The Best of Fence was a manufacturing business. It made fence components for a variety of standardised designs and sold these to building contractors across the country. Like most other industries, the building industry had been hit hard by interruptions to global supply lines. However, The Best of Fence made a point of getting everything locally, so they had been much less badly affected. That meant they now had plenty of eager customers. Business was good, but management wasn’t able to take advantage of that as much as they’d have liked to; customers were telling them their lead time was a problem. The Best of Fence tried to compete on responsiveness, rather than price, so it did a bit better on lead times than most of its competitors. Even so, it would still take four weeks between a job being received and it actually getting to the customer.
Reviewing the situation with their accountants, the accountants pointed out that a lot of money was tied up in stock, mostly parts and materials used in manufacturing the fence components. This confused the owners of The Best of Fence; they needed all that stock to ensure on-time delivery. Freeing up that money wasn’t simply a matter of running down its stocks; most of the parts and materials were actually in use – they were part of jobs sitting all through the factory in various states of completion.
So, how to improve the lead times and not put delivery at risk? A lot could happen in four weeks. Perhaps there was a spell of bad weather, or a customer’s build was falling behind schedule – if that happened, obviously the contractors couldn’t install the fence on schedule, and the order had to sit somewhere while The Best of Fence waited for the conditions to be right. Or maybe a rush order was received from an important customer, and it jumped to the head of the queue. When that happened, everything else got delayed. Despite using local suppliers, sometimes supplies got delayed, and this meant other jobs got started in the meantime – “We might as well make a start on the next order” – which meant that when the supplies arrived, the older jobs had to wait a bit longer while the new jobs got finished off, and the operators were free again to go back to the deferred jobs. And so on, and so on, and so on… when you considered all the things that could delay a job, four weeks was actually pretty good!
The only reason the system worked as well as it did was that the factory supervisors ran themselves ragged keeping everything moving. They were constantly checking which jobs were going to be late, how late they were going to be, and how upset the customer would be as a result, and shuffling around the job priorities to take that into account. This frustrated the operators – they often didn’t even bother looking at the printed schedules for jobs, because they knew the schedule would be changed soon anyway, so it just seemed pointless. The business was pretty chaotic a lot of the time, but it somehow managed to keep that four-week lead time fairly consistent.
What The Best of Fence could see was the piles of work-in-progress in the operations, and that jobs in operations spent a lot of time waiting, queued behind other jobs that were being priority processed. Just how long were jobs sitting not being worked on? Anecdotally, it seemed ‘a lot;’ a job could sit for days before advancing. Everyone had a pile of jobs in front of them.
Having slept on it, and argued about it over the course of a week, the management team at The Best of Fence decided to reduce the amount of work-in-progress that it had. Various ways of achieving that were suggested, including taking on temporary staff to increase their output for a time. In the end, though, they went with something simpler: “Let’s just stop sending jobs to the factory for a while.” Jobs were still being accepted by sales, they were just held back at release until the jobs in operations had cleared out a bit.
Simply stopping the release of jobs turned out to be harder than anyone had expected. The managers felt a lot of pressure to do ‘just one more,’ the supervisors tended to get all upset when it looked like the operators might have to stand around idle, and the declining volume of jobs in play spooked the operators. It took management some effort to convince the operators that it was expected that they might run out of work temporarily, that it was okay, and that instead of asking for a new job to be released, they should find where the backlog of jobs was being held up to and help out there if they could.
But, as they continued to flush work out of the system, the managers also started to notice that the remaining jobs were moving faster than expected. Having some empty space in the factory made it easier to move things around, and less damage was occurring.
Eventually, the number of jobs in the system had halved, and the continued decline threatened to starve the system entirely. The managers at The Best of Fence started releasing work into the system again. The first few jobs that got released also moved fast: because the queues in front of many of the operators were small and they could start on the new job as soon as it got to them; and because jobs were in the system for less time, they were also being interrupted less by any of random delays that could happen. That meant less resequencing by the managers, which allowed the operators to follow the printed schedules more easily. It was actually quite startling how fast things could move when the factory was nearly empty! It was also quite startling how much cash that had been tied up in parts and stock showed up in the bank; there had been a lot of money tied up in stock.
Eventually, though, the factory started to fill up again, and the rate of completions started to drop back down. Nobody had expected this – it had been assumed that once the system had been flushed, things would continue to move through it smoothly once jobs were being released again. Obviously, there was a trick to making it work right. The managers shut down the release of work once more. This time, though, they kept a much closer watch on how many jobs were in the system at the same time. Once the factory started to clear out, the managers purposefully started releasing a trickle of jobs rising, once work-in-progress had reached the agreed level, to match the output.
Reflecting on this, The Best of Fence noted that precisely how many jobs to release was a subject of some debate. Some people said they should release enough to keep everyone working, but when they tried that the clogging and queuing returned, even if it took longer to become bad. Eventually, they decided to release only enough to be sure their most constraining resource would always be busy. They had to accept that some other operators would be idle sometimes, but that was felt to be less of a problem than having the entire factory clogged up with work that couldn’t be finished!
This was where it all started coming together. With fewer jobs in the system, there was less money tied up in parts for those jobs. And there were other benefits, too: all jobs moved faster than they used to, and more consistently. The managers also had an easier time finding problem jobs and keeping everything moving.
Now that things were moving faster, it turned out that The Best of Fence could also reduce their quoted lead times for jobs. It used to be that jobs had to be released four weeks ahead in order to be ‘usually’ done on time, but now ‘usually’ was more like two weeks internally. Just reducing the amount of work-in-progress had allowed the business to cut its lead times in half, which was a competitive advantage that it would be hard for other businesses to match. This was a strong price point with customers and allowed The Best of Fence to price two different lead times into the market for a tidy profit.
The biggest management challenge now was dealing with what they called the ‘sneakies:’ jobs that somehow sneaked through into production. This usually happened because someone was trying to be helpful, or was doing a job for a friend, but the result was that the system would start clogging up again. Sneaking a job into the system didn’t get that job done faster, in fact, it slowed things down for everyone! But this was a different challenge, and compared to how things used to be it seemed easy to solve.
Reducing the amount of work-in-progress seemed at the start as though it would only be helpful to the accountants and would make delivery performance worse. In fact, it provided benefits right across the business. Now, The Best of Fence could had a competitive advantage and they fully planned to exploit the gain!
