The Golden Thread

Proposed future state VSM 

Over the last couple of posts, I’ve been working through an example of using a high-level value stream map (VSM) to uncover opportunities that deliver actual practical actions for change. The VSM showed how the imbalance between site and factory leads to increasing inventory and costs as several building sites are needed to match the factory output. This led to the proposal ‘balance site and factory lead times’ as a breakout objective for a Hoshin Kanri X matrix, along with a future state map and the strategies, initiatives, and measurements needed to deliver results.

This example showed not only how these powerful tools work, but also how quickly and easily they can deliver results. Instead of spending months working to populate a VSM with data, you can start focusing on changes that deliver an impact, then using that change process to fill out the details. Focusing your resources where they are needed, as opposed to risking a model line that struggles to demonstrate the power of lean due to surrounding constraints.

In the last post one of the processes targeted for reduction was ‘hand over’, driving it down from 2 weeks to less than 2 days, our Hoshin plan highlighted strategies and initiative that will start to deliver this, such as, right first time and eliminating of snagging rework. But another factor is making sure all information is in the right hands at the right time ready for the final hand-over. Failing to do this can lead to significant waste causing delays to the project and preventing completion. Waiting on information is a common issue throughout the whole value stream and is the biggest cause of delays I’ve witnessed. Issues involve missing factory drawings, specific details or dimensions, and material specifications causing direct delays to purchasing and production. But the handover of information is also critical for approvals such as warranty, fire sign off, and building control, without which, stops the project hand over.

In the previous post I mentioned briefly that a full VSM is more than just mapping the process flow, and the example is technically an unfinished document. Mapped across the top of the diagram should be the flow of information, what is needed, who needs it and why it’s needed. Understanding this is absolutely critical and spending time to fill this out can massively reduce risks later in a project. As with all Lean processes it’s best to start closest to the customer and work backwards as you expand the VSM and once complete you’ll have your ‘golden thread’.

The Golden thread is one of those buzz words I’ve raised as an issue before, it’s often spoken about without any real specific detail on how to do it or what it is or should be. Without understanding what you need and why there is a risk of gathering information for information’s sake, which is over processing and so waste, ultimately the information flow from the VSM shows what’s needed for a complete golden thread. 

Information Flow added to the high level VSM

Using the future state VSM from the last post we can sketch out key information milestones and link them to their source. For this example, we’ll focus on the handover requirements as part of our Hoshin plan and expand out some of the VSM stages. A good tool to support, with this is known as SIPOC (suppliers, Inputs, Process, Outputs and Customers). By producing a SIPOC diagram on the handover stage we can start to expand on the processes that need to happen and list out what the process customers’ needs are. In turn that will show what inputs are needed and then who the suppliers are.

Example SIPOC of the handover process

This diagram is far from comprehensive and the next step would be to work with all the interested parties to ensure it's accuracy and expand the processes into their own diagrams. But even at this generic level it has value linking the suppler of information to what it's used for, knowing that we can make sure we know how we're going to do that and ensure a smooth handover. The diagram also demonstrates that the information falls into two main supply categories;

• Design information
• Evidence of the build process

In theory the documentation for design and specification should be easy, coming from the design phase of the process much earlier and at the point where a building is nearing completion it should be finished and readily available for handover, but what about the evidence?

Traditional clerk of works and building control processes struggle with the offsite setting due to structure of the work progress. Instead of building batches, in phases, in one location where the inspector can visit and sign off a single phase of work across a project in one go. The volumetric offsite process should be a flow process with modules at all stages of work very quickly from manufacturing start to site install and finishing. In the example we’ve been looking at, it takes around a week from the project start to the first module ready for delivery. So, it’s vital to capture your evidence along the way as smoothly and efficiently as possible.

Logically then a digital system would seem to be the way to go and there is no shortage of available companies that have products that can help. The difficulty here is that it again gets into a world of abbreviations, acronyms and buzzwords. Terms like IR4.0, MES, QMS, IOT, RFID, RFI, NCR, BIM, ITPs and more. This can be a minefield of options where one software package does about 40% really well, has addons for another 30-40% and just doesn’t have anything for the remaining. On one project when we reviewed options for external systems it involved 5 different companies/software packages that needed to work together. While one company was going to be our main supplier and responsible for coordinating the rest this does bring a risk of delays for updates and support, or worse if there is a falling out between the suppliers. On the flip side I’ve working in factories that run on printed forms made in excel, manually completed and scanned. Having to update and format the Excel file for new projects was time consuming and frustrating, so I turned to macros to automate this process, saving a week or more of admin for each project. That still didn’t solve the biggest issues though as no matter how fancy or advanced a system is, if they aren’t used you won't have the data you need to deliver those hand over processes.

“We shouldn’t be marking our own homework”

Last year I wrote about culture, and this is a prime example of a culture of ‘not our job’ causing problems, a production team that wants quality to have sole responsibility for filling out inspections will then be stood around waiting (and likely complaining) for the inspectors to carry out the inspections.

This is where the Lean pillar of built-in quality is so important; no one should know the job better than the people who do it. The workers on the shop floor who do a job day in, and day out should know what good looks like better than anyone. When completing a quality sign off record and taking pictures, they’re presenting the work as completed. It’s not marking the homework it’s filling out the answer sheet saying you are done and have completed the task to the required standard. Likewise, the quality team are not there to mark your work but are an ally, proofreading your submission, a second pair of eyes to help avoid mistakes passing forward and ensuring the required information is captured clearly. Even a dedicated quality bay is there to back up the signoffs and function as a safeguard against issues going down the line, getting harder to spot and exponentially more expensive to put right.

So, who’s marking that homework? First up is the warranty provider and building control; they are the ones who sign off the building as safe and complete. Then should the worst happen incident investigators will go over that paperwork and records with a fine comb with serious consequences if it fails to pass muster.

Ensuring a strong cultural attitude towards quality is the first step towards successfully delivering a golden thread and being able to ensure that documentation is completed ‘right first time’, ‘on time in full’. This then opens the ability to submit that documentation through the project at milestones, reducing the amount that needs to be reviewed at the final handover helping to deliver the targets of the X matrix. To aid with this the system should not be a burden on the workforce and this is where modern digital platforms begin to shine because they can add advantages that outweigh the manual actions of filling out forms. But with so many options how do you choose? 

Keep it simple remembering that the core aim of the system is to gather the evidence. Get that specification worked out first then decide what else to add with a reason for each feature you choose to add. It doesn’t have to be all singing all dancing on day one, and in the past, I developed a working system in Microsoft Access that offered several benefits over the printed forms

• eliminated the need for printing and scanning
• could run reports to pdf for checking work in progress and completion
• supported production control and kept track of module progress
• gave live updates and tracking to a big screen
• could be saved as an app so you don’t even need a 365 license for the whole shop floor

The extra features for live updates and production control benefited the production supervisors, production engineers and quality teams by removing a significant amount of admin work, freeing up their time to be out on the shop floor supporting the team.

MS Access Proved surprisingly capable as a live digital system

Not everything is perfect though and the downsides of this approach included, 

• user lag, the system could be a little slow to run and needed optimisation
• multiple app instance feeding into a core database on the server common drive caused some development headaches
• it couldn’t be used to take photos though it can track and manage photos taken manually on the device
• it’s still very manual for the shop floor staff
• it’s limited by the features of access and visual basics which are pushed very hard already to tackle the 1st two points
• Limited or no integration outside of the database to wider business systems like planning or purchasing/financial
• Internal resource needs to maintain the system and there was a lot of bug chasing to start with

Despite this, it makes a great option for prototyping and testing what works and allowed for a lot of flexibility and rapid iteration as it could be configured and modified in the factory. And if your going to spend time in front of an MS office program why not get the most out of it?

The next step up is an external specialist supplier and there are a range of standard systems on the market that do a fantastic job delivering the benefits listed and overcoming the majority of the downsides. Considerations do need to be made towards the 40/40/20% and multi package systems mentioned earlier but working with a supplier can get a system up and running quite quickly. One problem in particular though is, building site integration, most off the shelf factory systems are aimed at traditional manufacturing environments they may not be suitable for the site team, and vice versa.  Additionally, the systems can still be a burden on the shop floor workers both from the time to fill out forms and working around large panels or modules requiring the return to a workstation or looking after a tablet. So, it’s important to look at options and possibilities to capture the information as smoothly as possible or better yet take advantage of technological advances so that working with the system is easier than not.

Two ways to achieve this involve

1. Delivery of drawings and build information. This can be through the tablets and screens which offers the ability to zoom in on details and have quick links between files, they can also use layers or step by step instructions to simplify drawing to only what’s needed for each step of a process. There are even projection systems that can display the information onto the workspace.
2. Automatic quality measurements, such as dimensional checks or squareness, by having the system replace a task it makes the operatives life easier and advances in photogrammetry and AI are making this more affordable for more tasks. 

Finding the right system that already has these integrated and able to interface with traditional construction design systems can be difficult. The size of offsite components is likely to need a bespoke solution and in the case of AI and photogrammetry will need a supplier who’s willing to work with you to develop these new technologies and adopt them live into their software quickly.

It was a combination of 

• the risk factors, 
• the desire for a single end to end system 
• from the first component parts being picked in stores, to running the reports for final handover and 
• the ability to stay at the cutting edge of technology 

that lead to Lighthouse choosing to develop it’s own bespoke system. This had the added advantage that it opened up funding opportunities and support from universities through knowledge transfer partnership programs and innovate bids. We successfully deployed the system both in the factory and on site for module delivery and had Lighthouse continued to operate, the system had a lot of promise, but it was a lot of work and required more dedicate resource as other distractions slowed development.

Lighthouse MES

However, you choose to do it Building a solid golden thread process is massively beneficial and after training is my number 2 priority for investment and ideally planned and started before opening the factory and above any form of automation. After all, the same missing documentation and evidence will hold up hand over and project progress for £100 thousand or £100 million factory set up. Driving up inventory and then costs, as destructive inspections are called for on a finished building, absorbing time from supervisors, project staff and design team that could be better spent on the next project.

Now I did say ‘in theory design is easy’ and I’m sure a few people reading this will be sceptical and maybe even rushing to the comments. Afterall, like the missing evidence, lack of material specification and design info can also stop any factory and project. In this context though at hand over the design process should be long finished and producing as built drawings is rework and therefore waste. Getting design right first time, on time in full, plays a critical role in the golden thread and project success. So next time we’ll be looking at the design process, kit’s of parts and standard details.

And until then if you’d like to discuss more about the golden thread, would like support with your system or talk about systems you would like to share, please feel free to leave a comment or message me on LinkedIn.

Bringing the factory to site

 

80 Mod Project rough sketch Value Stream Map (VSM)

This VSM from my last post was developed to highlight how, with a 20 minute white board sketch, clear insights can used to direct a company's strategy. It helped highlight the issues with the factory focus of recent years where getting the factory output running has been the main focus without the same level of focus on the site install and completion of projects, ensuring keys are reaching the hands of their customers.
The take away from this quick VSM clearly showed the inventory levels in red at each step of the process, with a factory outputting 3 modules a day and delivering a project every 6 weeks. This requires 4/5 active sites running at any one time and up to 400 modules of unfinished inventory to consume that relatively low factory output. The historic approach from all the factories I've visited and worked in has been to try and increase factory output, often with targets to double output in this quarter or that. If successful with this example it was budgeted to cost somewhere in the region of £4 million and site was just left to the same methods and processes, meaning that we would have needed 8-10 building sites open at one time. So that's 8-10 site teams, scaffold installs, cranes, logistics support, QS, procurement, design support, on hire equipment, and worst of all risks, problems, weather, and delays.

When we use Goldratt's 3 metrics to see if a company is 'making money',

'Increase throughput while simultaneously decreasing inventory and operational expense'

It's clear that doubling the factory output isn't going to work and it's no surprise that the losses from the likes of L&G, Ilke, TopHat, Urban Splash, Swan, Lighthouse, and the rest of the mid 2010's wave of offsite manufacturers, adds up to easily a £1 billion if not more.

So what do we do? the first step is to go back to the Value stream and spend a bit of time proposing a future state VSM. Something like this.

Future state VSM for site process

Setting a target reduction in lead time of 15weeks, from 21 week to 6 weeks to match the factory we are aiming to balance site with the factory.
In a Lean transformation it is important to set ambitious targets that give a true competitive advantage and develop a sound strategy and plan to achieve them, the tool commonly used for this is Hoshin Kanri also know as Hoshin planning. It's powerful for strategic planning as it also ties into the lean culture and when done right helps guide the whole company in the right direction. Robert Camp's book the "lean leader" covers the Hoshin process in a good level of detail that's easy to follow and I highly recommend it.

First up is to set Breakthrough objectives, in this case we'll only look at the one and that is to balance site and factory lead times. From that we need to create strategies each with appropriate tactics to deliver their breakthrough objective. I discussed the first obvious one in my last post, the elimination of the 4 weeks of snagging. Snagging is rework, it's pure waste and Lean's 'built in quality' pillar offers the solution, practices like 'in station control' and 'error proofing' works well to ensure a right first time approach and remove the need for snagging. This can be delivered through training, short improvement events (Kizan events) to tackle common problems, and quality gateways to ensure no fault is passed forward. One of the big benefits of reducing the number concurrent active sites is that you can really focus on developing a crack install and finishing team, and if you're not already working like this at the factory it's great time to get the benefits both at the factory and onsite. Plus you can deliver both for less than £40k a fraction of the factory ramp up mentioned earlier.

So a strategy of 'eliminate snagging' gets the Tactics of Training, a focused Kizan burst (multiple events in a short space of time), and quality gateways. Lastly for snagging will be the Measurement targets, zero snags past the gateways and 100% defect recording at the gateways and anything afterwards, with 100% investigations on any snags that do slip through.
This all sounds like a tall order but I've actively delivered these results in a factory when implementing a quality gateway before the finishing line. I've never been prouder at work, than when I got to tell a decorator that "he was giving me the best module I've ever inspected" as he smashed my targets. The next day though someone else gave me an even better one really showing what was possible and we where only a portion of the way through the process.

The thing about Hoshin planning is that the strategies don't have to deliver the whole breakthrough objective in one go, the objective timeline could easily (and likely will) span multiple years. So a strategy or it's tactics can be focused on taking a step toward that goal. In this case we want to reduce 5 weeks install and 10 weeks of finishing to 28 days. That sort of change, is not going to happen over night, and will require steps in the right direction. In Lean this is know as the PDCA cycle as described by the American Society for Quality (ASQ)

The Plan-do-check-act Procedure

Plan: Recognize an opportunity and plan a change.

Do: Test the change. Carry out a small-scale study.

Check: Review the test, analyze the results, and identify what you’ve learned.

Act: Take action based on what you learned in the study step. If the change did not work, go through the cycle again with a different plan. If you were successful, incorporate what you learned from the test into wider changes. Use what you learned to plan new improvements, beginning the cycle again.

With that in mind a sensible mid point is to bring the VSM closer to the future state diagram, by separating internal and external finishing. Completing the external finishing alongside the install all within a target of 6 weeks without any weather downtime and zero water ingress.
Now I mentioned a couple of posts back that I'd been talking with Paul Richards from GUR and he has exactly the kind of solution that could offer these results whilst forming the foundation for further strategies and PDCA cycles. 

Thunderhaus by GUR (photos curtesy of GUR)

The Thunderhaus is more than just a building site in a tent, it is intended to bring the factory conditions to the building site. With a clean and stable flooring, access walkways, integrated gantry cranes and storage it opens up so much potential.

Inside of the Thunderhaus (photos curtesy of GUR)

Immediate results protect the site from weather, a gantry crane designed for module handling and safe access systems. The conditions from the basis for so many Kizan events and the implementation Standardised work and 5S directly on site, factory style jigs and fixtures along with quality systems and further 4.0 solutions. Then in the future the possibilities extend alongside the best of factory innovations as a solid platform for robotics and automation. Every improvement in the factory there is the potential for matching ones onsite, to help keep the balance once it's achieved and ramp up from that 3 modules a day.

With deployment and development of the Thunderhaus as a tactic to help install and protect the modules a second tactic can be used to develop the process towards a full external finish at the end of 6 weeks is needed. This could well involve looking at DFMA and premanufactured Value (PMV), often there is an attempt to get as much external work completed in the factory, but this then leads to issues on site. Insulation and external cladding installed in the factory needs module to module zip up which can be a jigsaw of bits or a lot of cutting down standard sheets. Complex factory roofing work directly installed to the modules, can also cause issues in the factory. One tactic to help alleviate these problems could be the use of 2d panellised facades, using a specialist supplier. A company who's expertise is in a watertight envelope has a lot of potential, installed with the cranes, it's design wouldn't need to match to the module layout giving a lot more architectural flexibility.

A combination of Kizan and DFMA is also likely a key part to reducing the finishing time from 10 to 6 weeks for the intermediate step and beyond. I've worked on a number of solutions, material selections and design choices and then can be quite long projects. At this point getting the basics right is key and the same tactics for zero snagging can have a massive impact on lead times for finishing works, especially well managed quality gateways so it's worth rolling the 6 week finishing and 0 weeks snagging together at this stage.

These quality gateways are also key to the reduction of the hand over time, from weeks to just a couple of days, This is where the value of industry 4.0 and digital solutions can really shine, as information is updated to the client as each stage is finished. In my next post I'll be looking at the top line of the VSM, the tracking of information, and the golden thread in more detail. In my experience it's the failure to understand what's needed, when it's needed and where it comes from that caused all of the biggest delays and problems on projects.

For now though the Breakout objective, Strategies, Tactics and measurements are recorded in an X matrix. Following this, project teams and accountable people are assigned to the tactics, they then develop the plan to deliver and the key performance indicators (KPIs) that will track results. Throughout this PDCA cycles will run and the information in the current state VSM will get updated with more and more details to help identify further strategies, tactics, and measurements until the Breakout objective is achieved.

And if you find these articles interesting and would like to discuss more or are interested in support for your offsite Hoshin Kanri, please feel free to leave a comment or message me on LinkedIn.

Delivering the Value

Recapping the previous post, we explored the misconceptions surrounding Lean and its applicability to offsite construction. Unlike the mass production of Henry Ford, Lean was created specifically to manage lower volumes and higher variability. We also discussed how the ideal Lean system can be adapted in the beginning, with continuous improvement to move towards that ideal.

Many definitions of Lean exist, but few provide context for our industry, it’s not surprising that these misconceptions exist and the lack of clarity can lead to confusion. That’s why I’m passionate about getting back to the basics and digging into what makes lean work. This series aims to set down some key ideas to follow and give them context within our industry. These key ideas are intended to help start the journey into lean principles and build a strong foundation, both in practice and company culture. 

In their book ‘Lean Thinking’ Womack and Jones[1] introduce the ideas that form the core of lean along with guidance through the process, ‘Lean for Dummies’[2] defines this as a 4 step process. 

1. Scrutinize the value stream
2. Keep it Flowing
3. Pulling through the system
4. Striving for Perfection

Earlier this year Autodesk released their guide[3] with 5 steps (they separate ‘identify value’ from the ‘value stream mapping’ (VSM) in step 1 above). The Autodesk construction cloud has a number of good resources on the subject. I may not totally agree with everything they say but it’s worth reviewing.

However, no matter how you approach it, the first step is simply to understand what you need to do to deliver the customer what they want. This value is defined by the customer and what they are willing to pay for as defined by Womack and Jones[1].

‘A specific product (a good or a service, and often both at once) which meets the customer’s needs as a specific price at a specific time’

The value stream is the process that answers the question, the customer wants something how do you get it there? Or in simple terms what activities are needed to take the raw parts and convert them into the final product. The terms value adding (VA) and non-value adding (NVA) activities are used to describe these activities, defined by Sayer and Williams[2] value added activities must meet three criteria

• The customer must be willing to pay for it
• The activity must transition the product or service in some way
• The activity must be done correctly the first time.

Install of Modules on site, VA or NVA?

Value stream mapping is the documentation of this process, when discussing how important considering the whole system is I briefly introduced VSM in part 2 of ‘Exploring the lean house’. The common approach is detailed and in-depth. Typically, at the beginning of a lean transformation this may be a section of the manufacturing process know as the model line. This gives a demonstration of the power of Lean, and the success of the model line then garners enthusiasm and commitment from the wider company. Those involved and trained on the model line will take their learning and expand lean practices through the company. This typical transformation is commonly applied to an established company in an established industry. 

The construction sector however is relatively new to industrialisation with more in common with Industry 1.0 than modern manufacturing. So, the question is how do we get started? Like the situations in my last post there are benefits to bending the standard approach to get started and then use continuous improvement to move towards the ideal.

“Let’s carry out a detailed value stream map to identify the low hanging fruit”

This is a genuine quote from a former co-worker, it’s clearly contradictory but helps illustrate the issue with detailed VSM projects early on. By definition ‘low hanging fruit’ should be obvious so spending time detailing out a VSM to find it is unnecessary. Then, solving those issues will likely have a big impact on the VSM details, making all the studies, time data and detailed evaluations obsolete. Much better would be to sketch out a high level map to help prioritise the issues, and as you work on each problem record your results and improvements to increase the detail you hold in the VSM.

This approach offers several advantages. Firstly, you can quickly sketch the entire system at a high level. This allows you to prioritise issues and focus on the areas that will have the most significant impact on project delivery. By asking the question does the solution increase 'throughput' while simultaneously decreasing 'inventory' and 'operational expense'? Starting with challenges closest to customer handover helps ensure earlier solutions have a real impact on overall efficiency, preventing bottlenecks and unnecessary inventory. Then, as you tackle each problem, record your results and improvements, adding valuable detail to the VSM while also getting the benefits of the improvements.

Adopting this approach at a high level in the business will also promote a good culture, when making critical decisions on project timescales it’s good practice to ask the questions around the three metrics and if this choice deliver to our customer sooner? It’ll help cement a problem solving culture, one that won’t self-sabotage by pushing inventory into the system just to be seen to be doing something.

Lastly there is another layer beyond the processes that need to be completed, across the top of the VSM information flow is recorded, where it’s produced and where it’s used. By understanding this and working from the customer end, delays and problems can be avoided. Particularly around final sign offs from warranty providers, building control and the local fire brigade. When you know what they will want to see you can make sure you are collecting that info and evidence in advance without the need for potentially destructive inspections.

80 Mod Project rough sketch VSM

If we look at this on an example project, an 80 model block of flats, in a factory with a solid design that’s understood and known it’s not too difficult to get to an output of 3 modules a day. With a 6 day lead time for the first module on the line to leave the building those 80 modules can be built with a 2 week pre-flight and 6 weeks of main production. Typically that 2 weeks prep time overlaps with the end of the previous main production run so having a project finished every 6 weeks isn't out of the question.

Then a typical install currently can manage 5/6 modules a day for a floor a week and then the roof (5 weeks), this is followed by finishing work both internal and external, which can easily take 12 weeks or more and then snagging for another 4 weeks and a couple of weeks for final handover. Now this is some very rough timings and possibly quite generous to many with the current industry performance. But that site time of around 21 weeks means that to consistently use the factory output at least 4 active building sites need to be running at any one time.

All of that means a large stock of unfinished inventory (320-400 modules), lots of concurrent operational expense and poor throughput. Using the VSM approach it very clearly highlights that we need to focus our efforts on the site side of volumetric and MMC in general. This whole exercise took about 20 minutes and it very clearly illustrates an obvious risk, yet companies setting up factories have spent jaw dropping amounts of money on factories and factory infrastructure and clearly haven't been able to grasp this concept. From this simple VSM we can see that investing in a factory that produces 6 modules a day by automating will have a small impact of at best 3 weeks to the total lead time while driving up inventory as now 8 sites would need to be active, whereas targeting the non value add of Snagging has the potential to eliminate 4 weeks, reducing inventory while costing orders of magnitude less to achieve. And none of this even accounts for the most unpredictable and very costly risk to the site side, our delightful British weather.

Snow falls on site

In my next post I'm going to explore how the lean company can approach these risk and problems utilising this knowledge to tackle the low hanging fruit. How the senior team can set out a companies goals and strategies with Hoshin Kanri, targeting the model line to make the biggest bang and even look at a fantastic solution and foundation for revolutionising the site process.

And if you find these articles interesting and would like to discuss more, please feel free to message me on LinkedIn.

[1] Womack J.P., Jones D.T., (2007) Lean Thinking, Simon and Schuster.
[2] Sayer N., Williams B., (2012), Lean for Dummies 2nd edition, John Wiley & Sons, Inc.
[3] Autodesk, (2023) Discover Lean Construction: Why and how your construction teams should go Lean, Autodesk construction cloud

Building a Lean House - Bending the rules to get started

Several times recently, I’ve had discussions where people are surprised that Lean manufacturing is something that would really benefit the offsite sector. Prior to that I’ve had colleagues lamenting the lack of a standard product and how we need have a standard module design, house type or flat. The perception is that lean is about producing a high volume of matching products, “you can have any colour you like as long as it’s black”.  Thing is, that quote is from Henry Ford and the world of mass production. Lean manufacturing, on the other hand, originated in a much smaller market with diverse product offerings. Lean was born in post World War 2 Japan from the principles of lower volume, high variability with significant space and investment constraints preventing Kiichiro Toyoda from simply copying western plants. I find that this tendency to lump all manufacturing together and the lack of a clear distinction between mass production and lean manufacturing produces the misconceptions around lean and offsite. This is why I used this idea to open our exploration the lean house over the last two posts and proposed a clear philosophical line between industry 2 and 3 as well as the technological one. 

Following this, I delved into the power of Throughput, Inventory, and Operating Expense as key metrics, and how Value Stream Mapping helps identify areas for improvement. We've also discussed the ideal of Single Piece Flow, Just-in-Time (JIT) and the reduction of waste for a truly efficient system. I finished up pointing out that Toyota had decades to build TPS and have been refining it for the last 40 years. Achieving perfect single piece flow and JIT can be a challenging feat, especially when starting out. Real-world manufacturing environments often face demand fluctuations, equipment breakdowns, and supplier issues.

So, how do we get started on our Lean journey without getting overwhelmed? I’m a keen advocate of the idea of getting started even if it’s not perfect and then using problem solving and continuous improvement to ramp up. In this pragmatic approach we can bend or even break some of the rules to allow our starting point. But ultimately when deciding where to bend a rule or even go and break it you need to be clear on your destination and the core principles or culture you’re building.

One common example of this is the use of buffers, where a stock of inventory is kept between processes to ensure that the follow-on process won’t be forced to wait should there be an issue downstream. In the offsite volumetric factory this is common between the subassembly of 2D elements such as floors, walls and ceilings and the box build process.

Buffers also help to balance uneven workflows, a prime example in volumetric construction being the site install of modules. A factory will build up an inventory of modules which then get dispatched to meet the install plan, often the whole build is ready at the factory before site install is started even when this might require space for 100 modules or more.

The same goes for stock levels, much of the construction industry supply chain isn’t ready to delver anything close to just in time so the offsite factory requires space to have sizable deliveries and store it as it consumes the materials into the product.

It is reasonable to deliberately create a controlled form of waste (inventory) to prevent another form such (waiting) if by doing so enables you to keep your flow steadier and get the product to the client sooner. The problem comes when the inventory isn’t well controlled by overfilling it or not recognising it as waste that needs to be reduced. Looking at the example of the subassemblies, should there be an issue upstream which prevents the buffer being drawn from, pushing more inventory into that buffer and overfilling it is overproduction so now you’re generating two wastes in an attempt to prevent one. Managing your buffer levels is therefore critical and setting clear minimum and maximum levels controlled with a pull-based system like Kanban will help with this control. One thing to be very clear on is how you manage these levels as output ramps up. It’s easy to set targets for a day’s worth of production and as you ramp up output this level will increase with it, which from 1 module a day to 2 is easy but 2-4? What about 16 modules a day? Your now firmly in the realms of overproduction with oversized buffers and the associated risks such as storage costs of that inventory.

Part of your plan for ramp must include for a clear assessment of maintenance needs, risks and existing performance. If you have evidence that means you are regularly needing buffers at your current rate for a full day’s downtime you’re not ready to ramp up, make your next improvement project focus on the need for the buffer so you can keep it the same size or only slight larger after ramp up as it was before. Then once the ramp up is delivered and running, plan to target your buffers further so you can reduce them especially if they were increased in the ramp up process.

While the use of buffers is about breaking the ideal Just in time system using small batches breaks the pure single piece flow. I always used to joke with a manufacturing director that his batch build processes built a restaurant in 8 week but with a flowing process I could do it in 8 days. Which with the correct prep, jigs and fixtures and process development a flow would be able totally achievable. However we still have some arears of the process where batching has benefits to help us 'get started', the most common is to manage drying times. Tape and jointing or Tiling have very long waiting periods where further work can’t be achieved, so while setting up it’s reasonable to designate these tasks to a batch process. Once you’re up and running improvement projects and work in these areas should be carried out to reduce the size of these batches and there is a real R&D potential for industry to look at replacing old fashioned traditional materials in these areas. 

Like the high speed wheel change in F1 SMED allowed for the efficient change over between products  

By taking this approach the goal is to move towards the more ideal Just in time single piece flow system and is why understanding the underlying principles of Lean is so important to our industry instead of just knowing the tools or jargon. Lean is more than Just in time and Pull is more than just Kanban both things that are more common misconceptions like those above. When you dig into the history of Lean it was SMED or Single Minute Exchange of Die that was the real game changer. It allowed Toyota the ability to develop single piece flow in an industry that traditionally used batch sizes in the hundreds of parts. It moves away from the idea of economies of scale to ones of whole system efficiency and allowed Just in time to come about. In our industry we don’t have complex setups between products, instead we are often faced with difficult product balance as some parts of a building are just fundamentally more complex than others. So, we need to be going back to the basics of the lean house and finding our own enabler like SMED was for Toyota while taking advantage of over a century of learning manufacturing offers us.

One particular challenge unique to our industry is the transition from factory to site, be it a small plumbing subassembly, 2D cassettes or fully fitted volumetric modules. Getting the keys into our customers hands relies on us managing this process, there is no point in producing the best factory products in the world if they go to site, get damaged or wet and take months to finish. To really reap the benefits of offsite and MMC getting the site side sorted and aligned with lean thinking is our enabler. Having recently spoken with Paul Richards from GUR build we discussed this at length and how important it is to bring site thinking into alignment, because right now that disconnect often causes conflicts with factories and the MMC processes. There is so much that can be gained by bringing ‘the factory to the site’ as opposed to a tendency to have a building site in a shed for the factory.

Over the last few posts we've explored many of the principles and ideas underpinning lean, today I applied some of those in a practical manor to a portion of the system. In my next post I’ll be taking this further going into the steps to lean implementation, looking at the practical application of those steps and how they can apply both to the offsite factory and the building site.

And if you find these articles interesting and would like to discuss more, please feel free to message me on LinkedIn.

Exploring the Lean House, Part 2

In Part 1 we Looked again at the Lean house. Explored the limitations of traditional manufacturing approaches and how they can negatively impact a company's financial performance. We delved into Eliyahu M. Goldratt's insightful work, particularly his concept of Throughput, Inventory, and Operating Expense. These three metrics provide a powerful lens for evaluating a system's efficiency and profitability. And I finished part one with a question.

“What would happen to these metrics if we never stopped building or constantly pushed production at individual stations?”

Let’s start today with the answer.

Throughput: the best outcome for throughput is that it won’t change. While it seems like everyone is busy and work is constantly ongoing, factories have bottlenecks, be it in a work station, sales demand or logistical there is somewhere that at 100% isn’t as good as the rest.

Manufacturing isn’t a race it’s more like a group hike, and the job isn’t completed until every activity in the process is done. Having some of the activities running faster doesn’t finish the hike any sooner and 'get the job done' but it can cause a heap of problems.

Inventory: One of those key problems here is a build-up of inventory, be it finished good or work in progress (WIP) and levels will typically skyrocket if production keeps pushing out products without considering actual customer demand or process limitations and downstream holdups. This ties up valuable cash flow and drives up storage and movement costs. In the worst case, as inventory builds up defects and other problems are hidden.

Operating expense: More inventory leads to more expenses; in the absolute best case this will only be the logistical costs above. But often the build up of inventory disconnects the work in progress from the next steps in the job which will lead to finding defects late and will affect more parts as they build up. Causing rework or scrappage costs and so operational expense increases and delays pile up which actively reduces throughput.

Like a restriction in a busy road pushing more traffic into the jam never gets anyone to their destination sooner so modern smart motorways actively reduce the speed up the road from a traffic jam, reducing the amount of traffic that joins it so that when the restriction is removed the jam will clear sooner. The focus therefore should not be on keeping the rest of the factory running but on resolving the issues and removing the restriction to allow the processes behind to flow.

In his further work Goldratt emphasises controlling production rate to manage inventory levels. Where Goldratt proposed careful control of material release as the drum beat for the factory, in lean manufacturing this is done using pull flows. This ensures that work only progresses when there's demand from the next step in the process. In the ideal single piece flow each workstation only has the single WIP product and the lowest amount of materials available to complete that workpiece, each time a product passes to the customer in a sale it opens a space for the work behind it to progress, which in turn opens a space behind it and so on down the line back to the first workstation which can then begin a new order. Likewise, when materials are consumed at the workstation they are then replenished, Toyota use Kanban cards as a method of signalling stores that there is space for replenishment and this ideal process in called Just in time.

This approach is based on considering the whole system, the key to the definition of throughput is that it’s based on the final handover to the client or sale and not the beginning of production. The steps that make up the whole process from end to end is called the value stream and laying this out in a process flow is the first step in building up the value stream map (VSM). As a whole system map, the processes mapped out aren’t just the individual work centres but also includes movement of product, quality control and anything else that needs to happen to get the product to the customer. The VSM should also show the flow of information needed to complete the product and time each steps takes. Lastly, the value stream map helps identify if a process adds value to the final product from the customer's perspective or is considered non-value added. Non-value-added activities consume resources but don't directly contribute to what the customer sees as valuable.

These activities are classed in lean as waste and fall into 8 categories, Transportation, Inventory, Motion, Waiting, Overproduction, Over-processing, Defects and Skills otherwise known as TIM WOODS.

• Transportation: Unnecessary movement of people, materials, or products.
• Inventory: Having more stock than is immediately needed for production.
• Motion: Any unnecessary physical movement by workers that doesn't add value to the product.
• Waiting: Idle time for workers or machines due to delays, breakdowns, or lack of materials.
• Overproduction: Producing more than is demanded by the customer, leading to excess inventory.
• Over-processing: Performing unnecessary steps or using overly complex methods in production.
• Defects: Any product that does not meet quality standards and needs rework or scrapping.
• Skills: Underutilising the skills and knowledge of employees.

By understanding the value stream and using a pull flow, effort can be focused on the issues that prevent the smooth flow of production to the client. Understanding the wastes allows for targeted improvements, achieving a more efficient manufacturing system through continuous improvement.

Rounding off the lean house tour Standardised work forms a part of the foundation in the lean house, if 3 people do a job 3 ways you need 3 different improvements for that task to benefit every time. Standardised work allows for the improvements to help the process every time and with everyone working in the same manor. They all have the opportunity to suggest and help implement changes that will benefit  everyone working with that process. 

This whistle stop look around the lean house over the last few posts has covered at not only the features but on how and why they work. Together these will increase throughput, reduce inventory and operational expense thereby making the company Money, achieving the Goal.

The just in time process of Toyota took decades to develop to where it was in the 80s and they have had a further 40 years of refinement, it’s impossible to deliver the perfect single piece flow line without years of trial and error, next time I’ll take a look at the steps to develop these processes and how you sensibly mange breaking the rules to get started, and have something successful to improve on towards the ideal.

Exploring the Lean house, Part 1

In my previous post I discussed the importance of culture and how it forms the foundation of the Lean house visual, I hinted at the type of culture that needs to be fostered and how training and management leading by example from the very top are critical for a real delivery. This week I’m going to expand on this foundation and explore the lean house further, after all I wouldn’t want culture to just be another amorphous buzzword. 

When we talk about culture we talk about behaviours and responses. With a strong culture these are predictable, be it negative behaviours such as a disrespect for processes and procedures or knowing that owning up to mistakes leads to being yelled at.  Management often fosters these types of environments by only wanting to hear good news and not listening to anything that goes against this, surrounding themselves with people who just agree or make impossible promises.

Positive cultures allow people to take ownership of their mistakes and acknowledge that the majority of the time it’s systems that fail and not individuals. A culture of trust and a willingness to hear bad news or negative feedback, tackling problems head on as a team in a supporting environment.

In Lean, this culture of ownership, trust, and open communication is essential. It fosters a continuous improvement mindset and empowers employees to identify and eliminate waste. This supportive environment is what allows companies to truly embrace the principles of the Lean house, despite seeming counterintuitive to those with a mass production background. 

Not all manufacturing is equal

I’ve made this statement or similar a couple of times in my previous posts and have been an advocate for getting back to the basics almost as long as I’ve been in this industry. During my masters dissertation I took this to heart and delved into the history of manufacturing right back to it’s roots in the industrial revolution. I’ve mentioned Industry 4.0 as a buzzword in my first post and it’s a label for the 4th industrial revolution. The 4 IRs are explained by Vaidyaa et al in their paper ‘Industry 4.0 – A Glimpse’ [1] (figure 1)

The 4 Industrial Revolutions [1]

From ‘Glimpse’ the revolutions are largely described by the applications of specific technologies and that would be the familiar terminology for many, in my dissertation I argued that alongside technology the philosophy and approach also changed with each revolution. Specifically, that during Industry 2 the ideas and methods of Mass production evolved and were delivered most famously by Henry Ford and then with industry 3 Toyota’s Lean manufacturing came to the fore. Arguable we are still in the transition 50 years later, as many companies look to use lean manufacturing but struggle to shake off the shackles of Mass production, something I’ve seen time and time again in the offsite industry.

Every manufacturing manager I’ve worked with has uttered the words ‘we can’t stop building’ or something similar at one point or another and yet time after time delays and issues build up and rework is needed to put things right. Along with this is a reluctance to have people or machines idle, a classic approach from mass production is to judge the efficiency of a factory by the local efficiencies of each individual area pushing to keep everyone and everything busy 100% of the time and measuring them all on how close they can get to that. The problem with these two ideas from mass production, is that they simply don’t work, because if they did the west wouldn’t have struggled so much in 80s and 90s against the newer Lean manufacturers of Japan. If they worked, it would be an American company at the top of the automotive industry and not Toyota who make more money than their top 5 competitors combined [2].

When we look again at the Lean house (figure 2) it highlights a number of items that directly opposes those two ideas. Built in quality highlights the need to stop the line and solve issues not allowing them to be pushed forward, while the principles of Flow and Pull and the focus on delivery to customer require a ‘whole system approach’ which goes against the in station 100% efficiency targets and can actually mean working less efficient in some areas is better for the overall system output.

A diagram of the three pillar Toyota house of Lean management. [3]

To understand why this works we can look outside of the core of lean teachings or that of Toyota, in 1984 Eliyahu M. Goldratt publish their book ‘The goal’ [4] 6 years before Womack et al would publish 'the machine that changed the world'[5] and 20 years before the Toyota way[2]. 

Published as a novel following the manager of a struggling American plant ‘The goal’ proposed that the purpose of any company was simply to make money. While it may be argued that there are other goals by company owners, if the company isn’t making money it’ll struggle for growth or to have the resources for its ultimate purpose.  Using a mentor character in the book to guide the protagonist Goldratt poses a series of questions to encourage the characters (and the reader) to work through the logic of their proposals. Starting with identifying the goal the question is how would a factory be able to measure if it’s making money? Financially investors or company owners can achieve this by 

“Increasing net profit, while simultaneously increasing return on investment, and simultaneously increasing cash flow.”

Goldratt goes on to translate this to the shop floor using three very specific definitions:

• Throughput: the rate at which the system generates money through sales.
• Inventory: Inventory is all the money that the system has invested in purchasing things which it intends to sell.
• Operating expense: Operational expense is all the money the system spends in order to turn inventory into throughput.

Each one of these definitions practically measures financial performance of the business,

“Throughput is the money coming in. Inventory is the money currently inside the system. And operational expense is the money we have to pay out to make throughput happen. One measurement for the incoming money, one for the money still stuck inside, and one for the money going out.”

From these definitions you can directly link the factory performance to practical measurements and targets to take you towards the goal. If you can increase throughput while reducing you inventory and operating expenses you know that you’ll be able to increase net profit, while simultaneously increasing cash flow and return on investment.

I still have a lot to discuss on this and that includes how this links back to the classic tools of Lean including the value stream map, flow, pull and the 8 wastes, however It’s also getting to be a very long post so I’m going to split this post into two parts and like Goldratt’s mentor character I'll pose a couple of questions to help tie this together.

As we've seen, the 'keep building' mentality and focus on individual station efficiency are hallmarks of Mass Production and I made the bold statement that they ‘simply don’t work’. Keeping people busy at their stations is often described as 'getting ahead', usually in areas that can already run faster than processes later down the line. Let's explore the impact of these approaches on Goldratt's three key financial metrics: Throughput, Inventory, and Operating Expense. What would happen to these metrics if we never stopped building or constantly pushed production at individual stations? Does it make money? Please feel free to leave a comment here or on my LinkedIn with your answers and thoughts on this.

And if you find these articles interesting and would like to discuss more, please feel free to message me on LinkedIn.

[1] Vaidya S., Ambad P., Bhosle S. (2018) Industry 4.0–a glimpse. Procedia manufacturing. 20. pp.233-238.
[2] Liker, J. K. (2020). The Toyota Way, Second Edition: 14 Management Principles from the World's Greatest Manufacturer. McGraw-Hill Education.
[3] Fekete, M. and Hulvej, J. (2014) Lean management as a house from the past to the present. Comenius Management Review. (8) 2. p.5-16.
[4] Goldratt, E. M., Cox, J. (2016). The Goal: A Process of Ongoing Improvement. Taylor & Francis.
[5] Womack, J. P., Jones, D. T., Roos, D. (2008). The Machine That Changed the World. United Kingdom: Simon & Schuster UK.

The Power of People and Culture

Last week, we explored the challenges plaguing the offsite construction industry. While the specific triggers for company failures can vary, my experience reveals a consistent pattern. Many problems stem from a critical disconnect: The inability to translate manufacturing concepts into practical workflows and actions.

This disconnect fosters a reactive environment, leading to firefighting and “sticking plaster" solutions, which tigger rework and delays, increasing costs and throttling cashflow. This series of reactive solutions creates a phenomenon I call "process debt." Just like any debt, process debt accumulates over time as companies prioritise pushing projects through over establishing efficient workflows. The more companies rely on "sticking plaster" solutions, the greater the process debt becomes. This cycle of rework, delays, and ineffective solutions perpetuates, eroding trust and creating a reputation for unreliability. The offsite construction industry, battered by a series of bad news, is suffering from a lack of confidence – both internally and among potential customers.

I went on to speak about the need for a clear and well-defined plan and finished on the fable of the Tortoise and the Hare, slow and steady progress often trumps a hasty approach. We need to shift gears from the "hare" mentality to the "tortoise" approach. By creating actionable plans rooted in Lean principles and efficient workflows, we can demonstrate expertise, build trust, and deliver to our customers. However, there is one step that needs to be considered and understood first.

The importance of culture

A diagram of the three pillar Toyota house of Lean management. [1]

The Lean house visual is often used to demonstrate Lean principles and the hierarchy of importance, like with any building that starts with its foundations. In Lean implementation, culture is the foundation. The most meticulous planning won't succeed if people don't believe in the plan and understand its purpose. Core ideas like pull flows and pausing production to solve problems won't be effective if workers react to difficulties by reverting to bad habits under pressure to keep things moving. Therefore, addressing this foundation is crucial, long-term planning must be prioritise over short-term shortcuts, and the company leadership must be involved, willing to adapt and learn, and lead by example through their actions and decisions.

Once a company chooses to adopt this approach the first step to consider is training, I’ve often joked that I’d value £100,000 training budget more than a £100 million investment in automation and plant. Empowered and skilled people can achieve more than idle machines waiting for materials and information, or worse, overproducing parts that then sit unused until downstream processes are ready. Therefore, even as a company develops its plan, it can begin the coaching and learning process.

At Lighthouse, we trained all our production supervisors and manufacturing engineers as Lean practitioners, alongside members of our design, procurement, and project management teams, with attendees ranging from directors to apprentices. It will always be a source of pride of to see the buy-in and enthusiasm from those on the training after a year advocating for the implementation. The icing on the cake? While our senior team worked on the value steam mapping project and business processes, the Lean teams delivered immediate improvements that exceeded the training’s cost.

However, fostering a culture of continuous improvement requires more than just training. We need to address specific aspects of company culture that might hinder progress. Hierarchies that discourage information flow and employee participation, or a fear of failure that stifles innovation, can all impede Lean implementation. Crucially, Leadership must actively foster a culture of trust. Empowered employees who feel valued are more likely to contribute ideas, take ownership, and strive for continuous improvement.

Planning for Success: Tools for a Culture of Continuous Improvement

Lean also provides powerful planning tools to ensure your team is working towards shared goals. Such as Hoshin Kanri, a strategic planning process that helps companies translate their vision into actionable steps at every level. This approach, combined with a strong foundation of trust and continuous improvement, empowers teams and unites the company towards achieving its strategic goals. I’ll talk through Hoshin Planning in a later blog after discussing a couple more of the key portions of the Lean House next week, including the value stream and waste.

[1] FEKETE, M. and HULVEJ, J. (2014) Lean management as a house from the past to the present. Comenius Management Review. (8) 2. p.5-16.