Three Types of Improvement

There are three types of improvement. Additive, multiplicative, and exponential. In this post, I will briefly describe each of the three using an understaffed call center as an example.

When presented with a given challenge, most of us think additively. That is, we believe adding more resources to the problem will solve it. Imagine if you will, a call center that has more calls than it can handle. Customers wait for their call to be answered, get upset, and hang up. They try again later. Now, the call center reporting tool shows that two calls were trying to come through. To respond, the call center manager asks her staff to give more. "Does anyone want overtime?"

This might seem like an appropriate solution. It will eventually reduce the number of calls since there will be fewer abandons. The problem with this additive approach (adding extra people) is that it isn't sustainable. The cost of additional labor increases significantly, and the marginal benefit of adding human hours reduces overall productivity. It becomes more and more costly to have diminishing returns.

Now let's imagine our call center manager has more experience. She knows that overtime might be a short-term solution, but it will never solve the problem. She thinks bigger. Instead of adding resources, what if she multiplies them? Multiplier solutions are what we tend to think of as process improvements. Our manager might look at incoming call patterns to identify times to optimize her agents' schedules. Instead of adding agents, these optimizations improve multiple agents at once. Or, she may consider improving the process her agents use. Perhaps changing the call script allows her agents to reduce call times by 10 seconds. 10 seconds per call adds up. Multiplier improvements quickly solve problems, and make the most of current resources. Not bad.

But what if our call center manager is a bit of a strategist? She knows not only the resources she has and how to best use them, but can think strategically about the cause of her problem. Exponential improvements create a self-sustaining returns over time. They consistently get better with no maximum benefit (technically, most are asymptotic and approach an end of improvement at some point, but I digress).

One way to get these flywheel effects is to layer multiplier improvements so that they multiply each other. Improving call scripting, optimizing agent schedules, etc generate second-order effects. The shorter calls means more calls get handled. This leads to fewer abandons, and improves the forecast for the day. The improved forecast allows for better staffing. Better staffing leads to less stress. Less stress leads to shorter calls (fewer complaints) and better customer service. Better service leads to fewer calls. Fewer calls leads to...

People in crisis think additively. Most people have the capability to think muliplicatively. Very few people think strategically to simple exponential improvements.

But there is one more level. A master strategist gets to the root cause of their customers' concerns, complaints, longing and needs. There is one thing better than perfectly efficient calls...no calls. Imagine a company that takes such great care of its customers that they don't have to call customer service.