Evidence trumps opinion every time – there is no substitute for an experiment

I have spent a lot of time looking at processes in the last few years and (blowing my own trumpet a bit) I have become pretty good at spotting where the throughput and productivity issues will be, and suggesting countermeasures to improve the situation.

As a scientist by training I understand the power of experiments, I use them to support my predictions and I’m not often surprised by the outcome – does anyone hear a pride before a fall moment coming?

Yip – major wake up time – my prediction was wrong!

The process under question was a simple packing and labelling for shipment. The products were aftermarket automotive spares which fitted into a range of 5 different sized boxes. Quantities per order line ranged from 1 to 20 pieces. Each order, typically of 8 lines, is picked from the warehouse and presented to the packers in a tray for them to label the part, insert the part into the correct box, label the box and insert it into a shipping outer container.

It was a nice, simple, process step but as packing was the bottleneck for the business in the busy summer months it needed some attention. Space was tight and the layout created a lot of transport and movement waste for the packer with bending from the waist and over shoulder height lifting, double handling etc. So clearly there was room for improvement.

The warehouse leader was convinced that two people on a packing station would deliver more than twice the throughput of one person and looking at the ergonomics of the layout I agreed with him. Good team work would eliminate some of the movement waste and some of the double handling.

So with the enthusiastic engagement of the team, we ran an experiment to compare productivity of 1 man vs 2 man teams. We made sure the starting conditions and finishing conditions would be the same and both packing lists were identical with a mixture of low volume and high volume parts. We ran the experiment fully expecting the 2 man team to finish in less than half the time. WRONG!

The 1 man team completed it in 36 minutes, the 2 man team did it in 24 mins – a win for throughput but a major failure for productivity. The 1 man team packed 40 items in 36 minutes or 0.9 minutes per item, the 2 man team did 40 items in 48 minutes or 1.2 minutes per item – 33% less productive.

I was surprised because I was sure that the 2 man team would work together and almost automatically eliminate waste – they didn’t. In fact, they got in each other’s way and slowed the flow down. What was even worse was that I was sure that the 2 man team would struggle with the small volume and really get ahead on the high volume parts of the packing list.

The complete opposite happened. With the small volumes, because they were forced into one piece flow, the 2 man team were almost balletic in their fluidity! As soon as they got to a batch of 20 it was very lumpy with boxes lying all over the place causing waiting time for one of the team.

So the team wanted to make some changes and run the experiment again.

I learnt a lot from this:

  1. No matter what you think you know, reality will surprise you – experiments beat opinions every time.
  2. Getting an unexpected result is just as good, if the team designs the experiment.
  3. The important thing is to keep making changes to the process in controlled experiments to test results.

As I write the team are on with trying new layouts and lots of experiments to measure the outcomes – and because they are going to use good experiments and have the team involved I know they will get to the best result – on their own.