Myth Busting: KPIs Are Always Helpful
Many believe that if you can’t measure it, you can’t manage it. That’s just not right. Measurements can only count what is countable – dollars, production numbers, headcounts, timeliness, etc. They can’t count the achievement of objectives unless those objectives are purely numeric in nature.
For instance, you can count attendance or ticket sales for an event. It might appear to be a huge success by that measure, but is it a success if those attending didn’t enjoy the event? We would very likely consider our high schools successful because of diminishing dropout rates, but that just means people attend. Do they learn anything useful? Are they employed after leaving? Do they go on to higher education?
In addition to being misleading, measures can also get us into trouble. In one delivery company they measured package sorting machine downtime as an indicator of maintenance effectiveness. If downtime exceeded a set limit (30 minutes in that case), then the incident was reported to operations and the maintenance manager was held to account for it and ended up somewhat publicly embarrassed over the incident. The maintenance manager had great incentive to minimize downtime. What did he do?
Rather than focus on reliability (or uptime), he focused on the downtime. How can he minimize that? Keep spare parts for every conceivable job beside every machine. This had the effect of ballooning the size of spare parts inventory at each site. There were over 20 sites! Millions of dollars were sitting on shelves in cabinets beside machines to minimize repair time. It worked, too – downtime went down. But was it necessary?
Those machines actually ran for only 4 to 5 hours a day. They were idle for the rest of the time. With many redundant machines in each facility it was no big deal to route packages through other lines if one went down. Total sorting time still remained low. Repairs were being rushed and excessive sparing resulted from this single well-intentioned but misguided measure.
Metrics need to be managed carefully. They require balance. The effect of the downtime measure could have been buffered by another, equally important measure of inventory turns, also managed by the same manager. It would have forced the managers to think of how to achieve a balance. Of course that would also force the organization to manage across traditional functional boundaries – maintenance and stores.
Stepping back from the immediate measure, they might also have realized that downtime was actually unimportant to them. They had loads of spare capacity.
Production is often driven by monthly, weekly, daily and even shift quotas. They must produce to a target to meet sales objectives. There is a lot of pressure to keep machines running, even if they are not running well or they are due for needed servicing. Skip the servicing and they risk having the machine fail with an attendant loss of production. Maintenance arrives to do their servicing and production refuses to allow them access. A few days later the machine is down and production is pressuring maintenance to repair it quickly so they can meet their targets. The production targets, and likely a lack of flexibility in how long they can take to recover losses, are behind the run-at-all-costs mentality that invariably gets them into trouble.
In lean manufacturing environments, they know to “slow down to go faster”. Slowing production to sustainable levels means fewer breakdowns, less downtime, increased run times and less variability. That all translates into going faster (i.e., producing more). They are encouraged to take small hits to production to correct minor defects because they know it means running more and better. They allow time for servicing and machines don’t end up over-stressed and broken. The quotas exist abut there’s flexibility in them to allow for minor ups and downs.
In those first two cases, KPIs actually made things worse. And this is seen time and time again. Poorly thought-out measurement schemes, often limited by artificial departmental or functional boundaries, push behavior in the direction of optimizing a score while sub-optimizing overall performance and/or costs and/or risks. In the lean case, the KPIs exist, but with more balance and flexibility. They look at the bigger picture and don’t put on departmental blinders.
We are obsessed with measuring yet we seldom step back, look at the big picture and determine what is best for the business – ultimately what is best for our customers. It means we all participate in delivering what the customer pays for and we all contribute to how effectively and efficiently we do it.
KPIs are indeed useful, they are a powerful tool, and they can be easily misused. Sure, use KPIs, but be careful and really think things through as a cross functional group before you apply them.