KPIs are not adding value.

In my organization (service industry), There are three KPIs :
OTD (On Time Delivery) Target 90%
FPY (First Pass Yield) Target 95%
EV (Effort Variance). Target +-20%

For the last 2-3 years we are constantly over-achieving Target and never missed a KPI target.
However, on-ground customers do not agree with the quality levels. The employees are devoting a lot of time to fulfilling the KPI targets as well.

How can we reduce overburden on employees and do we need to think about new KPIs. Basically what needs to be done to match reality with numbers?
Do you see such scenarios in other places as well?

8 Replies
Hi Deepak;
My first impression is that your targets are too low. If you constantly achieve them, shouldn't you increase them? This would be in line with continuous improvement. In my organization our target was 95 % OTD and when we reached that, we increased it to 96%. If we achieve that consistently this year, we will increase to 97% next year.
Hi Deepak,
It may be worth making a review and revision of your KPIs part of your annual review cycle. In my factory we review performance to targets each December and set new KPI targets. If we hit the target some of the time but not all of the time, we tend to leave it the same, but if we consistently surpassed the target (good or bad) then we revise it to either make it more reasonable. For example, if OTD was 90% and we were only achieving 80%, we might adjust the target down to 85% to make it less out-of-reach for the coming year... on the flip side if we were constantly hitting 95%, we might make it 97% to challenge the team to do better. Of course, the voice of the customer is also important in considering your Quality targets.

This has really helped highlight good weeks/months/quarters versus bad ones, as when our goals were unreasonable in the past, failing by a meter and failing by a mile seemed the same. Constant failure to perform can lead to a "well what's one more failure" mentality that's not healthy. We also had goals set at times (typically by outside sources) where a cycle time target was less than say 8 hours and our process for whatever reason was finishing in perhaps 2 hours, so the urgency to engage in continuous improvement evaporated.

Best of luck!
There's another possibility you might check out. Are people gaming the KPI's? I mean, are they taking actions which were not foreseen by the person who set up the formula, which result in the numbers always looking good even if performance in the real world is not so good? Maybe that's not the problem, but it could be worth checking what data you collect and how you calculate the KPIs, just to make sure.
John Elwer
4 Posts
First, if you consistently meet your goals, you can tighten them up. OTD at 95%, FPY at 98%, etc.

Another consideration is the use of continuous data vs discrete data. On time delivery is not continuous data, but time between shipment errors is and can give you more usable info.

All goals/objectives are only useful if the company takes action on goals not met to improve the process. Otherwise it is a waste of time to collect the data or an empty achievement for brown-nosing.

Reading about the Lean/Six Sigma can give more insight on what to measure, why, and how. I think is a great site.
Meta Brown
8 Posts
The most significant thing you've said here is " on-ground customers do not agree with the quality levels." It's the customer's perception of quality that matters, so focus on that.

What are you doing to understand the your customers' expectations and concerns? Do you have a system for tracking complaints? What kind of customer research are you doing? How do you know that the customers are not satisfied with the quality of your offerings - what are they telling you?

Just looking at your KPIs, I have concerns. A 90% target for OTD implies that it's OK with your management if one in 10 deliveries is late. That's a lot of late deliveries. What are the consequences for your customers if a delivery is late?

A 95% FPY - I don't know what you produce, but it sounds like there could be a lot of waste in that process. And you may not be catching all the problems - if defective product slips through to the customer, it might be expensive for them.

It would help to know more about what your organization is doing to manage the quality of your processes and products. For example, do you use statistical process control? Experimental methods? Checklists?

Duke Okes
168 Posts
Agree with what others have said ... 90% OTD would not be considered a good result, and it appears your "on the ground" customers have indicated so. Wonder what the cost is of having only a 95% FPY?

Also, these are outcome indicators which can't be managed directly. Do you have leading/controls indicators at key points within the processes that allow detection and response to variances that contribute to less than 100% results? Do you use Paretos to identify major contributors and work on them on a project-by-project basis?
If your metrics are not matching customer perception, you've got the wrong metrics. These are internally focused.
Your customer doesn't care about your FPY or EV...unless it translates to lower prices for them.
Einstein said "Not everything that can be counted counts, and not everything that counts can be counted."
Metrics need to be customer focused and targets should be set based on the customer/market needs (not what you think you can achieve).
For instance if you ran a fast food restaurant, you would measure customer wait time.
If your current time is 15 minutes and your nearest competitor is 5, you're going to lose business. Your goal should be 5 minutes or less to stay in business.
Goals should also be set for the system, not individuals. See some of Deming's books.
Lots of great feedback in the comments - FPY and EV are internal metrics that the customer generally doesn't see. When you say "do not agree with quality levels", you certainly need to dig to understand what that means and what the expectations are from the customers (and not just one customer's view). If you don't understand that, you can't close reality to expectation. Another watch item is "averages" - take a look at the data to see how much variation you.