Spike Detection and Levelling Application

Nigel found Sarah shuffling through an improbably large stack of papers.

“Sarah,” he said firmly, “We need to do something about the late jobs.”

“I know,” Sarah growled in frustration. “It seems like I spend all my time reworking schedules, and everything is still late! This morning I had to tell the sales team to stop selling for a week, just so my people can catch up a little! That was not a fun conversation to have, believe me.”

Nigel nodded sympathetically. “I’m sure it wasn’t. But… this might be more serious than the sales team having to work on their suntans for a week. It’s playing havoc with our financial situation, too. We have to pay to expedite work, we have to pay penalties when jobs are late, and our sales margins are falling. Customers are demanding discounts because they think we are unreliable! This is killing us, Sarah!”

Sarah shoved the stack of papers to one side and turned to face Nigel. “I know, Nigel. Really and truly, I know. That’s what all these papers are about. We don’t have a smooth flow of work, alright? Some weeks are busier than others, sometimes the sales team has good days and bad days, and so on. So, I look at all the orders that are coming up, and if I see a big one that’s going to be due soon, I schedule it to start as soon as possible. Too many times I have seen big jobs like this, which look good on paper, wreck everything as we muscle them through operations. We can’t go on making these trails of carnage in our business!”

Nigel frowned. “Yes, that’s common sense. Why do we end up with late jobs, then?”

Sarah sighed. “When there’s a big job coming, we schedule it to start as soon as possible. But that means other jobs, due at the same time, get scheduled to start after the big job… and sometimes, those jobs that get started later have due dates before the big job. But we can’t start on them before the big one is finished, so they just get delayed unless we pay extra to expedite something.”

 “Okay, but you’ve got people standing idle at times. Can’t they just start on these big jobs, no matter what the schedule is doing, so that we don’t have everything being late all of the time?”

Sarah frowned at Nigel. “Do I have to tell you about constraint scheduling again?” she asked.

“Actually… yeah, maybe you do. Or at least tell me why it’s relevant to this problem.”

“Well, you know the glass ovens, the ones that always have a backlog to clear. They are our constraint and we are scheduling the jobs they work on, so that they don’t get overwhelmed, and they need to produce jobs in the right sequence or we can never deliver on time. That’s the theory, anyway.” Sarah explained.

Nigel shook his head, confused. “So, what does that have to do with getting anyone idle to start on these big jobs? If they’re idle, they’re not the constraint, and they may as well be doing something useful.”

“Because the job doesn’t stay with them! Sooner or later, these ‘non-constraints’ finish what they can do, and then the constraint operation gets a huge job dumped on it out of nowhere!” Sarah said. “And that blows the schedule to tiny bits. Other jobs which are due in the oven before it aren’t ready because the big job having been pushed in front of them. So many times the oven operators have complained that they can’t clear the ‘must process this today’ jobs because the big one gets priority. It damn well overloads the constraint, and you know this. As if that wasn’t enough there’s no way the big job can get done on time anyway, because extra time for it wasn’t allocated in the schedule!”

Nigel opened his mouth to object, then closed it again. The situation Sarah had just described clearly wasn’t an improvement on the current situation, and Sarah was equally clearly doing her best to keep things running. “So… are we stuck with it, then?” he asked eventually.

“No, not at all!” Sarah was starting to get a bit sarcastic. “If you can just get sales to let me know about orders six months in advance, and with due dates at least three months out, and then not change anything, I reckon I can sort out a schedule that can get everything done on time. Otherwise, yeah, we’re stuck with it.”

“And there’s really nothing that would help with the scheduling issues?” Nigel persisted.

“Not unless you’ve got a couple of big buckets full of time you can pour all over this dumpster fire,” Sarah said. “If you don’t, then –“

“Hang on a second. That might actually be the answer…” Nigel said slowly.

Now it was Sarah’s turn to be confused. “What do you mean?” she asked.

“Buckets!” Nigel said excitedly. “Big buckets of time! That’s what we’ll use!”

Sarah shook her head. “Nigel, I have no idea what you’re saying. Can you break it down for me, please?”

“Even better! We’ll do that too, and put the resulting smaller jobs, the ’chunks’ in the buckets! Sarah, you’re a genius! We know how much the ovens can process in a day, we just need to make sure the jobs are chunked to fit in these one-day buckets, with no overload.” Nigel was almost laughing with delight, but he eventually calmed down enough to explain.

“Sorry about that, but it really is exciting. Look, the reason you want six months’ warning of jobs is because you’re trying to schedule the constraint for everything that comes in, in due date order, so it completes in time, right?”

“Yes, that’s right.” Sarah nodded slowly. “But I still don’t see what– “

Nigel cut her off. “And the reason that never works is that due dates change and big spikes of work arrive unexpectedly, and you spend all your time creating schedules that are out of date before you’ve finished.”

“Yes, that’s what I’ve been saying. What’s your point?”

“We use buckets,” Nigel explained.

“One bucket is what’s due for release in the next day, another is for what’s due in the following week. We don’t fiddle with those two unless there’s a genuine emergency. Another bucket is for what’s due in the next two weeks, we pick from that to choose what goes into the next week’s bucket. And the third bucket is for the next month, say – the exact numbers aren’t too important. What is important is that you can still schedule the constraint, and you’ve got advance warning of anything big coming up for release and whether it will overload the schedules when released. That gives you time to break the bigger jobs up into chunks – great idea, by the way, I’m glad you mentioned it – and schedule them, so that we don’t get huge spikes of work wrecking our lovely system. Even if we have to call a customer occasionally and reschedule the job, that’s still better than delivering late.”

Sarah nodded. “That sounds good. We’d have to try it, and I think there’d be a lot fewer problems with lead time if we did that, and we wouldn’t get overloaded anywhere near as much.” She turned back to her papers with a renewed sense of purpose and waved a hand dismissively at Nigel. “Get out of my hair. I’ve got some buckets to build.”

Over the next month, Nigel started to see the results of the bucket idea – or, as he described it in meetings, their ‘Spike Detection and Levelling’ system.

It hadn’t worked out well at first. Somehow things still got overloaded, to the point shere Nigel had thought they would have to give up on the whole idea. But Sarah could be very determined when she was hunting something down, and eventually, she discovered that the sales team had been closing large multi-day jobs and giving customers a discount based on order size. Even worse, sometimes they made the sale by promising a rush order. Fixing that meant making the buckets visible to sales and editing the rules around discounting. And after that had been fixed things improved rapidly.

Sarah’s predictions had been correct, the lead times are a lot more reliable now, and the factory seems to get overloaded less. Even big orders – spikes – are getting done on time, there are fewer of them reaching the ‘release today’ bucket, and they are no longer affecting the completion of other jobs anywhere near as badly.

From his perspective as finance manager, Nigel could see other benefits too. Spending on expediting was down significantly, and sales were back selling, income was improving slowly and steadily;  customers were realising that Glassware Solutions really could be relied upon.

Perhaps most interestingly, Nigel had found the sales manager looking thoughtfully at the bucket diagrams a couple of days ago. She had commented that there were some gaps in the buckets, and she wanted to see if her sales team could reliably fill them up further in advance.

All things considered; it was like finding a pot of gold at the end of a rainbow. Nigel was starting to think that there might be some helpful spirits out there as well.