lora Cecere reports that fast-moving consumer goods manufacturers are now earning about 50 percent of revenue from the tail. Even food and beverage companies can have 36 percent of their revenue from the tail. 

As the long tail grows and fragments in response to demand volatility and complexity, planners see more periods of zero or intermittent demand and spiking “lumpy” demand. Traditional models that worked fine for high product volumes and predictable demand don’t do well in this environment. Many actually deem long-tail items with very low demand rates to be “unforecastable”. This is why lots of companies today are either augmenting or replacing them with more sophisticated tools that come with demand sensing, machine learning and other advanced capabilities.

Mastering the long tail calls for new approaches that can manage the complexity inherent in variable demand and high service levels.  These include:

Demand modeling:  Understanding unique demand distributions of each SKU-Location at the order line level. Traditional “single number” forecasts and forecast accuracy metrics aren’t very meaningful in the long-tail environment.

• Service level planning:  Inventory optimization must incorporate the understanding of demand distributions and a model of the supply chain configuration and its variability.

• Execution to plan:  Replenishment plans must incorporate an understanding of high demand volatility to provide degrees of freedom for replenishment, along with intelligent and exception-driven alerts.  Traditional single number DRP approaches won’t work in this environment.

Avoiding the path of least resistance

Future supply chains will grow more complex. Long tails will grow longer. If a company with a seemingly simple product assortment like Dart has a long tail problem, chances are yours does, too. You can tackle this problem. The tools are available, the processes are tried and tested and there is a growing army out there gaining practical experience curing long tail problems.

Don’t take the path of least resistance and cut off your long tail. Optimize it.

Source: ToolsGroup

A Different Approach to Fixing Your Long Tail