What's Wrong with a 95% Service Level?

MDM has syndicated MIT senior lecturer Jonathan Byrnes' blog focused on profitability management. Byrnes has worked with several businesses across industries, including distribution. MDM published a two-part interview with Byrnes in October 2010. He shares best practices based on his experience here.

Customer service level is a core measure of company performance. This measure seems completely obvious: after all, if you increase your line-fill or case-fill from 94 percent to 96 percent, your customers will surely be much more satisfied. And satisfied customers will buy more from you, increasing your profits. Right?

This seemingly simple logic has critical flaws that are very important for your company. Is your objective simply to maximize average customer satisfaction? Or is your objective to maximize your company’s profitability and growth? These are not necessarily, or even often, linked.

Think about the following questions, in the context of a company with a 95 percent service level.

Which customers get the bad 5 percent of your service measure? How would you feel if your best customers – the customers in your company’s “sweet spot” (those that produce the most profit and have the highest growth potential) – received unreliable service 10 percent of the time, while your marginal customers enjoyed 99 percent service? How would you feel if both your best customers and your marginal accounts received the same poor service 5 percent of the time – after all, this is what the measure indicates.

How bad is the 5 percent deficiency? Is it hitting key customers running a VMI or cross-dock system with one week delays, or is it hitting customers carrying ample inventories with an overnight delay? Is it hitting critical products or commodity-like products with many substitutes?

What price are you paying for making promises you can’t keep? If you always made realistic customer service (order cycle) promises that you could keep virtually 100 percent of the time, would your customers trust you with a slightly longer order cycle, or are they simply insisting on very tight cycles to give themselves “breathing room” for your service deficiencies?

When you focus on aggregate measures of customer service, in reality you are maximizing what’s easiest to measure, not what gives you the most profitability and lucrative growth. This is another artifact of the Age of Mass Markets, when companies distributed as widely as possible, customers had plenty of inventory, and computers were in their infancy.

Service differentiation is a much more effective way to frame and measure customer service. In a nutshell, you should make different order cycle promises to different customers depending on your customer relationship and the nature of the product.

Consider a simple 2×2 matrix, with core and non-core customers, and core and non-core products. A core customer is a significant steady customer, often one whose supply chain is integrated with yours. A non-core customer is a smaller, occasional customer, or even a large, high-potential customer that uses your competitor as its primary supplier. A core product is either a high-volume product or a critical product with no substitutes. A non-core product is a slow-moving product, not critical, and often with ample substitutes.

If you think about each quadrant of the matrix, it becomes clear that each quadrant’s customer/product characteristics logically suggest a different order cycle.

The core-core quadrant requires fast, completely-reliable service, and here a 95 percent service level is particularly deadly. For non-core customers ordering core products, you might offer guaranteed 2-3 day service, but always keep your promises. This will allow you to produce 100 percent reliable fast service for your core customers (from local stock), while gaining the leeway to source product from centralized stock to meet the occasional spikes in demand that non-core customers sometimes produce with occasional large orders. The alternative is to carry a huge amount of costly safety stock in the effort to treat all customers the same.

There are important exceptions. If a non-core customer is a high-potential account that a sales rep is working intensively, it can be bumped into the core category, as long as you reexamine this after a period to see whether the relationship has changed.

The service differentiation logic is similar for your non-core products. By definition, non-core products are those not needed urgently. When a core customer orders a non-core product, most of the time it can be sourced from local stock, with an order cycle of perhaps 1-2 days.

The big cost drain occurs when non-core customers order non-core products. Here, you will need a massive amount of local safety stock to meet an aggregate measure like 95 percent service level (with short cycle time). The correct course is to make an order cycle promise of perhaps 3-4 days, so you can source the product from a centralized pooled inventory.

Three important customer service principles emerge: (1) you should match your order cycle promise to the nature of the customer relationship and the product characteristics; (2) you should make different promises to different segments of customers for different segments of products; and (3) most importantly, you should always keep your promises.

The right measure of customer service is not aggregate line-fill and case-fill: it is always keeping your promises. Service differentiation will enable you to maximize your profitability and to develop high sustainable profitable growth. The classic dilemma of trading off between cost and service is an obsolete, misleading concept. It assumes that all customers and products are the same.

You can have your cake and eat it too if you have clarity, focus, and follow-through – and you let go of the tacit goal of trying to be everything to everyone.

Think about another common measure of customer satisfaction: net promoter score. This is a commonly-used measure that compares customers who would recommend you against customers who would not. The problem is that this is another aggregate measure with all the flaws of an overall measure of customer service level.  (Aggregate surveyed customer satisfaction has the same problem.) Clearly, the customers in your “sweet spot” should have net promoter scores off the charts, or your islands of profit will be sinking beneath the waves.

What about the rest of your customers? If you implement service differentiation effectively, they should also have a very high net promoter score. They will have a clear understanding of their relationship with you, they will trust you to always keep your promises, and they will know exactly what they have to do to change their relationship and therefore their order cycle.

In business, the worst news is not knowing what to expect.

Jonathan Byrnes is a senior lecturer at MIT and author of the recent book, Islands of Profit in a Sea of Red Ink. He is president of Jonathan Byrnes & Co., a consulting company with which he has advised over 50 major companies, medical institutions and industry associations. Contact him at jlbyrnes@mit.edu.