
Good jobs are also designed through operations
Job quality is also produced through the organization of shifts, tasks and room to respond. Productivity improvements and working conditions are outcomes best examined separately.
Better pay changes an important part of the employment relationship. It does not, on its own, resolve a last-minute schedule change or two essential tasks competing for the same time. That does not diminish the pay improvement. It makes visible that job quality is also decided through the organization of operations: what work we accept, how much capacity we reserve and which interruptions we treat as normal.
The distinction has a consequence for those running a company. Some of the conditions they offer are defined neither in compensation policy nor in a Human Resources initiative. They are produced when deliveries are coordinated, shifts prepared or changes authorized that others will have to absorb. The OECD job quality framework (opens in a new tab) separates earnings quality, labor market security and the quality of the working environment. It is not a management recipe, but it prevents the judgment from being reduced to a single dimension. We can improve one while leaving another untouched, or even making it worse.
That relationship between employment and operations is central to Zeynep Ton’s proposal. In an excerpt published by MIT Sloan in 2023 (opens in a new tab), she presents the good jobs system as a combination of investment in people and decisions about work. A well-paid, well-trained workforce cannot compensate for every operational design, nor does a well-organized process replace the need to invest in those carrying it out. The complementarity matters: changing only one part may leave unresolved the condition that prevents the other from being put to use.
The work that disappears from view when time runs short
In a store, helping someone who asks a question and correctly replenishing a product are different contributions to the same service. One happens in front of the customer; the other makes it possible to find what they need without asking. If both compete for the same people’s time, always favoring visible demand can leave undone the work that prevents future interruptions. The problem may look like a succession of people needing help when part of it lies in work that was never completed.
Ton’s study of labor, quality and profitability (opens in a new tab) allows that less visible part to be examined. The working paper, revised in 2009, uses data from 268 stores in one chain between 1999 and 2002. It approximates the quantity of labor through spending on wages and benefits, not a direct count of hours or people. It finds associations between that measure, process compliance and profitability; it does not find an equivalent significant relationship between service quality and profitability. Its design does not isolate the effect of raising wages or produce an estimate transferable to any company. Its value lies in not limiting labor’s contribution to customer interaction: it also includes ensuring that products, records and processes are in a condition to sustain it.

This reading changes what we look for when reviewing a working day. If the team spends part of it locating incorrectly recorded products, repeating checks or redoing a delivery, measuring how hard people work is not enough. We need to know what generates that demand and which part could be avoided. Adding people may be necessary; simplifying an instruction or correcting a source of errors may be necessary too. There is no automatic equivalence between fewer hours and greater efficiency, any more than between more staff and better operations. The comparison must retain the work needed to produce a usable service.
Predictability can also improve operations
Schedules reveal another point of connection between working conditions and performance. The experiment by Kesavan, Lambert, Williams and Pendem at Gap (opens in a new tab), published in 2022, evaluated a multicomponent intervention in 28 US stores over nine months between 2015 and 2016. The package included repeating shift time slots across weeks and an application for offering or taking shifts. The intention-to-treat analysis found a 5.1% increase in productivity, measured as sales per labor hour, with higher sales and fewer total hours. The result applies to the set of practices studied, not the isolated effect of announcing a shift earlier or a test of Ton’s entire system. It provides evidence that organizing schedules with workers in mind can improve, rather than necessarily harm, the commercial result.
That allows us to discuss an often implicit idea: that a company’s flexibility depends on keeping people’s time available until the last moment. An organization needs to respond to real variations in demand. The question is how much of what it presents as unpredictable comes from the customer and how much from deliveries, promotions or internal decisions that could be coordinated differently. Stabilizing those conditions does not require knowing the future; it may allow less of the variation to become an urgent change in working hours.
Training people to take on different tasks can help cover variations during a shift. But that versatility needs an organization that allows it to be used: criteria for changing tasks, adequate preparation and understandable limits on the responsibility assumed. It does not turn one person into simultaneous coverage for two demands. Nor does it justify always leaving whoever learned more functions to absorb the unexpected. The valuable resource is being able to reorganize the contribution, not erase the cost of displaced work.
Ton also includes operating with slack among the choices in her model: having room to respond to demand, absences and development. That requires discussing capacity which, viewed only through immediate utilization, may look surplus. If every minute is already committed, an explanation to a new colleague or an exception requiring data to be checked competes with a pending delivery. Slack has a cost and must be justified according to the work; removing it from the calculation does not eliminate the situations that need it.
Two outcomes worth keeping in view
A productivity improvement does not, by itself, settle job quality. Sales per hour may increase because work is done better, but that indicator does not capture the full experience of those sustaining the result. We also need to observe what changed in their conditions: predictability, workload, opportunities to learn and the ability to respond without omitting an indispensable part of the work. Commercial and employment improvements can reinforce each other; neither should certify the other without examination.
It would also be a mistake to demand that every working condition justify its value exclusively through an immediate financial return. Being able to organize one’s life with some predictability or work in decent conditions is not valuable only when it increases sales. Evidence of compatibility with better results helps challenge a false opposition and seek more sustainable designs; it does not turn every right or basic condition into an optional investment subject to that calculation. Economic viability deserves serious discussion, but does not determine on its own what employment is acceptable to offer.
For a company choosing to invest in better jobs, the work does not end with announcing the benefit. It includes reviewing which demands are simplified, how changes are coordinated and how much room remains to do what matters well. That review cannot remain solely with whoever administers the benefit if the decisions generating urgent demands remain beyond their reach.
The schedule we promise, the tasks we fit into it and the room we leave to resolve them are also part of the job we offer. We cannot present those decisions as a purely operational matter and then attribute their consequences to people’s motivation. Job quality also bears the signature of those of us who organize the work.





