As new home sales decline across the country, solving the trade labor shortage has become a lower priority for most production homebuilders.
Most builders recognize that the strength of their production apparatus atrophies the longer it lies dormant, but how many are taking the necessary intermediate steps before they determine how to restore its production capacity?
Many optimistic innovators are developing robotics, advocating for immigration reform, designing methods for offsite construction or building the infrastructure to train the next generation of tradespeople.
These efforts are necessary and noble, but proponents of the Theory of Constraints might note that they represent the fourth step in Eli Goldratt’s Five Focusing Steps: identify, exploit, subordinate, elevate and repeat.
How might Goldratt view the situation if he were alive today?
Reality check at the jobsite level
Even our application of the first step of 5SF would be considered inadequate.
Sure, we’ve identified that labor is constrained, but we haven’t developed a reliable method to quantify our labor market’s capacity. We don’t know what it’s capable of producing at any point in time. This has significant implications for later steps.
Every hour a skilled laborer spends on work other than constrained work permanently reduces the system’s capacity. In 5SF terms, “exploit” means removing all waste from a constraint.
Fragmentation creates process waste through context switching and windshield time. A framer who builds the same familiar plans in a single community tends to perform better than one who constantly shifts between builders. An excavator completing a one-hour task can get more done with a geographically optimized route.
Process and information failures cause dry runs, rework and late changes. Inconsistent cadencing decisions can make it impossible for trades to maintain consistent crews that achieve better first-pass yield.
The purest form of exploitation would be a production machine designed around what trades can actually produce. Builders would need systems that support a shift from trades chasing dates to builders planning around dedicated trade capacity. Trades know how many sheets of drywall their crews can hang in a day.
The paradigm that aligns with optimal production capacity is one in which builders ensure there’s enough house to supply the crews, not enough crews to cover houses.
Understanding a starts pipeline
Subordinating to the constraint, 5SF’s third step, is legitimately difficult to execute. Subordination means releasing only as many starts as the production system can handle. This decision can be made only when division leaders accept the unpleasant reality that cramming more starts into a constrained system will not increase closings.
The system can support only what it can handle, and starting anything beyond that capacity only increases your cycle time, your WIP, and your likelihood of being stuck with excess inventory when the market inevitably winds down again.
It’s impossible to make reasonable starts decisions without first identifying the precise capacity of your production apparatus.
Notwithstanding the incredible pressure divisions face to meet their closing commitments, subordination faces a more practical problem: it shares its labor resources with every other builder in its market.
Even if they’ve already done considerable exploitation work to make their jobs the most attractive place for trade partners to send consistent crews, they’re still subject to being knocked around by builders who are less inclined to subordinate their own starts pace.
Business risk versus competitive risk
At some point, the trades risk losing the business of the builder who practiced less self-restraint in their starts pace and fell way behind on their schedules. That builder may pull crews away from consistent work to get the errant builder caught up.
The continuing trend of builder consolidation amplifies this issue. The largest builders have always unwittingly benefited from smaller builders, who serve as the stalking horse for their own cadence inefficiencies.
With fewer builders in the game, there will be no one left to absorb the impact of inconsistent starts pace. The remaining builders will feel the impact of their decisions in the form of elevated cycle times, as the trade base is forced to build a larger buffer to handle the variance in incoming work.
The shared-resource problem plagues subordination efforts, creating a moral hazard for which no solution has yet emerged. The eventual solution is likely to follow one of two paths: mandated coordination or self-selection.
Mandated coordination might resemble OPEC or the FAA’s airport slot-allocation system. Perhaps HUD or another government entity mandates the creation of a starts-pace cartel to ensure a consistent housing supply. This approach would undoubtedly ruffle feathers.
Self-selection’s upside
The self-selection approach would be fueled by improved data on production capacity from next-generation production systems.
Just as demand-based pricing for lift tickets encourages skiers to self-distribute across dates and helps control capacity on the mountain, scheduling tools built around crew resource allocation data could show builders the cycle-time cost of starting houses in one week versus the next.
Reckoning with data reality vs. intuition
5SF’s fifth step is to repeat the process and prevent inertia from becoming the new constraint.
Currently, the sales pace is the constraint. Our production machine has been kept on life support by BTR and brief bursts of spec-home optimism, but without tools to test its capacity, we’ll have no idea what it can produce when sales return. We’ll likely still believe we’re constrained by sales when we reach the maximum capacity our neglected machine can handle.
Our industry’s approaches to addressing the labor constraint are genuinely unique, creative, and inspiring. But we should bring that spirit of innovation to the intermediate steps we can tackle now.
