20 Things I learned talking to Modular Insiders

November 15, 2019 Author: Julian Bowron, President, Vector Praxis

Two months ago I emailed my contacts to let them know I was looking for new opportunities. Based on the report created by Constant Contact, about half of the recipients opened the email, and only 6 reported me for spamming. Thank you for that!

As a result of reaching out, I enjoyed many great conversations with colleagues working in the Modular and Offsite Construction Industry. The companies they work for (or own) range from recent start-ups to large established players, with most in the commercial segment.

The most noteworthy thing I discovered is that the companies in our industry are almost all “IP Islands” surrounded by a wall of NDA’s. As a result, many of us are facing similar challenges without access to solutions that others have found. In fact, I heard the same stories multiple times from people in 6 countries on 3 continents.

It’s not surprising that no one is in possession of all the answers. In the hope of saving colleagues from this “Groundhog Day” existence, I am sharing the best of the warnings and advice that arose out of these conversations.

BTW: If you shared, don’t panic, this post has been cleaned of any identifying or proprietary information.

And finally, if you find the tone a bit fraught, it’s because the industry is pushing the envelope on many fronts, resulting in many challenges and corresponding amounts of introspection.

Here we go! The headline-only version is 1st, the full version is second

Headlines Only

1. Modular Construction is not a fix for the shortage of construction managers.

2. In the start-up phase, Modular Construction is not a fix for the skilled labour shortage.

3. Mass customization gets me a ton of customers, but my costs soar.

4. Only go down the modular road if you have the backing, patience and discipline to make it past the 3^rd year.

5. Multi-lateral construction process moved under a roof doesn’t deliver conflict-free, lower cost construction.

6. You can't boil the ocean with a candle.

7. “Perpetual Crisis” is never a successful management technique.

8. Mistakes happen. It's how you deal with them that matters.

9. Software isn’t the answer to all life’s problems.

10. Precision is now free.

11. Upstream suppliers who embrace modular and panelized execution are flourishing.

12. Few companies in the modular space have in-house R&D, some are all R&D.

13. Don't automate a process you aren't already good at.

14. The correct production solution is often not obvious.

15. You can be a modular builder without a "railway" production line.

16. Rain falling on a partially finished building is one of the industry’s biggest challenges.

17. Don't, no never ever don't leave a problem for the site.

18. You are running a business. OPEX and CAPEX apply.

19. A rendering or prototype does not a technology make.

20. Values are more important than ever.

Full Version

1. Modular Construction is not a fix for the shortage of construction managers. Conventional or Modular; Construction ain’t easy. To be both faster and better than conventional construction, modular construction requires managers that are both more skilled and more disciplined. In other words: By itself, modular construction is not a fix. Like any other methodology, it must be delivered by staff with the necessary skill sets.

On top of this, decision-making for modular construction has to take place faster and earlier. Because of this “need for speed”, the effect of mistakes will multiply rapidly. If you don’t have the experienced staff to swamp issues with management as they arise in real time, the consensus opinion is that you are better off subcontracting your projects to a conventional builder and going for a beer.

2. In the start-up phase, Modular Construction is not a fix for the skilled labour shortage. As with all manufacturing, the start-up phase of a modular enterprise requires highly-skilled personnel in fields such as product design, plant design and commissioning, production planning and production troubleshooting.

To get rolling, you need at least one experienced tradesperson willing and able to teach every production team (MEP, drywall and paint, framing, roofing, etc.). If you can’t access high-functioning staff during the start-up phase, you won’t be able to create a high-functioning company.

So be sure to locate your plant in an area with high educational attainment. Keep in mind that those older workers so often mentioned as an untapped labour pool want creature comforts like good lighting, clean washrooms and safe working conditions. If you are a progressive employer, the enjoyment factor will work to your advantage when searching for experienced staff.

3. Mass customization gets me a ton of customers, but my costs soar. Offering architects and customers whatever they want is fine for dimensionally neutral features such as the colour of cabinet doors and floor coverings, but it’s infinitely more difficult to offer a panoply of amenities, floorplans, facades, roofs, rated assemblies, seismic and hurricane ratings, etc.

The inevitable result of offering all things to all people in all places is best summed up by the Business 101 truism: “We lose money on every job, but we make it up on volume”. The most successful / highest volume modular companies operating today are focused on a narrow range of module form factors and market verticals, and their factories are designed around those products.

4. Only go down the modular road if you have the backing, patience and discipline to make it past the 3^rd year. Because Modular execution requires detailed planning and it requires setting up and running an efficient factory, the first batch of projects delivered by a Modular Construction start-up can be more work than building the same jobs conventionally. The landscape is littered with companies that ran out of gas and died 3 years in.

5. Multi-lateral construction process moved under a roof doesn’t deliver conflict-free, lower cost construction. The most workable approach to modular construction puts as much of the project team as possible on the same side of the table and repeats the process again and again. In-house resources are the only way out of the Groundhog-Day cycle of rebuilding, re-educating and then dismantling teams. Outsourcing key roles reduces agility and de-skills a company leaving it dead from the neck up.

“Having everybody on the same side of the table” means working with an architect right from the massing and site-planning phase, performing detailed design, routing and conflict detection in collaboration with a production-savvy design team, coordinating like mad and performing the work of all trades with in-house forces. And then repeating the cycle with the same team. If you are your own client and self-performing the work: Even better.

6. You can't boil the ocean with a candle. It’s great to be ambitious. But when the investors and / or a commissioned sales force are given the reins of a modular start-up, they often push the new company to undertake multiple, large, overly complex projects before the organization has the necessary capacity to execute. This pressure can arise from an impending IPO or unrealistic projections developed to attract investors.

Don't fall into this trap and take any job just to have a job. Be choosy and critical regarding project scope, read every line in the contract and above all don’t over-reach. See lesson 7!

7. “Perpetual Crisis” is never a successful management technique. If everything is a priority, then nothing is a priority. Never let your pipeline get to the point that staff is overwhelmed. The result is panic hiring of expensive or unqualified staff, un-bid subcontracting, massive overtime, detailing and manufacturing errors and the inevitable damage to reputation and morale.

8. Mistakes happen. Its how you deal with them that matters. Ever since the phone was invented, a culture of secrecy has been hard to maintain and it’s not getting easier. Don’t make a bad situation worse by promoting a troubled project while pretending everything is fine. Customers and potential partners doing due diligence quickly figure out what is going on and end up questioning your honesty.

Better to be candid, reference failures openly as “Lessons Learned” and highlight your company’s capacity to improve.

9. Software isn’t the answer to all life’s problems. As many of you know, I am a strong advocate of MCAD, whatever-D BIM, single-source-of-truth models and numerous other hip acronyms and concepts. So, I am well aware that software can dramatically improve design process and purchasing management, then produce the “cut files” that run CNC machines and robots which churn out precision parts.

BUT…..the latest software isn’t a panacea and no software can solve management issues. As a corollary, 3D printing sure is cool, but the deposition rates for 99% of materials used in construction are still far too slow for commercial applications. (However, patterns and tooling components are well suited to 3D printing.!)

Keep in mind that the modular industry functioned 40+ years ago when all drawing was done manually.

10. Precision is now free. It’s less work to make an accurate component than a crooked one, it’s easier to install precise components in a precise volume, and it’s easier to assemble a building composed of accurate modules. Widely-available CNC machines are built to deliver tolerances measured in thousandths of an inch.

Don’t believe production managers that tell you precision is expensive or impossible to achieve.

11. Upstream suppliers who embrace modular and panelized execution are flourishing, and those that don’t are flatlining. Ask any light steel or wood framing company which has shifted to panelizing its product! In North America, panelization continues to outstrip volumetric modular as a delivery method.

12. Few companies in the modular space have in-house R&D, some are all R&D. A remarkable number of companies still make their products the same way they did when they were founded, in some cases 40-50 years ago. On the other hand, at least 5 start-ups are 3D-printing tiny houses with advanced materials like ceramics. Somewhere in the middle between trailing and bleeding is a good balance.

13. Don't automate a process you aren't already good at. If you don’t thoroughly understand a process, it’s likely that those expensive robots you bought are making the same mistakes, only much faster. And the real bottlenecks have been overlooked.

Before automating a process, build some “manu-matic” (manually activated) fixtures, get some product out the door and learn how the process works. When you have achieved smooth workflow and a good product, look at the statistics you have been gathering (you are gathering statistics, right?!) study the process, look for bottlenecks and focus automation where it will do the most good.

14. The correct technical solution is often not obvious. I heard many examples of “misdirection” arising out of rushed or ill-informed troubleshooting. EG altering welding sequence or part detailing can mitigate welding-induced distortion, which is an easier implementation than “10X more clamps". Maybe welding isn't even the right process and fasteners would be better? In-person, shop floor meetings that bring designers and production staff together in an ego-free zone are the best way to find root causes and develop a focused fix.

Above all don't be afraid of changing course when the evidence is clear that a course change is needed. Just because you've invested $10m and baked in an approach doesn't make the approach the right choice.

15. You can be a modular builder without a "railway" style production line. The eternal debate between "bring the work to the workers" vs. "bring the workers to the work" rages on. A typical plant produces modules with a spectrum of complexities. Modules with kitchens and bathrooms take longer to complete than living / bedroom modules. Confined linearly on tracks like beads on a string, all modules move at the same speed and have to dwell at all stations.

If time-to-shipping is dictated by the modules that take the longest to complete....well, you get it. It's just as viable to assemble product in stations and leapfrog finished modules to shipping using a crane or to tow them on wagons or wheels with a forklift. You are still a modular builder!

16. Rain falling on a partially finished building is one of the industry’s biggest challenges. Rather than “hoping for a stretch of good weather” during building assembly, develop a detailed precipitation management plan and build or buy the right equipment to ensure water can’t get into a partially assembled building.

17. Don't, no never ever don't leave a problem for the site. Remediation at site is at least 10X more expensive than the same work performed in the plant. Ensure that staff are empowered to point out problems both in the plant and on the site, without being labelled troublemakers. If a problem is discovered, push stop and resolve the problem in the factory before you ship.

18. You are running a business. Just because modular is cool, doesn’t mean that old-fashioned concepts like OPEX and CAPEX are no longer relevant.

Because of the crisis in conventional construction, many start-ups and investors have gotten caught up in the hype around modular. Over enthusiasm often results in a failure to develop accurate costing and pricing models prior to spending astounding amounts of money and baking in a specific strategy.

As an example, “smart buying” can save 10-15% on material costs, but won’t save 30% and may result in oodles of inventory and a corresponding reduction in agility. Likewise, automation may reduce labour costs, but the CAPEX can add $4000 or more to the cost of every module produced, wiping out labour savings.

Successful manufacturers understand their costs and study well-informed and detailed financial and production models prior to making significant investments.

19. A rendering or prototype does not a technology make. A great rendering or a beautiful prototype of a super-hip apartment can be very useful in attracting investors and potential customers, but it doesn't mean the intended structure can be permitted or even built. Before investing in a modular start-up, have a critical look at the compliance and deployment roadmap, then build a realistic cash flow study based on the barriers to deployment that the venture will face.

20. Values are more important than ever. Modular Construction is an industry of the future and is just as reliant on knowledge workers as Silicon Valley. Unemployment is at historic lows. If you are a member of the executive team, the most important task you have is to attract and retain a skilled and dedicated workforce.

The best available personnel will come from many countries, they will have a wide range of lifestyles, and if you are very lucky, they will have ideas you need to hear. And they will consider themselves worthy of respect and know that they are mobile.

In this context dictatorial, sexist, racist, nativist or homophobic values are not just wrong, they are counterproductive. Running a Dilbertian dystopia has never been less viable than it is today.

Copyright Julian Bowron 2019

September 2019 Author: Julian Bowron, President, Vector Praxis

Following is a list of common deficiencies seen in stainless steel work, likely causes and suggested fixes.

Parts, especially parts made from different thicknesses of material, look very different when viewed side by side on the jobsite

• The polished surface is actually composed of parallel rows of tiny grooves whose sides reflect light. If the grooves on two adjacent parts are slightly different in depth, or in length, or have a different underbrushing (1st pass) they reflect light differently and so they look slightly brighter or darker from a given angle. As a result a ½” (12mm) baseplate will look different than a panel made from 16G sheet.

Two pieces of the same material from the same supplier don’t match when installed side by side

• The brushing proceeds from one end of the coil or sheet to the other, so the sanding medium is duller at the end of a run than at the beginning, resulting in a variation in texture, especially when seen from a distance.

• Specify that material be sequenced (in a process similar to the colour matching of wood veneers)

Brushing goes various ways on a group of related surfaces

• The drawings did not indicate which direction the brushing should go, so the supplier saved material by rotating parts during the layout for cutting, or “nesting”.

Oil canning is visible on the face or on edges, especially from a distance

• Just like glass, a sheet of stainless steel with a slight warp acts like a lens with a very long focal length. This means that what looks good close up can look awful from across the street. The solution is to use thicker material and ensure edges are “controlled”, meaning framing, brake-forming or hemming.

• As a bonus, brakeforming tends to pillow thin sheet in the middle or unify it, producing a more appealing and uniform reflection. This principal is developed to an art with rigidized metals.

There is an obvious variation in appearance on the face opposite studs, at welded corners or along welded seams

• Re-polishing after welding is difficult on #4, use XL Blend S instead. If welding is not actually required, it is best to substitute mechanical fasteners or adhesives.

• Polishing is incomplete or the worker has used the wrong media and / or tools, so more work might fix the problem.

• The polishing media has to be carefully chosen to match the brush on the sheet and applied in a sequence using tools with characteristics similar to the machine that originally applied the factory finish to the material.

There are dull or intermittent marks running the length of parts parallel to the edges

• During processing steps such as brakeforming, no die protection was used and the tools on the machine compressed the fine ridges of the brushing. Use a hard urethane sheet over the dies.

Visible welds mar the appearance and / or blue / gray heat marks are visible

• Eliminate welding if possible, or ensure welds are positioned in accessible locations so they can be ground and polished out.

• Use pickling paste to eliminate the thermally-induced discolouration

Corner radii are too large or too small, or there are undesirable facets

• Each brakeformed radius / sheet thickness combination requires a tool worth thousands of dollars. Very tight radii also require the sheet to be scored on the rear face with specialized equipment.

• Large radii can be rolled on specialized intermittent rolls by skilled operators but are often “bumped”, resulting in facets.

• Require samples in the contract and confirm the fabricator can deliver what you want.

Brown rust marks appear on the work a few weeks after installation, and / or the work corrodes badly to the point of failure

• Alloy 304 is not actually “stainless” in an outdoor, curbside environment and will develop stains under normal urban exposure. To be truly stainless, the steel must have some Molybdenum, designated by the middle number, meaning that 316 is acceptable.

• Another possibility is contamination by mild steel particles which occurred in processing or in the suppliers shop. Sources include forklifts used directly on plate without a pallet, nails in skids, grinders which had previously been used on steel, wire wheels, etc. Unfortunately, it is not possible to remove contamination that has worked in to the surface, but light contamination due to dust may disappear in a few months of steady rain.

• From Anton Nelson of Dewhurst Macfarlane and Partners: The reason some stainless “rusts” is because the heat of the welding causes the chromium in the stainless to come out of solution and form chromium carbide. The chromium carbide is prone to corrosion which is why you sometimes find rust marks around welds in stainless steel parts. It is rust, but not iron oxide, the chemical which is the result of rust in non-stainless steels. I don’t think chromium carbide has the same detrimental structural effects as iron oxide, but it is unsightly. The way to avoid this problem is to use low carbon stainless such as 304L or 316L if possible when you are doing welded architectural connections.

Brushing does not extend in to corners

• This is a design problem, as the tradesman must have clear “on and off” access to a surface to be polished. Inside corners etc, are impossible to reach, so parts requiring polish to the inside of corners must be mechanically fastened so parts can be polished before assembly.

The author would like to thank Bill Butt of Excelsior Metal Polishing, Anton Nelson of Dewhurst McFarlane and Partners and Davide Parisi of Vaughan Metal Polishing, all of whom generously shared their knowledge.

Copyright Julian Bowron 2019

October, 2019 Author: Julian Bowron, President, Vector Praxis

Two years ago I attended the opening of a 50-story “LEED Gold” office tower. After the mayor and officials left, a group of us walked the ground floor of the brand new building. Because we were all in the fabrication business, we couldn’t stop looking at the details. I found a shocking number of deficiencies in the stainless steel work.

Following is a partial list of the defects I saw on this one jobsite:

1. Stainless elevator door frames with visible polishing marks at the miters.

2. Cove base with visible polishing marks at every splice.

3. The stainless steel grate under the revolving doors was made of unsupported 1/8” stainless with wide spans in the slotted pattern, so the weight of a high heel depressed it until it was permanently bent.

4. Flat stainless column covers on the exterior appeared badly “oil canned”.

5. Curved column covers on the same column had a checkerboard appearance.

This article is intended to help designers avoid these issues and provide background to facilitate troubleshooting in the field.

Defects in stainless steel can't be camouflaged by coatings

Stainless steel has a problem in common with many high-end materials: Stone, stainless steel, bronze and glass seldom have a thick or opaque coating, so the effect of every process is visible, making it hard to hide bad workmanship. Worse, polished stainless steel sheet is shipped with masking, so defects are not visible until the masking is striped. Because the work is expensive and hard to fix, when bad workmanship makes it to a site despite a designer’s best efforts, suppliers are reluctant to admit fault. It then falls to the designer to “call BS” or to identify the source of a problem and suggest a fix.

The most common stainless specification is “304 Stainless Steel, #4 brushed”. Behind this spec lie pitfalls that can result in a job which complies, but looks rough, patchy and just plain old cheap. Workmanship aside, specifying exotic finishes which are subject to interpretation, like “angel hair” or “non-directional”, can produce even more variable results. A list of some things that can go wrong with stainless steel work is found below, together with the likely causes and suggested fixes, but first, a brief explanation of the stainless finishing process is required.

Both stainless steel sheet and plate are rolled in lengths and widths much greater than the material which eventually finds its way in to most fabrication shops. In a stainless mill, billets weighing 25,000 lbs (12,000 kg) or more are rolled hot, then the raw material enters two different streams; Plate and sheet. Material destined for sale as plate is further rolled to thicknesses from ¼” (6mm) to over 6” (150mm) and sizes from over 10’ x 20’ (~ 3m x 6m) to 4’ x 8’ (~ 1.3m x 2.8m).

Surface quality of plate varies enormously and most plate suitable for use in polished architectural applications comes from modern mills in North America or Europe. Low quality plate best suited to industrial applications can have significant warpage, roll marks and improperly alloyed iron, which causes staining of finished work and contamination of processing equipment. Even high quality plate can meet some specs with visible “rescue grinding” where defects were removed.

It Starts With Coils

Stainless sheet suited to architectural applications is produced and stocked in coils 4’ (1.2m) or 5’ (1.5m) wide and come with a “B” or “Bright” finish which looks like a cloudy mirror (2B is most common). Coils are “sheeted” at service centres, meaning sheet lengths are only limited by minimum orders and the fabricator’s ability to handle the material.

This sheeting process is why it is expensive to specify a small number of parts 8’ long with a 1” return all around, because making the parts requires material 8’ 2” long, meaning almost 2’ will be lost from the next stock length up, which is a 10’ sheet. By contrast, if thousands of parts are involved, the fabricator can order the material sheeted to the length required, resulting in less waste.

Polishing or "Brushing"

If a polished finish is required on sheet (the industry uses the term “polished” for all finishes, including “#4 brushed” and “mirror polished”) the finishing is applied on the entire coil in a continuous process using a multi-headed belt sander, prior to cutting in to sheets. The masking is applied immediately after the polishing, before the sheets are stacked. If a finish is required on plate, the finish is applied by special-purpose sheet-fed machines, or by a manually-operated overhead gantry polishing machines. As a result of the manual processing, high-quality finishes on plate are much harder to achieve than on sheet and supplier qualification is paramount.

The key decision to make when specifying a finish is whether the stainless will be used as whole sheets, without the need for re-finishing, or whether it will be joined by welding or otherwise worked and must be refinished after fabrication. This choice is required because a #4 finish is cheap, but is created by machines whose characteristic long surface markings cannot be restored or blended by hand, making invisible re-work impossible.

Only finishes such as Excelsior Metal Polishing’s “XL Blend S” or Vaughn Metal Polishing’s equivalent can be duplicated with hand tools and should always be specified for work which requires blending. The same principals apply to round and square tube: Chose #4 if the parts will be cut and used, but upgrade to XL Blend S if polishing is required after fabrication.

Non-Directional or "Angel hair" Finishing

Non-Directional or Angel Hair finish is commonly specified for stainless steel paneling used in high-traffic public environments, because it can be restored on site by semi-skilled workers using an orbital sander. Applying this finish directly on raw 2B sheet produces a dull result, so this finish looks best when applied over a #4 finish. To produce the 25 tonnes of 316 stainless plate with a non-directional finish we needed for a job in the Jamaica Airtrain Station, our polisher (Vaughan Metal Polishing, VMP) started with a series of progressively less aggressive grits culminating in a #4 pass, which removed the scale and leveled the surface.

We saved VMP an enormous amount of time applying the random final pass by rigging 6 heavy-duty air-powered random orbital sanders to a frame. The VMP polishers moved it by hand in a carefully timed pattern over the 20’ (6.5m) plates to apply the uniform and repeatable non-directional finish. We knew that since the oscillating radius of the sanders used was ¼” (6mm), a similar sander could be used for repairing damage on site.

Because plate can vary in surface quality and be damaged by handling, when writing the specification for polished stainless work that will be fabricated from plate, it is worth including a line requoring the fabricator to select the plates for their finish quality, so they cannot claim that finish defects are the result of industry norms. It is also advisable to suss out whether the supplier is accustomed to handling materials with delicate finishes, since a correct set up is time-consuming and expensive to implement.

Celebrating Defects is OK

Finally it is worth noting that architects led by Frank Gehry, have sometimes chosen to celebrate defects such as the wrinkled appearance that results from wrapping flat metal sheet around compound curves. Making a positive statement with a ”defect” in workmanship, is actually a very difficult stunt to pull off and is not recommended for amateurs!

Copyright Julian Bowron 2019