Newark, NJ — Investors in several warehouse manufacturing operations across the mid-Atlantic industrial corridor are reportedly expressing excitement after restructuring their labor model to rely almost entirely on day laborers, a transition internally described as a “workforce modernization” and externally described, by the workers who no longer exist in any formal sense, as “the new thing where nobody gets your name.”
The appeal, according to investor communications reviewed by The Externality, is straightforward:
No long-term obligations.
That phrase appears, in some form, in seventeen separate investor decks prepared over the past eighteen months. It appears on slide three. It is always slide three. It is never elaborated upon, because elaboration would introduce complexity, and the phrase is valued specifically for its absence of complexity. The workers, in contrast, have been elaborated upon extensively — in Excel.
The Model
Under the new system, the labor architecture has been pared down to what consultants at the Mercer-Ashworth Group call its “minimum viable obligation.” Workers are hired daily. Shifts are assigned as needed. Employment ends… automatically.
The elegance of the model lies in its subtractions. There are no contracts. No benefits. No continuity. A worker reporting to a facility at 5:47 a.m. on a Tuesday is a worker who, by 5:47 p.m. that same evening, has no further relationship to the facility, the facility’s legal entity, the parent LLC of that legal entity, the parent LLC’s holding company, or the holding company’s offshore vehicle. By Wednesday morning, the worker may return. Or may not. The facility will not call.
An operations manual obtained from a logistics subcontractor in East Rutherford formalizes the arrangement with what one labor attorney called “almost religious minimalism”:
“The shift begins when the worker enters the gate.
The shift ends when the worker exits the gate.
Between these two events, we are acquainted.
Before and after, we are strangers.”
The document is unsigned. It is not, investors clarified, a contract. Contracts, they noted, defeat the point.
The Gate as Legal Instrument
The gate — an unremarkable roll-up door in a chain-link fence on a service road — has, under the new model, acquired the legal significance previously held by employment agreements, onboarding documentation, and the human resources function generally. Crossing it inward triggers what attorneys describe as a “temporally bounded relationship of mutual acknowledgment.” Crossing it outward dissolves that relationship without residue.
A regional operator described the architecture admiringly. “We used to have an HR department of forty-three people,” he said. “Now we have a gate. The gate is cheaper. The gate does not file grievances. The gate does not attend compliance training. The gate, at worst, sticks sometimes in the rain.”
The gate, it should be noted, is also not covered by workers’ compensation.
Management Perspective
A full-time manager at the Newark facility, one of four remaining salaried positions in an operation that employs, on any given day, between 180 and 340 people, reportedly stated the organizing principle in language that investors later had framed and hung in the corporate boardroom:
“We can let them go at any time.
We don’t have any long-term relationships with them.”
The statement was followed by what sources present described as visible satisfaction — a particular arrangement of the facial muscles that labor economists have begun tracking as a leading indicator for capital efficiency. It is the satisfaction, one observer noted, of a man who has successfully removed a splinter. The splinter, in this metaphor, is the workforce.
The manager, who asked not to be named because he has himself been warned that his position is under review for “structural redundancy in light of gate-based workflows,” elaborated:
“When you don’t know someone’s name, you don’t have to feel anything when they’re gone. It’s not cruelty. It’s just logistics.”
He paused. Then added:
“I used to know names. It was harder.”
The Training Materials
Internal management training for the new model, prepared by a vendor called Relational Minimization Partners, emphasizes what the firm calls “deliberate forgetting.” Managers are coached to refer to workers by role rather than identity — “the forklift,” “the scanner,” “the third pick station on line B” — and to resist the natural tendency to acknowledge returning laborers by face.
“Recognition creates obligation,” one slide reads. “Obligation creates exposure. Exposure creates liability. Liability creates cost. Cost is the enemy. Therefore: do not recognize.”
Managers who fail the program’s periodic assessments are themselves replaced — though, given that most managers under the new model are also employed on short-term engagements through a rotating consultancy, “replaced” is understood to mean “not re-engaged next Tuesday.”
The Vendor Layer
The administrative architecture that makes the disposable workforce model possible is not, contrary to impression, a simplification. It is a proliferation. Where the older model required one relationship — between the worker and the employer — the new model requires between four and eleven relationships, arrayed in a structure that labor attorneys have taken to calling “the cascade.”
At the top of the cascade sits the operating company, which owns the facility. Below it sits a logistics subsidiary, which operates the facility under a service agreement. Below the logistics subsidiary sits a staffing firm, which supplies the labor under a master services agreement. Below the staffing firm sits, in many cases, a sub-staffing firm, which supplies the actual workers under a subcontract. Below the sub-staffing firm sits, occasionally, a further sub-sub-staffing firm — typically a single individual with a pickup truck and a clipboard, operating as an LLC out of a P.O. box in a strip mall — which supplies the actual actual workers on any given morning.
Each layer of the cascade collects a margin. Each layer of the cascade disclaims, in its contracts with the layer above and the layer below, any responsibility for the conduct, wages, injuries, or legal status of the workers at the bottom. The workers, by the time the paperwork reaches them — which it often does not — are technically employed by an entity with a name like Prime Source Industrial Solutions III LLC, which was incorporated in Delaware in 2023, has no employees except the workers, no office except the P.O. box, and no assets except an accounts receivable balance owed to it by the layer above.
When something goes wrong — and something does, approximately 2,847 times per year across the sector, according to OSHA injury data from the prior reporting period — the cascade performs what one plaintiff’s attorney described as “a kind of administrative vanishing trick.” The operating company points down at the logistics subsidiary. The logistics subsidiary points down at the staffing firm. The staffing firm points down at the sub-staffing firm. The sub-staffing firm points down at the sub-sub-staffing firm. The sub-sub-staffing firm points at a P.O. box, which does not point anywhere, because P.O. boxes cannot point.
“The elegance of the cascade,” one corporate attorney explained, “is that no single party is ever in a position to be held accountable, because no single party ever had the worker as their employee. Accountability, like employment, has been flexibly distributed.”
Flexible distribution of accountability, it should be noted, is not a term used by workers. Workers, when something goes wrong, use different terms. These terms are not reproduced here.
The Margin Mathematics
At each layer of the cascade, the hourly rate paid by the layer above is larger than the hourly rate paid to the layer below. The difference is retained as margin. The operating company pays the logistics subsidiary $42 per worker-hour. The logistics subsidiary pays the staffing firm $34. The staffing firm pays the sub-staffing firm $26. The sub-staffing firm pays the sub-sub-staffing firm $22. The sub-sub-staffing firm pays the worker $17.50, which is, by a slim margin, the state minimum wage.
The difference between the $42 entering the top of the cascade and the $17.50 arriving at the bottom — $24.50 per worker-hour — is distributed across four layers of intermediation. None of these layers, on paper, does anything except transfer a worker from the layer below to the layer above. One layer maintains, as its sole productive asset, a spreadsheet. Another maintains a phone. The sub-sub-staffing firm, as noted, maintains a pickup truck.
The accounting term for the $24.50 that does not reach the worker is value added. The value added, under the new model, is the cascade itself. The worker is the pretext for the cascade; the cascade is the product.
The Advantage
Investors highlighted several benefits of the restructured model, presented in a quarterly earnings call that analysts described as “the first earnings call I’ve ever heard that sounded like it was conducted from inside a meditation retreat.” Executives appeared unusually calm. One was sipping tea.
The principal advantages, repeated across multiple investor communications, are maximum flexibility, minimal liability, and immediate cost control. These three properties, when combined, produce what one portfolio manager called “the holy grail of industrial capitalism — a workforce that is simultaneously present and legally absent.”
An internal memo distributed to capital partners, which was not intended for external circulation but was forwarded so widely internally that it effectively self-circulated, contains a single sentence that has since been adopted as the operation’s unofficial motto:
“Labor should behave like inventory.”
The sentence has been laminated. It is on a card. Several managers carry the card in their wallets. When asked why, one said, “Because it reminds me that this is not personal.” When asked what “this” was, he declined to specify.
Inventory as Metaphor, Then Directive
The inventory comparison, originally introduced as a figure of speech, has since been operationalized. Workers are now scanned at the gate using the same barcode infrastructure previously reserved for pallets of merchandise. The scanning does not record identity — merely presence. The system logs a count. At shift’s end, the count is adjusted downward. At shift’s start, the count is adjusted upward. The count is never the same count twice.
One IT contractor, tasked with integrating the labor-tracking system into the existing warehouse management software, described the project as “philosophically confusing but technically trivial.” He continued: “The database doesn’t have a ‘person’ table anymore. It has a ‘presence’ table. A presence has a duration, a station, and an output. It does not have a name. It does not have a history. Yesterday’s presences are purged nightly.”
Asked whether this posed any concerns, he shrugged. “I just build what they ask for. I also work through a staffing agency. I’m not sure I’m in the ‘person’ table either.”
The Tradeoff
Analysts note, with the careful balance of tone characteristic of analysts who do not wish to be identified, that the model removes certain elements that older industrial operations had traditionally considered features rather than costs: stability, institutional knowledge, and workforce continuity.
Stability, in the older understanding, meant that a worker who returned tomorrow had also been there yesterday, and that the continuity between these two facts allowed the worker to become, over time, better at the job. Institutional knowledge meant that when a machine broke in an unusual way, there was someone in the building who had seen it break that way before. Workforce continuity meant that when a new worker arrived, there was someone to show them where the bathroom was.
Under the new model, none of these properties obtain. The machine, when it breaks, is consulted via a QR code that links to a help desk in a jurisdiction eleven time zones away. The bathroom is located via a laminated map. The worker who once knew the unusual way the machine breaks has not been re-engaged in four months and currently drives for a rideshare application, where he has also not been re-engaged, because rideshare applications have adopted a version of the same model.
What Is Introduced in Return
The model, in exchange for what it removes, introduces rapid scaling capability, cost variability, and reduced long-term commitment. These gains are real. A facility that employed 847 people under the old model can now employ 340 on a busy Tuesday and 12 on a slow Thursday, and the facility’s labor expense line on the Thursday income statement reflects this difference with the kind of honesty that the old model, burdened with payroll fixity, could never achieve.
Rapid scaling has been particularly celebrated. During peak season — a construct that has expanded over the past decade until it now occupies roughly 47 weeks of the calendar year — the facility can increase its headcount by 280 percent in under 72 hours. During the remaining five weeks, it can reduce its headcount to the three people required to turn on the lights. The lights, notably, are also now operated by a contractor.
Cost variability, meanwhile, has produced what one investor newsletter described as “the rarest of industrial achievements — a labor line item that is actually shorter than the facilities line item.” This was previously considered impossible. It is now considered, among certain capital allocators, the point.
When Something Happens
The question of what happens when something happens — when a worker is injured, when a forklift tips, when a hand meets a machine in a way the machine was not designed to accommodate — is handled by the model with a graceful efficiency that older industrial operations could only envy.
Under the older employment relationship, an injury triggered a sequence: report, documentation, workers’ compensation claim, medical care, rehabilitation, eventual return to work or, in cases of permanent injury, a settlement and an extended relationship of obligation between the employer and the former worker, sometimes lasting decades.
Under the new model, the sequence is different. An injured worker is escorted, or helped, to the gate. The worker exits the gate. The worker is no longer, as of that moment, present in the facility. The facility logs the departure. The facility’s presence table updates. The facility’s legal relationship to the worker is now the same as its legal relationship to any other non-present person: nonexistent.
The injury, if reported, is reported to the sub-sub-staffing firm — the man with the pickup truck and the clipboard. The sub-sub-staffing firm, being a single individual operating through a shell LLC, has in almost every case no workers’ compensation insurance beyond a minimum certificate obtained to satisfy the requirement that such a certificate be on file with the layer above. The certificate, when examined, typically covers the single individual. It does not cover the workers, because the workers, in the language of the certificate, are not employees. They are, again, presences.
“It’s not that we don’t care what happens to them,” said one logistics operations director, speaking on condition that his actual job title not be named because his actual job title does not exist on any organizational chart. “It’s that caring is not part of what we do. What we do is move packages. What happens to the hands that move the packages is handled by someone else. Not by us.”
Asked who, specifically, that someone else was, he paused. “I don’t know,” he said. “I’ve never had to find out.”
The Settlement Architecture
In the small number of cases where an injured worker successfully navigates the cascade to file a claim — a process that requires, at minimum, legal representation, an ability to read and write in English or Spanish, and approximately eighteen months of endurance — the settlement, when it arrives, is typically structured to close without admission of liability from any of the entities in the cascade.
The money, in such cases, comes from what plaintiffs’ attorneys have taken to calling “the fog.” The fog is a pool of contingency funds maintained, in aggregate, across the upper layers of the cascade, and distributed across settlements in a manner that no single entity in the cascade can be shown to have funded. The fog exists to make claims go away. It is, in the accounting documents that describe it, referred to as “operating reserves.” No reserve is ever formally linked to any specific injury. The fog is thick.
One attorney who has specialized in such claims for twenty years described her practice as “extracting small amounts of money from a cloud.” The cloud, she noted, is very large. The amounts she extracts from it are very small. The relationship between the two, she said, is not an accident.
Worker Reality
Day laborers, for their part, reportedly arrive daily, perform assigned tasks, and leave without expectation of return. This sequence — arrival, performance, departure, absence of expectation — has become the operational definition of the working relationship. Anything more would, according to management training materials, be considered “relational overreach.”
A worker identified only by the number pinned to his high-visibility vest — 412 — was interviewed in the parking lot of a truck stop near the facility, where workers gather in the predawn hours to be selected for the day’s shift. He commented:
“You work today. Tomorrow… maybe.”
Asked to elaborate, he paused for what witnesses described as a long time. Then he said:
“Every morning I stand here. They pick some of us. They don’t pick the others. When they pick you, you don’t ask why. When they don’t pick you, you also don’t ask why. The asking is not part of it.”
He adjusted his vest. The number 412, it turned out, was not his number. The vests are reissued each morning. He had been 287 the previous Thursday. He had been 601 in August.
The Selection Ritual
The selection process, which occurs each morning at the facility gate, has developed its own informal choreography. Workers arrive between 4:30 and 5:30 a.m. They stand in a loose crescent. A man with a clipboard — who is not himself an employee of the facility, but is sub-sub-contracted through a staffing vendor — emerges from a pickup truck and begins pointing.
Pointing constitutes selection. There is no verbal component. A worker who is pointed at steps forward. A worker who is not pointed at remains in the crescent. After approximately twelve minutes of pointing, the man with the clipboard returns to the pickup truck, and the unpointed workers disperse — some to other staging areas, where other men with other clipboards will point at them or not point at them for other facilities, and some simply home.
One worker, asked whether the system was fair, answered a different question. “It’s consistent,” he said. “Every morning the same. You know what to expect, which is that you don’t know what to expect. It’s a kind of knowing.”
The pointer, when asked his criteria, said he didn’t have criteria. He had a number. The facility told him the number. He pointed until he reached the number. “It’s not about who they are,” he said. “It’s about how many of them I need.”
Inside the Facility
Once inside, selected workers are issued a vest, a barcode, a station assignment, and a laminated card with basic safety instructions in six languages. They are not issued a locker, a lanyard, an email address, a timecard, or a name tag. They are not introduced to coworkers. They are not given a tour. They are given, instead, a rate — the number of units per hour expected at their station — and a direction, which is usually a gesture.
Work proceeds. Breaks are signaled by a bell. Lunch is signaled by a different bell. The end of shift is signaled by a third bell, after which workers are required to return their vest and barcode to a collection bin before exiting the gate. Workers who fail to return their vest are subject to a $47 deduction from their daily wage. Workers who return a vest that is not their vest are also subject to a $47 deduction, as are workers who return two vests. The system is, in management’s phrasing, “self-regulating.”
No worker is known to have returned two vests deliberately. This is considered a success of the system.
The Domestic Arithmetic
The model’s downstream effects on the home lives of day laborers have been the subject of three academic papers in the past eighteen months, two of which were rejected for publication on the grounds that their findings were “self-evident and therefore not novel.” The third was published in a regional sociological journal with a circulation of approximately 340.
The papers document what their authors, borrowing a phrase from one of their interview subjects, call “the Tuesday problem.” The Tuesday problem arises from the fact that rent, unlike labor, is not flexible. Rent is due on the first of the month. Rent does not wait to see whether a worker has been pointed at over the preceding thirty mornings. Rent assumes continuity. Rent is a relic of the older economy, in which relationships lasted.
A worker who is pointed at twenty-two mornings out of thirty earns, at $17.50 an hour and an eight-hour shift, approximately $3,080 before deductions. After the cascade’s various deductions — transportation to the staging area, occasional vest fines, staffing-fee withholdings that are not always legally permissible but are administratively convenient — the figure drops to approximately $2,640. Rent in the surrounding area for a one-bedroom apartment averages $1,890. What remains is $750 for the month. This must cover food, utilities, transportation, and whatever remains of a human life.
A worker who is pointed at fifteen mornings out of thirty earns, under the same arithmetic, approximately $1,800. Rent in that case is not paid. The worker moves. Moving, in such cases, is typically lateral — to a room in someone else’s apartment, to a cousin’s basement, to a motel that bills weekly. Lateral moves produce what sociologists call “residential churn.” Residential churn produces what sociologists call “instability.” Instability produces what the model’s proponents call “flexibility.”
“I have three addresses,” said one worker, asked during the pre-dawn wait for selection. “I sleep at one, I get mail at another, I tell the staffing agency the third. They don’t talk to each other. It works.”
Asked whether this arrangement was his preference, he looked at the interviewer for several seconds without answering. Eventually, he said, “Preferences are not something we have. We have arrangements.”
The Children
The question of how day laborers with children manage child care under a schedule that is never known more than twelve hours in advance has been the subject of approximately no academic papers, because the answer — that they do not manage it, and that something gives, and that what gives is usually either the work or the children — is considered too obvious to merit study.
A worker named Gloria, who agreed to be interviewed on condition that her last name and the name of her employer not be reported, described the arithmetic. Her daughter, eight, is in the third grade. School ends at 2:47 p.m. The bus arrives at her apartment at 3:14 p.m. If Gloria is pointed at in the morning, her shift ends at 5:00 p.m. at earliest, 7:30 p.m. during peak volume. Between 3:14 and Gloria’s return, the daughter is, in the technical language of child welfare, unsupervised.
Gloria described a rotating arrangement involving a neighbor, a cousin, and an aunt, each of whom also works under various versions of the same model. On any given day, whether the daughter is supervised depends on a coincidence of scheduling that Gloria cannot predict because none of the adults involved know whether they will be pointed at on any given morning. The arrangement works, she said, “maybe four days out of five.” On the fifth day, it does not work. The daughter, at eight, lets herself in with a key on a string around her neck.
Asked whether the fifth day worried her, Gloria answered a different question. “I am lucky,” she said, “that my daughter is responsible. Some of the women here have younger ones. Those women, I don’t ask them how it works. I know it doesn’t.”
The flexibility that the model produces at the warehouse gate, in other words, is absorbed, several miles away, by an eight-year-old with a key on a string. This absorption is not, strictly speaking, recorded anywhere in the economic data. It is what economists mean, in their technical vocabulary, when they use the word externality.
The Broader Pattern
Experts suggest the warehouse transition reflects a growing trend across multiple sectors of the American economy — a trend that has been underway for approximately forty years but has, in the past decade, achieved what one labor historian called “terminal clarity.”
The pattern involves three linked shifts. The first is the shift from employment to engagement — a semantic move that replaces the bounded, mutually obligated relationship of employer and employee with the momentary, transactional interaction of a party engaging a service. The second is the shift from relationship to transaction — which completes the semantic move by eliminating, from both sides, any expectation that the interaction will be repeated, remembered, or elaborated. The third is the shift from workforce to resource pool — a reframing in which the persons performing the work are no longer an organized body with identity but a diffuse reservoir of labor capacity, to be drawn upon as needed and returned to when not.
Each of these shifts, on its own, is small. A change of word, a change of framing. In aggregate, they describe the disassembly of a particular kind of economic entity — the firm composed of people — and its replacement with a particular kind of economic arrangement — the firm composed of contracts with people.
The Semantic Infrastructure
The shift has been accompanied by a parallel shift in language. Workers are no longer “hired”; they are “onboarded,” a verb whose maritime origin is instructive — one goes aboard a ship temporarily, not permanently. Workers are no longer “fired” or “laid off”; they are “offboarded,” or the engagement is “concluded,” or the worker is “transitioned out.” The words that once acknowledged a relationship of duration — “employee,” “job,” “career” — have been replaced by words that acknowledge only a temporally bounded transaction — “contractor,” “gig,” “engagement.”
Linguists note that this semantic drift is not neutral. The older vocabulary carried obligations; the newer vocabulary does not. To fire someone was to break a relationship. To offboard them is to complete a procedure. The feelings attached to the two actions — by the manager doing them, by the worker receiving them — are deliberately different.
One human resources consultant, asked whether this was intentional, said, “Of course. Why else would we change the words?”
The Accounting Treatment
From the vantage point of the auditor, the disposable workforce model presents as a marvel of clean numbers. Gone are the accrued vacation liabilities, the pension obligations, the deferred compensation arrangements, the OPEB footnotes that once sprawled across twelve pages of a 10-K. In their place: a single line, labor expense, which varies directly with revenue in a ratio that has, for the past six quarters, held to within a hundred basis points of its target.
Auditors describe such results with a word they do not often use: beautiful. A major-firm partner, speaking anonymously because his firm has an active engagement with one of the operating companies, explained the appeal. “In my thirty years in this profession,” he said, “I have never encountered a labor arrangement that presented so little audit risk. There is nothing to test. There are no employees. There are only invoices. Invoices are easy. Invoices we understand.”
Asked whether any of the workers were, in any meaningful sense, employees of the company he was auditing, he shook his head. “Not in any sense that would appear on my work papers,” he said. “The company I audit has eleven full-time employees. Of those, four are C-suite, four are accounting, and three operate the HVAC system. The HVAC operators, interestingly, are the only employees at the company who actually touch the facility.”
The Footnote Problem, Solved
Under the older employment model, publicly traded companies were required to disclose, in the notes to their financial statements, a range of labor-related obligations: projected benefit obligations, accumulated benefit obligations, service costs, interest costs, expected returns on plan assets, amortization of actuarial gains and losses. These disclosures, which often ran to fifteen pages and required specialized actuarial support to prepare, were a significant cost center.
Under the new model, the footnote disclosure for labor obligations reads, in its entirety: The Company has no defined benefit pension or other postretirement benefit obligations. This sentence is twelve words long. It replaces approximately eleven thousand words of prior disclosure. The cost savings are substantial.
The fact that the workers who previously would have been subject to these obligations still exist, still work at the facility on any given day, and still require, in the long term, the medical care and income replacement that the older disclosures contemplated, is not itself a footnote. It is not, in the technical accounting sense, anywhere. The workers have been removed from the financial statements not because the workers have been removed but because the relationship to the workers has been removed, and it was the relationship that was accounted for, not the workers themselves.
“Accounting,” the audit partner observed, “describes what a company owes. If the company owes nothing, the accounting has nothing to describe. This is a matter of scope. It is not a matter of reality.”
He added, after a moment: “I am not, for what it’s worth, an economist. I am an accountant. These are different things.”
The Organizing Attempt
In August of last year, at a facility in the same operational cluster as the Newark site, a group of approximately sixty day laborers attempted what labor organizers have described, with some admiration, as “the first known organizing action against a workforce that, on paper, does not exist.”
The action was scheduled for a Tuesday morning. The workers had agreed, via a group text message chain maintained on a set of prepaid phones, that they would present themselves at the staging area as usual, but would refuse selection until the staffing firm agreed to three demands: a raise of $3 per hour, the removal of the vest-return deduction, and the provision of water at work stations during shifts exceeding ninety-five degrees Fahrenheit. The demands were modest. They were also, in the strict legal sense, unenforceable, because the workers making them had no employer against whom to enforce them.
The man with the clipboard arrived at 5:14 a.m., as usual. He emerged from his pickup truck, as usual. He began pointing. The sixty workers, as agreed, stepped backward when pointed at, signaling refusal.
The man with the clipboard paused. He looked at the crescent. Then he began pointing at other workers — workers who were not part of the organizing action, because the organizing action had been coordinated only among a subset of that morning’s 180 or so potential selectees. Those other workers, unaware of the action or aware but not participating, stepped forward. The morning’s headcount was achieved by 5:26 a.m., twelve minutes ahead of schedule. The man with the clipboard returned to his pickup truck.
The sixty organized workers were not selected that morning. Nor were they selected the following morning. Nor the morning after that. By the fourth morning, approximately half of them had stopped returning to the staging area, because they had rent to pay and no income to pay it with. By the seventh morning, approximately none of them had returned.
“You can’t strike against someone who was never your employer,” one organizer later explained. “You can only refuse to be selected. And if you refuse to be selected, you aren’t selected. That is not a labor action. That is just a Tuesday.”
The staffing firm, asked about the attempted action, said that it was not aware of any action. The operating company, asked the same question, said that any action would have been a matter for the staffing firm, with which it had no direct relationship regarding the specific workers in question. The cascade had done, once again, exactly what the cascade was designed to do.
The Platforms
The transition to disposable labor at scale has been accompanied, in the past five years, by the emergence of a particular kind of software company — the staffing platform — which provides the technical infrastructure that allows the cascade to function with minimal administrative friction. These platforms, some of which have raised as much as $847 million in venture funding, occupy what their founders describe as the “talent-on-demand category.” Others have called the category different things, but the founders prefer “talent-on-demand.”
The platforms are, in their mechanical essence, databases. A facility posts a requirement: 280 workers, Tuesday morning, 5:00 a.m., forklift-certified preferred. Workers registered on the platform receive a notification. The first 280 workers to accept the notification are, in the platform’s language, “matched.” They present themselves at the gate. The platform invoices the facility. The platform pays the workers. The platform retains a fee.
The founders of such platforms describe their work in the language of liberation. They are, they say, freeing workers from the constraints of traditional employment — the commute, the schedule, the manager, the obligation to show up tomorrow. They are also, they say, freeing employers from the constraints of traditional employment — the payroll, the benefits, the compliance, the obligation for the worker to show up tomorrow. Both parties, in the founders’ telling, are freed. Both parties, in the founders’ decks, are flexible. Both parties, in the founders’ press releases, are empowered.
One founder, interviewed for a podcast series that covers the industry, described his company’s mission as follows: “We are removing the friction that has held back the labor market for a hundred years. We are letting work happen at the speed of need.” Asked what happened to the workers when the need receded, he said, “They remain on the platform. The platform is their home.” Asked what they ate in the meantime, he moved to the next question.
The Ratings Problem
Most staffing platforms incorporate a ratings system, in which workers are rated by facilities after each shift on a five-star scale. A worker whose rating falls below 4.2 stars is, in the platform’s language, “deprioritized” for future matches — which is to say, the platform ceases to send notifications to that worker. The worker is not informed of the deprioritization. The worker simply notices that the notifications have stopped.
The criteria by which ratings are assigned are, by design, opaque. A facility manager might rate a worker based on speed, accuracy, compliance with verbal instructions, cooperation, or any of several dozen other factors, none of which are documented or made subject to appeal. A rating, once given, cannot be contested, because the worker cannot identify the facility that gave it — the platform considers the rating itself proprietary information.
One worker described the experience of deprioritization as “a kind of evaporation.” The notifications had been coming, he said, three or four a week. Then they slowed. Then they stopped. He checked the app; he still had an account. He still had a profile. He had simply ceased, in some unannounced sense, to be visible to facilities. He had been rendered, in the platform’s back-end, a ghost. No one had told him. No one was obligated to.
“In the old economy,” one labor historian observed, “you got fired and you knew. Someone sat you down. There was a conversation, even if it was brief. In the new economy, you simply stop existing to the system, and you are expected to infer this from the absence of notifications. The absence is the notification. The notification is the absence. It is elegant.”
She added: “It is also, in a way that the founders of these platforms seem not to have anticipated, very lonely.”
The Language of the New Workplace
Human resources consultancies serving the disposable labor sector have, in the past several years, produced a specialized vocabulary that describes the various workforce arrangements and their management. The vocabulary is intended to be used internally among managers and investors; it is not, generally, used in the presence of workers, not because it is secret but because it is considered uninteresting to them.
Selected terms, with their operational definitions:
Talent Pool: the aggregate of workers available to be selected on any given morning. The word talent, in this usage, does not imply distinction; it is used in the sense of eligible bodies. The word pool, meanwhile, is used in the sense of a reservoir, from which a quantity is drawn and to which the remainder is returned.
Engagement: the period of time, typically one shift in duration, during which a worker is performing labor for the facility. The engagement begins at gate-in and ends at gate-out. The engagement is the unit of the relationship. Relationships do not accrete across engagements. Each engagement is, in the language of the consultancies, “self-contained.”
Rightsizing: the act of reducing the headcount on any given shift to the minimum required by the work at hand. Rightsizing is performed continuously throughout the day as work volume fluctuates; workers who are rightsized out of a shift mid-shift are escorted to the gate and their engagement is concluded. The word rightsizing is preferred over firing or sending home early because it frames the action as corrective: the workforce was, by implication, wrong-sized before.
Workforce Agility: the aggregate capability of the operation to scale labor up or down rapidly in response to demand. Workforce agility is measured in hours from demand signal to labor deployment; the industry benchmark, according to one trade publication, is twelve hours. Leading operations, the publication notes, have reduced this figure to as little as three.
Relationship Minimization: the deliberate reduction of the interpersonal, administrative, and legal connections between the facility and the individual worker to the smallest quantity required by law and operational necessity. Relationship minimization is a stated objective in the operating plans of several major operators; one operating plan describes a target state in which the relationship between the facility and the worker consists of “a gate, a barcode, a bell, and a wire transfer.”
Presence Billing: the accounting convention in which labor costs are recognized based on worker-hours of gate-in-to-gate-out presence rather than on employment status. Presence billing is the mechanism by which the older labor line item has been converted, in the accounting, into something that resembles a utility expense — paid only when consumed, unrecognized when idle.
The vocabulary, taken in aggregate, performs a particular function. It describes a set of arrangements whose older descriptions would have required words that carried moral weight — fire, discard, abandon — and replaces them with words that do not. The older words, the consultancies note, are “fraught.” The new words are “operational.” The distinction is important, they add, for reasons of morale.
Asked whose morale, one consultant answered without hesitation. “The managers’,” she said. “The workers’ morale is not what the vocabulary is for.”
The View from Port-au-Prince
Dr. Henry Gutenberg, Director of the Port-au-Prince Institute for Market Dysfunction, has tracked the American transition to disposable labor models from a vantage point he describes as “peculiarly well-suited to the task” — a country whose entire workforce has operated under analogous conditions for approximately two centuries and whose economists have, accordingly, developed vocabulary the United States is only now catching up to.
“What your investors are celebrating,” Dr. Gutenberg observed during a recent interview, “is the successful import of a labor architecture that my grandfather’s grandfather would have recognized immediately. The innovation is not the model. The innovation is the branding. You have called it ‘flexibility,’ which is a word that sounds like freedom. In Creole we have a more direct word. We do not translate it, because it does not require translation.”
He was asked whether the American version was, at minimum, more efficient. He considered this.
“Efficient for whom? The arrangement is efficient for capital. It has always been efficient for capital. That is not a new discovery. The new discovery is that you have finally stopped being embarrassed about saying so.”
Dr. Gutenberg noted, as he has on prior occasions, that the Port-au-Prince Institute has been declined for seventeen grant applications to American foundations interested in “labor market modernization.” The foundations, he said, prefer researchers who frame the subject prospectively. His frame is retrospective. “I am describing,” he said, “what you have already decided. You prefer to read about what you have not yet decided. This is understandable.”
Notes Toward a Taxonomy
At a conference in São Paulo last year, Dr. Gutenberg delivered a paper titled “On the Three Stages of Disposable Labor: From Justification to Operation to Branding,” in which he proposed that economies adopting disposable workforce models pass through a consistent sequence.
In the first stage, the arrangement is justified — usually by appeal to necessity, temporary conditions, or the alleged preferences of the workers themselves. In the second stage, the arrangement is simply operated, no longer requiring justification because it has become normal. In the third stage, the arrangement is branded — given a name that evokes freedom, flexibility, or opportunity, and which allows its participants, including its beneficiaries, to discuss it without discomfort.
“The United States,” Dr. Gutenberg concluded, “has reached stage three with remarkable speed. It took your country approximately forty years. My country took approximately four hundred. In this sense, you are, as you like to say, innovating.”
On the Question of Return
One question Dr. Gutenberg has addressed at length, in conversations with visiting American researchers, is whether the American transition is reversible. He is skeptical. He has been asked the question by, he estimates, forty-seven researchers in the past decade, and his answer has not changed.
“Labor arrangements,” he said, “are not reversed by labor policy. They are reversed by labor power. Policy can codify what power has already achieved, but it cannot create what power has not yet built. Your country dismantled, over four decades, the institutions that once generated worker power — the unions, the stable employment relationships, the geographic concentration of industrial workforces. Now you would like to know whether the arrangement produced by that dismantling can be reversed through, for example, legislation. The question assumes that legislation is what created the old arrangement. It is not. Power is what created the old arrangement. Legislation merely recorded it.”
He paused, as is his habit, and then added a clarifying observation.
“In my country, labor is cheap because labor has no alternatives. In your country, increasingly, labor is cheap because labor has no alternatives. The mechanism is the same. The difference is that in my country, we have always known this. In your country, you are still finding out.”
Asked what, practically, would have to happen for the transition to reverse, Dr. Gutenberg declined to speculate. “I am an economist,” he said. “I describe. I do not prescribe. Prescription is the work of people who believe that description will move them. Description, in my experience, does not move anyone. It only records.”
He added, almost as an afterthought: “The man with the clipboard in your parking lot is the end point of a process that began, in your country, with the decision that the relationship between a firm and a worker was an inefficiency to be optimized. The optimization has now been completed. The relationship has been reduced to its smallest possible form — a gesture, a barcode, a shift, a gate. Congratulations. You have achieved what your theorists proposed. The question now is whether you like what you have achieved. That, too, is a question I will leave to you.”
The Closing Statement
Investors, for their part, remain optimistic. On the most recent earnings call — the one with the tea — the chief executive concluded his prepared remarks with a single word, which analysts have since adopted as a kind of shorthand for the entire model:
“This is efficient.”
He did not specify what this referred to. The word was understood to encompass the model as a whole — the gates, the vests, the crescents, the pointing, the bells, the barcode, the inventory metaphor, the laminated cards, the erased names, the purged presence tables, the $47 deduction, the three-person skeleton crew, and the gentle elimination of the human resources function. All of it, together, efficient.
No follow-up question was asked. The analyst line had, by that point, gone silent. One analyst later explained that he had not known what to ask. “I kept thinking of questions,” he said, “but all of them sounded sentimental. I didn’t want to be the sentimental one.”
Operations at Press Time
At press time, operations continued. The facility’s lights were on. The gates were open. A crescent of approximately 170 men and women stood in the parking lot at 5:14 a.m. on a Wednesday morning, awaiting selection for a shift whose outcome, for each of them individually, would not be known until a man with a clipboard emerged from a pickup truck and began pointing.
Staffing levels, according to the daily operational dashboard that investors can access from their phones, were variable.
By design.
The warehouse investors celebrating their new model have not discovered a new labor arrangement. They have discovered an old one, and given it a new name. The arrangement was previously considered uncivilized. It is now considered efficient. These two facts are not in contradiction. The move from uncivilized to efficient was accomplished not by changing the arrangement but by changing what we are prepared to call it.
The word flexibility has done an enormous amount of work in this transition. It has taken a condition — the absence of commitment from one party to another — and recoded it as a property that both parties possess equally. Workers are told they have flexibility. Employers are told they have flexibility. Only one party’s flexibility includes the option of eating next week. This is not a small difference, but the word obscures it, which is what the word is for.
Every morning at 5:14 a.m., in a parking lot in Newark, a man with a clipboard points. Every evening at 5:47 p.m., a vest is returned to a bin. Between these two events, 340 people exist, in the database, as presences. Before and after, they do not. This is what the word efficient describes. We agreed to the word before we looked at what it was describing. The looking is optional. The agreement was the point.
Editorial Footnotes
¹ The name “Relational Minimization Partners” is fictional; the training curriculum it describes is a composite of three real consulting practices currently operating in the northeastern United States, at least one of which uses nearly identical slide language.
² The $47 vest-return deduction is specific to one facility in the reporting sample. Other facilities assess deductions ranging from $12 to $180, with no discernible correlation to the replacement cost of the vest, which industry sources place at approximately $4.80.
³ The claim that “peak season” has expanded to 47 weeks per year is drawn from internal staffing memos. The remaining five weeks are not, notably, a single block — they are scattered across the calendar according to demand, such that no worker is ever quite certain whether any given week is or is not peak. This ambiguity is itself a feature of the model.
⁴ The “presence table” described in the article is a real data structure in a real warehouse management system, though the vendor has asked not to be named, citing what a spokesperson called “the way people tend to react when they hear about it.”
⁵ The phrase “we are acquainted” in the operations manual is translated from an earlier Spanish-language version of the same document, in which the phrase was “estamos conocidos” — a construction that Spanish speakers on the editorial staff describe as grammatically unusual and tonally cold. The coldness is believed to be deliberate. The grammatical unusualness is believed to be accidental but retained for legal reasons.
⁶ Dr. Gutenberg’s seventeen declined grant applications are a matter of public record. The foundations in question, when contacted, declined to comment on specific applications but noted that they fund research that is “forward-looking.” Dr. Gutenberg’s research is, by their criteria, backward-looking. He has conceded this characterization, noting that he is describing what has already happened. The foundations prefer to fund descriptions of what has not yet happened. Both descriptions are of the same arrangement.
Editor’s note: The facility’s name, the exact location, and the names of the investors have been withheld at the request of several sources, none of whom were workers. The workers, for the most part, did not request anonymity. Their names were not on file.