We often see in our cooperation with companies that in response to any gap in a Key Performance Indicator (KPI) people immediately start the root cause analysis and ask the miraculous why questions. In our experience, since the KPI level is too complex, it is almost impossible to find the causes of the problem without narrowing down the problem itself – in other words, without slicing the elephant. Now, let’s have a look at a simple example to see why the deployment of the strategic metrics is important and how it is done at the level of specific operations.
Reading time: 4 minutes
Author: Szabolcs Molnár, president, senior lean coach
Example: Pizza delivery
Let’s say you are the owner of a fast-growing pizzeria chain, and you are opening restaurants one after the other. Your biggest competitive advantage is speed of delivery. Your motto is “30 minutes from ring to bell” – this is how you advertise to customers, and it means no more than 30 minutes can pass between the phone call and the delivery, otherwise the pizza is on the house.
One of your main KPIs is revenue. Every month, your accountant sends you the list of your pizzerias’ revenue of the previous month, and one month you notice that the revenue of restaurant #23 has decreased by 17% compared to the previous month. You call the manager to inform him of this fact and ask him what the possible reasons for the decrease might be. He says he is sorry but does not know the answer. Now the manager has two options:
One of his two options is to summon his employees, inform them about the dismal results from the previous month and ask everyone to do better next month. Better, whatever that means. Even though they feel they are working better, revenue continues to fall next month.
Narrow down the problem: KPI – PI – I
Another option for the manager is to try to understand the origin of the difference in revenues. He knows that revenue is the result of the performance of a myriad of business processes. The first step is to look at the main processes of the restaurant and what indicators are specific to its performance. They are called Process Indicators (PI). Your restaurants’ most important process is the fulfillment of customers’ orders, and it should run like clockwork as it ensures your market advantage. Once examining this, the manager realizes that the planned 30-minute “ring-to-bell” lead time is actually 40 minutes, so currently there is a 10-minute gap between the actual and the planned or expected status.
Since this process is the combination of the performance of additional workflows, your manager continues to narrow down the problem. He examines each of the building blocks of the process, i.e. the individual work contents of each block. Their required timeframe will be the Indicators (I).
The 30-minute process is planned to look like this:
- 5 minutes: taking the order, preparing the pizza
- 10 minutes: baking the pizza
- 5 minutes: boxing, printing the receipt, mounting the motorbike
- 10 minutes: delivery
In comparison, this is what he measures in reality:
- 5 minutes: taking the order, preparing the pizza
- 20 minutes(!): baking the pizza
- 5 minutes: boxing, printing the receipt, mounting the motorbike
- 10 minutes: delivery
A deeper analysis shows that the difference is generated at the level of the pizza baking operation, which affects your lead time, which in turn has a major impact on your revenue: in an increasing number of cases you deliver pizzas to the customers for free because of the delay. In addition, some of the customers are leaving because other market players can also deliver pizza within the 40-minute timeframe.
Time for the whys: the point of cause
It is worth starting to look for the causes at the level of the pizza baking operation where it is easy to find out why a pizza takes twice as long to bake (e.g. one of the heating coils in the oven has burnt out).
But if we had asked why at the level of the revenue (KPI), we would have been left with uninformed guesses as to the reasons for the decline.
An additional advantage of having employees who are aware of the expectations for each activity (e.g. baking time/pizza or number of pizzas baked/hour) is that the first slow-baked pizza can be an early indicator that something is wrong.
This way, we can intervene immediately when one of the two heating coils has burnt out as opposed to noticing the problem a month later when it has already had an impact on revenue (KPI).
You cannot solve problems at KPI level
Of course, connections are not that simple in everyday reality, but the moral of the story is that the more we can keep our finger on the pulse of our processes, i.e. know what is happening now and what should be happening, the sooner we can intervene if something does not go according to the plan. It is also much easier to start troubleshooting because you can be hot on its trail.
Remember: you cannot solve problems at KPI level! It is merely a “historical” indicator that shows the performance of some strategic pillar of your company in the rear-view mirror. Moreover, we can only look in this mirror 12 times a year (presuming reports are generated monthly) and decide which way to steer the boat next month. However, these decisions will be based on assumptions only because in the absence of PI and I-level indicators we will have no idea what caused a KPI-level indicator to deteriorate.
