Katherine Van Adzin: Hi, everyone! Let’s start off this week with a brief introduction to our guest, Colin Milberg. Colin, tell us about yourself.
Colin Milberg: I am a Lean construction consultant. I've been involved in Lean construction since roughly 2000, when I left my job on the Big Dig to go study with Iris Tommelein and Glenn Ballard at UC Berkeley, and I've been involved in my Lean journey ever since. I was part of the faculty at San Diego State University, and then I decided that the industry was more in dire need of this education than the students coming out, because they were just getting corrupted going into a corrupted industry. My services would be better rendered and my skills would be better rendered working directly with industry, which is how I moved to consulting. I’m also an instructor in the AGC CM Lean certification program.
KV: Thanks for the background, and we’re glad you can join us today! Hal, could you start us off with a definition of takt time?
Hal Macomber: Okay. Takt time comes from the German word, “takt.” It basically means beat, meter or pace, it’s like a metronome or the pace that a group is playing at. In the case of a production setting, whether it's construction or it's manufacturing, the pace is the time between operations. That's what you're talking about when you're saying a pace. On a production line it could be that you get fifty-six seconds to install a steering wheel and then someone gets fifty-six seconds to do the next step in the line. Takt time at Toyota is established based on the rate at which they sell the product on that line, so they establish a pace based on sales rate, but that doesn't have to be the case. You can establish pace based on other goals. The intent, though, is that you design your production approach to a consistent pace. And we'll get into why that's the situation as we explore the production laws.
KV: In a case where takt time isn’t tied to the pace of sales, is it tied to demand for the use of the space by other teams?
CM: The demand of space by others is determined by what you come up with for your takt time. You back into the takt time that you want to try to achieve. Part of that is a discussion of logistics and a bunch of other stuff, which we can get into when we get into the production laws. But once the takt time is established, that is what creates the demand from one step to the next.
HM: For the built environment, there is a promise to turn over space. If you're doing a middle school, you might turn over classrooms before you turn over gymnasiums, or before you turn over something else. But you are starting with some kind of delivery of the product. In our case, it's usually a whole building, or for a roadway, one lane of traffic.
Take the the Longfellow Bridge connecting Boston with Cambridge, for example. They had established a turnover date of one lane of traffic at a certain point in time, but they didn't meet it because they found some additional problems with the structure. It should be tied to the value the customer is getting. Ideally, you don't turn it over after they need it, you turn it over earlier, as Colin said, if that's valuable to them. And if it’s not valuable to them then maybe you don't turn it over earlier.
CM: A perfect example of that is a school. I had an example of that with a client recently where they couldn't turn it over in February, with the manpower and the pace and the flow that they had going for the building, they could have met that, but that didn't help the school because they were going to make a shift in a February break. So they adjusted their takt and all of their crew sizes down to meet that date that was needed and release that manpower to other projects.
HM: It’s important to note that you don't establish takt by inspection, you establish takt by policy. You design your production operation or construction operations to meet that takt. Lahey Hospital is a recent example. Lahey was a project performed by Bond Brothers, it was a renovation of space. We decided that every single operation, once we got it back to bare walls, was going to be a one-day operation, and that whatever it was—layout, studs, blocking, pipe, whatever it was that was going to be done for a particular flow unit—was going to be a one-day takt time.
So first, we identified what the work operations would be. We sequenced the work operations, and then we said, "All right, now what would be the crew size that will allow us to do that in a one-day cycle?" I forget how many different handoffs there were, over two-dozen for one of the flow units. We were able to establish a crew size that could do it in a day, and they didn't have any time left over. The intent is to turn it over to the next performer every day at the same pace.
So Colin, would you like to explain the big benefit of working to a short takt time?
CM: Yeah, absolutely. I'm going to tie this to the first law, Little’s Law (or production law) which you could also think of as a law of batching, or a small batch, or reduce work in process. However it works for you, it's all basically saying the same thing. What it says is that the smaller the flow unit that you hand off to the next step in the operation, the less your total duration for your project or process will be, so you finish sooner.
And I can give a simple example of that. If we take a five-story building and five trades working their way through, and I define the flow unit as a floor, then I'm going by area here for a moment with a takt of a week, so each trade has a week to move through each floor. And so the first trade is going to finish at the end of week five. The second trade will finish at the end of week six, week seven, week eight, week nine. I have a nine-week schedule.
Without changing the manpower, if this is perfectly balanced, I reduce my takt time to a day. I'm handing off a fifth of a floor every day, my new flow unit is the fifth of the floor, not the whole floor. And I plan my work in such a way that I can release that to the next person, the next day.
I've reduced the handoff duration from five days to one day. It's a four-day savings. Four times four is sixteen days, and nine weeks minus sixteen days is five weeks and four days. To play that out: the first trade finishes at the end of week five, second one finishes at five weeks and one day, two days, three days, four days. There's the simple math.
The smaller you make your takt, the faster you will finish without increasing manpower, without accelerating, without any of those pieces.
HM: For any given trade, it was a week’s worth of time on the floor. The fifth floor finishes, so they still have five days on the floor, but they hand off to the next trade after one day as opposed to after five days. The first trade finishes on a Friday, the second trade finishes the next Monday, next trade Tuesday, Wednesday, Thursday, and we're all done in twenty-nine days as opposed to being all done in forty-five days.
It's a big deal, for a lot of reasons. One of the reasons it's a big deal is because of general conditions. The general costs of running a project occur over time, based on the amount of time you're running the project. The CM staff is out there as long as people are building, but if you're getting done faster, that's sixteen days from a nine-week schedule.
CM: Yeah, you saved over a third of your time. Your general conditions, the general costs of running the projects, the overhead of running the project, the general administrative costs like keeping track of people on the job—all that stuff, it's all reduced by a third, and you got the same project done.
You didn't cut back on any of the value-added steps. By the way, we learned to cut back on the value-added steps when we turn over more often. There are those secondary benefits. But the first order effect is the project is shorter, automatically, and a third shorter is a regular occurrence. And therefore you lower the overhead costs of doing the project, which isn’t value added. None of the work that the CM does is value-added, none of the work that administrative people do is value-added, none of the insurance that you're paying adds value. But now you're paying for insurance for two-thirds of the time as opposed to the whole time.
KV: How are you defining value-added?
HM: Anything associated with material in the field is direct value-added. On the flip side, there are three types of waste. Type one waste is anything that's necessary, but not value-adding, like climbing ladders, getting tools, taking measurements. There are all kinds of things that are necessary but you're not fabricating, you're not coding, you're not installing. There's a handful of things in the construction world that are value-added, and then there are a lot of things that are enabling. Type one waste is enabling actions.
Type two is basic waste. Moving stuff around that shouldn't have been there is type two waste. Every time we deliver all of the drywall to a project at once, you know it's not all installed the next day, so people are moving around to get the drywall, and sometimes moving the drywall because it's in the way of other work that needs to be done. Type two waste also includes rework, it includes waiting, that sort of thing.
And then type three waste is superfluous work. We coined superfluous work as type three waste. Superfluous waste is waste arising from the management of other work. For example, if you take a really long time answering an RFI. RFI is waste, it's do-over work, so we count RFI as type two waste. But instead of ask the question, get the answer, we have to put it in writing because we don't know when we're going to get the answer. We have to put it in writing, we then have to have a computer system to track it and a workflow to pass it to the right people. All of the work associated with the duration that it takes, and the handoffs, and the approvals, and all that goes along with answering that RFI is superfluous work. If you just answered the RFI, none of that would be needed. And that's type three waste.
All this waste goes down when you begin taking a takt time approach on projects. Now, I said there's a secondary effect. Colin, why don't you speak about how the more frequently we have handoffs, the better the opportunity is for learning?
CM: Each time you make a handoff, it's an opportunity to have a conversation with the person that you're handing off to, to evaluate and reflect on whether you did things in the most effective manner? Did you get what you needed? If we only hand off one floor at a time and there was something that didn't work, one, you have to go back and fix it, which is the failure demand piece, but two, using that same example, each trade only has an opportunity to learn from the previous one what worked for them or didn't work for them four times because they get handed off five times.
But if we're handing off a fifth of the floor today, that's five times five, so twenty-five learning opportunities where we have the conversation that there is in making a promise, right? You have a person who makes the request and the person who makes the promise, and that's a conversation that occurs. And in that conversation, you can have learning every time the promiser says, "I'm done." And the requester says, "Thank you. And yes, that was what I asked for.” So, you've increased those numbers of conversations by a factor of five, thus increasing the learning opportunity by a factor of five and the opportunity to do something different.
HM: One of the five big ideas is to tightly couple learning with action. Find a way to learn more by doing something and learning from it. Maybe you learn to work in smaller batches. Oh, well what pace should you work at? The fastest pace is not a day, the fastest pace is numerous times within the day.
CM: We were actually on a project, it was a locker room renovation and we did it by stall and it was two hours, and they were turning over one stall at a time to the next trade.
HM: What do we mean by flow unit? Flow unit can be general space, or it could be a room, a system or it could be a stall. In the case of the Lahey project, there were a number of flow units, but the primary flow unit was an examination room. And there were eighty-three or eighty-four exam rooms. The smallest crew that we were going to use was two people and they could do the blocking in seven exam rooms in a day. An exam room was a flow unit and our batch size for the day was seven exam rooms.
CM: I think people get confused when we talk about a flow unit and we describe it as a room or system because they think that all flow units have to be the same, right? A flow unit is a space. It doesn't matter if it's the same type of space, all that matters is that the sequence of steps that are going to move through that flow unit and add value to it are the same general sequence of steps. It could be an exam room here, and it could be a nurse's station there, and it could be an administrative office here, and those are all flow units going through the same process or flow.
Once we've defined what the process steps are, as long as those generically repeat as they go through, it doesn't matter what the nature of the physical space that we're defining as the flow unit is. What we might call a flow unit or area might be different, but it's still just the next flow unit.
CM: The second production law, some people call it the law of bottlenecks, or the law of balance, basically says we can only go as fast as the slowest element, so we all should be moving at the same pace through the flow units. Fundamentally, this is the production law that basically says you need takt.
Imagine for a moment as people are moving through a building or moving through the foundation work, if every single step were moving behind at exactly the same pace, both releasing and receiving a work area from the previous step at the same pace, then that release and receive would create a kind of continuous flow. You're releasing a physical work area that you can't logistically both work in the same time. As that release and receiving happens at the same time, it creates a force, a metronome step where everybody is moving together. If you're applying the law of balance, as I'm going to refer to it, you are creating a takt. There's no way around that, and that's what we need to acknowledge.
One of the most challenging things though is figuring out in a complex building or whatever else the case may be, how do we achieve that balance in the density of work for every single trip? It requires a lot of thought and a lot of planning up front.
HM: And it can lead to innovation. J.C. Cannistraro, a mechanical contractor that is a Touchplan user, has moved a lot of the fabrication and assembly steps off-site. They're doing off-site construction. People call it prefabrication, but the general term is off-site construction, so that now the only thing they're doing when they're installing the bathroom is they're bringing in large assemblies that need to be put in place and a couple of connections made as opposed to everything else inside the flow unit.
CM: If you have an activity that is a huge bottleneck, you can gain huge improvement by eliminating it by moving some of the work off-site and outside before the flow unit. And that's the real advantage of it from a project standpoint of prefabrication as a technique. It's one of the tools for balancing the work.
HM: A great example of this approach is what the Granger team did on their project for Sparrow Hospital’s pediatric nursing floor. There were two-dozen or so patient rooms, and their batch size was two. Every day, every operation was done in two rooms, but there were a couple of operations that took two crews. It took them all day, one crew all day to get the room done and they had a second crew in the other room all day working on that room. But the commitment was still that we're turning over two rooms every day to the next group. There was one crew that had half a day's worth of work in two rooms, and so in the morning, they did their half a day in the two rooms, in the afternoon they went to a different flow unit.
Another example is the Baltimore Harbor Tunnel, there were thirty-two fans that needed to be replaced. They're allowed to have four fans out of service on each side of the tunnel at a time, so one quarter of the fans could be out of service. They were going to do this as a batch of eight. I said, “No, we're going to do this as two batches of one, and every week take one off and bring one back on.” It took us five weeks to do it. Their schedule went from four years to two-and-a-half years to do that project by going to much smaller batches.
CM: If every day, everybody is releasing something, it’s much easier to keep track of and manage than many different people having different handoffs.
So let’s talk about the next production law, the law of variation. I'm going to call it the law of reliability to state the goal—we don't want variation. If we have variation in the steps in the process, we're going to end up having either work waiting for workers to move into the next step, or workers waiting to move into the next step because it's not ready, right?
Sometimes people confuse the law of balance to mean the average pace. That's not what we mean. We mean a definitive pace. If we're going to have that pulse so that everybody moves together, it can't have variation. Takt time planning really forces a slightly different way of thinking than some people approach the Last Planner® System. Lots of people have looked at it as, “We're going to try to continuously remove variation from our system, but we're going to assume that it exists, so we're not even going to go for a steady pulse or pace. We're going to build a bunch of buffers and we're going to try to slowly improve that to make things better. And we're going to choose fairly large batch sizes in that because we're not comfortable with the amount of variation that there is.”
When we take a takt time planning approach, we’re saying that we're trying to continuously improve to remove variation. We also talk about pull, which is the other thing that I want to mention. Pull is not a production law, it's a method of control. Pull is how we create signals to only do work at the pace at which the next person needs it. Pull as a control method is there when we assume that variation will occur. It says if one person slows down, we need a mechanism to tell everybody else to slow down with them, right? So that we're not running into each other. And if one person speeds up, let's try to speed up together.
HM: Yeah. So Dave, you and I've talked about takt planning off and on for over a year now. What questions do you have?
Dave Rolin, Touchplan’s Vice President of Software Products: What do you guys find is the biggest pushback to adopting an approach like this, or why isn't it an industry standard?
CM: In many cases, because they don't believe that the possibility of that reliability exists, they still view everything as one of a kind. There's still this idea that if I stay on the same task for a repeated period of time, I will be more productive and resource-efficient in that sense. But there's no reason you have to stick your whole crew on one activity at a time. Break it up. But they go to trainings where they're told, "Keep the crew on the same step one at a time to get your maximum productivity.” And there are still people out there telling them that. And that's what makes it difficult. There are all these preconceptions that we're fighting against.
Another factor is that on the CM side, many of them do a terrible job of procurement and make ready planning, and so they inject a ton of variation into the system, which causes the trades to have no faith that they can actually operate on this space.
HM: I'll add a couple things. Number one is that we really have a lot of resignation about what's possible with the circumstances today. For the most part, companies are bidding work and they're making a little bit of money or they lose a little bit of money, and they're bidding work based on the history of doing work. Well, when you look at the history of doing work, it's got a lot of waste in it, it's not reliably managed. There's nothing like the Last Planner System going on. There are no practices of making work ready, and so a superintendent’s perspective is that they spent over half of their day every single day following up on what didn't happen yesterday. And that's how they understand their job.
Their experience tells them that if they didn't tell people every day what they need to be doing, that the work wouldn't get done because people aren't responsible enough to get it done. But people do want to work reliably, they do want to do the work once, they don't want to do it over again. They do want to stay busy, they don't want to be hanging around. They can be trusted. In fact, the trades know more about their work than the superintendent does. We get an immediate improvement as soon as we stop doing three-week look ahead schedules by the superintendent, and we just do weekly work plans, day-by-day work plans by the foreman.
It's a kind of systemic resignation, Dave, and it takes a while to open some people up. Their real world is a challenge, it's an issue about people's identity. And what we're doing challenges identity.