House of Design, which ignited the component industry enthusiasm for robotics, has tragically gone out of business, but not all is lost. We have learned that robots can pick plates very effectively among other benefits, but we also learned that fact 15 years ago from Jim Urmson’s TCT Automation, and we can see it in action today on Randek’s Auto-Eye system. What HoD added was bottom-side and top-side pre-plating, which eliminates the labor of picking and placing plates and mitigates inherent plate placement inaccuracies. What HoD could not do when placing webs, however, is ensure that they would stay in the desired location until abutting webs were placed against them. Webs sometimes moved because they rested precariously on the upward-facing teeth of the back-plated chords. This inaccurate positioning could result in non-tight-fitting joints that had to be manually remedied. HoD could have fixed them in the correct location by impaling them on spikes, as Trussmatic does, forming tighter joints. However, the HoD process was still superior to doing it by hand.
What seemed out of character for a robotic process were the manual after-the-fact adjustments of joints and the hammering of plates; however, these actions take a fraction of the time that a manual process takes. And, of course, the inability of the end-of-arm tools to handle short pieces, like wedges, seemed to be a drawback, but the gantry operator maintained a stock of plates to deal with this shortcoming and took care of this not much differently than would be done manually.
Recent media articles seemed to bolster the effectiveness of the HoD system, including the following (parenthetical phrases taken from articles):
- June 29, 2022, HBS Dealer: BFS to add “8 more locations by end of 2023” (to the two already in place)
- December 18, 2023, House of Design Bulletin: “[Robotic] facility produces 158 board feet per man hour, well over the industry average of 100,” quote from Mike Farmer, BFS President of Commercial Operations
- June 11, 2024, SBCA Industry News: “Atlas Engineering Products Investing $7M in [3 HoD lines of] Robotics”
- June 25, 2024, HBS Dealer: “We’re going to be installing a second robotic line later this year,” quote from component manufacturer B.J. Louws.
Based on my observations of three Builders FirstSource facilities, I believe that the 158 board feet per man hour is achievable, so why did HoD fail?
Perhaps we missed the significance of the other two important articles in SBCA Industry News, because it seemed obvious that HoD would need an injection of capital just to expand their staff and inventory to fill the BFS orders:
- September 13, 2021: “Thomas H. Lee Partners Acquires House of Design”
- March 19, 2024: “Cullinan Holdings Acquires Majority Control of House of Design.”
From ex-employees, we have learned that one-third of employees were cut after the T. H. Lee purchase, and one-half were cut after the Cullinan acquisition. While this is unsubstantiated by management, it fits the experience of the Dorbyl acquisition of Alpine in 1998 for $50 million. Dorbyl quickly made huge staff cuts, and eventually flipped Alpine for $158 million. The reduction of staff is one of the only ways an unknowledgeable owner can stanch a cash drain, and is one of the inherent risks of a private equity purchase. Alpine, however, was able to survive because it had scores of people, like Dave McAdoo, who had decades of industry experience, while HoD had only a small handful of people who had one decade of experience. And in general, the HoD debacle underscores Warren Buffett’s disdain for private equity’s practices, especially investing in businesses that the buyer doesn’t understand. What could either T. H. Lee or Cullinan possibly know about robotics, except that it sounds sexy and the component industry is reputed to be in the Stone Age? Did these investors understand the time required to bring in sales (and cash deposit inflows) of new technology, or the diligence (and cash outflows) necessary to refine the technology?
I know how long it took to break even on MiTek’s first computerized component saws and first linear saws, and how unsteady sales are in the early years of adoption. I can only speculate that breaking even on a complete robotic system takes many times the support and expense as the standalone conventional truss equipment that I sold. Early on, it takes a large infusion of cash to fund such lengthy development. And infusing boatloads of cash is not the normal practice of private equity firms.
It is my opinion that the House of Design roof line was a viable and productive product that would have succeeded given adequate financial support and stable management, but HoD’s customers would render the best judgement. I would also suggest that this experience, painful as it may be, was just another step in advancing the art of building such a variable and highly customized product. The industry learned much about the future of robotics from HoD and continues to learn from Trussmatic, Randek, and other viable robotic systems. For my observations after visiting several robotic truss-equipped facilities, please see some of my previous articles in The Advertiser: