Every role has a Job-to-Market Distance, and almost no company measures it

Picture two janitors. One cleans a corporate office at midnight. The work matters, but the only people who ever see it are other employees. The other keeps the lobby and restrooms of a luxury hotel. Every guest walks through that work, and a slipped standard shows up in reviews, complaints, and return visits within days.

Same occupation. Same tasks. Completely different relationship to the market. The hotel janitor is a few feet from the customer's judgment. The office janitor is separated from it by walls the org chart never draws.

That separation deserves a name: Job-to-Market Distance, or JMD. It is the structural gap between the ordinary work of a job and the external loop through which a company senses its market, serves it, and hears back. Once you start scoring jobs by their JMD, a lot of familiar organizational problems start looking like one problem.

The org chart shows authority. It hides distance.

Companies sort jobs into front office and back office, revenue and support, customer-facing and internal. Those labels are old and comfortable, and they miss the property that matters.

Every organization runs a loop: a market signal comes in, gets interpreted, becomes work, the work produces a consequence out in the world, and the consequence comes back as feedback. Job-to-Market Distance describes where a job sits inside that loop. Some roles receive raw signal and watch their work land. Others receive instructions that have passed through five interpreters, and never learn what happened next.

Notice what the definition does not say. It says nothing about seniority, salary, title, or whether you ever talk to a customer. A site reliability engineer may never join a customer call and still be one failure away from every user's bad day. A salesperson can talk to customers all week and have almost no influence on whether the product delivers what was sold. A CEO can be market-near, or can live entirely inside internally prepared summaries. Distance is a property of how the work is wired, not of the title on the door.

The four dimensions of Job-to-Market Distance

JMD is not one measurement. A job's distance has four dimensions, and they move independently:

• Inbound signal distance: how many hands, summaries, and reinterpretations market information passes through before it reaches the job.
• Outbound causal distance: how many dependent steps sit between the work and a consequence the outside world can feel.
• Feedback distance: whether the outcome ever returns to the person in a readable form.
• Temporal distance: how long the full loop takes to close.

Inbound: what the signal survives on the way in

Say a customer keeps hitting the same problem. At one company, the report goes straight from the customer to the engineer who owns the feature. At another, it goes customer, account executive, customer success manager, CS director, product manager, planning meeting, engineering manager, engineer. Both engineers eventually get a ticket. They did not get the same information.

Every handoff compresses, and compression is lossy. Context, emotion, ambiguity, and contradictory details fall out along the way, and whatever an intermediary decides is unimportant may never travel at all.

Marketing researchers have studied the organizational version of this for decades: Ajay Kohli and Bernard Jaworski's classic work on market orientation in the Journal of Marketing defined it as generating market intelligence, spreading it across the company, and acting on it. JMD moves that question down to the level of a single job. How much transformation happens before market reality reaches this desk?

Outbound: what your work travels through before the market feels it

A bartender's output gets tasted in seconds. A checkout page change shows up in conversion data the same day. An analyst's report has to pass through a director, a budget recommendation, an executive committee, a resource allocation, and an operations plan before any customer notices anything. As that chain lengthens, three things fade: attribution, feedback, and the market's ability to falsify whether the work was any good in the first place.

Feedback: do you ever hear back?

Take two video editors doing identical work, a subject we know something about at MarketScale. One sees client approvals, audience numbers, revision patterns, and renewals. The other uploads the finished cut into an internal workflow and never sees it again. Their effect on the market might be the same. Their ability to learn from the market is not even close.

Time: a consequence that arrives late arrives weak

Time does the same quiet damage. A consequence that arrives in five minutes teaches you something. A consequence that arrives in five months is a rumor. Enterprise sales, capital projects, pharmaceuticals, and long-horizon R&D all live with slow loops, and that is fine, but it changes what feedback can do for the people inside them.

Jobs mix these four freely. An infrastructure engineer can have no customer contact at all and still score beautifully on consequence and feedback. A salesperson can score high on contact and low on nearly everything else. That is exactly why JMD is useful and the old labels are not.

Distance is not importance

Here is where the idea gets easy to misuse, so let's close that door now. JMD is not a value ranking. A cybersecurity architect is about as market-distant as jobs get, and also one of the people the company can least afford to lose. A corporate tax specialist can quietly preserve millions. A regulatory counsel may talk to regulators all day and to customers never, and be the reason the company still exists. Distance and strategic value are different axes, and collapsing them into one is how companies end up cutting exactly the wrong people.

What distance does reliably change is legibility. Work close to the market gets tested by reality: the client renews or leaves, the feature gets used or ignored, the room is clean or it is not. Work far from the market gets tested by other employees: manager approval, committee acceptance, process completion. Those are legitimate checks, but they are internal ones, and as distance grows, the risk rises that internal validation quietly substitutes for external validation.

High-distance work can hide your best people and your most pointless activity at the same time, and that is precisely the problem. The outside world cannot arbitrate between them.

Watching the work land changes the worker

The human evidence here is stronger than you might guess. In a field experiment published in Organizational Behavior and Human Decision Processes, Adam Grant and colleagues had university fundraising callers spend a few minutes with a scholarship student whose award their calls helped fund. Weekly phone time rose 142% and weekly money raised rose 171%. Comparison groups showed no such jump.

Grant's later field experiments with lifeguards and fundraisers, published in the Journal of Applied Psychology, found the same pattern: make the consequence of the work visible and effort follows.

Zoom out and the pattern holds. A meta-analysis by Stephen Humphrey, Jennifer Nahrgang, and Frederick Morgeson in the Journal of Applied Psychology, covering 259 studies and 219,625 workers, found that work design characteristics, task significance and feedback among them, explained large shares of the variance in satisfaction, commitment, and rated performance. A separate meta-analysis in the Journal of Management Studies tied meaningful work tightly to engagement and commitment. None of this proves JMD theory by itself.

Companies buy distance on purpose, and often they should

If distance were pure cost, the fix would be obvious and this would be a short article. It is not. Jay Galbraith's information-processing view of organizations, published in Interfaces back in 1974, made the point that people cannot all process everything, so companies summarize, specialize, escalate, and build hierarchy. Every one of those moves adds distance, and every one of them buys something: focus, scale, fewer interruptions.

Luis Garicano's work in the Journal of Political Economy modeled hierarchy as a knowledge machine. Frontline workers handle the common problems, experts handle the rare ones, and the company avoids teaching everyone everything. Richard Chase argued in Operations Research decades ago that customer contact itself carries an efficiency cost, because customers introduce variability into operations. Pushing every employee closer to the customer would be a terrible prescription. Deep work needs protection.

The newest evidence says both things at once. A 2025 NBER working paper by Michael Ewens and Xavier Giroud reconstructed organizational hierarchies from resume data across more than 3,100 US public companies and roughly seven million workers. Firms averaged around ten layers. More hierarchical firms showed higher operating performance, and higher administrative costs. It is a working paper, so treat the findings as early, but the shape of the tradeoff is exactly the one JMD predicts.

Even the layers themselves can prove their worth in measurable ways. Another NBER study, by Espinosa and Stanton, found that training frontline workers freed their managers from routine assistance, and a structural model attributed roughly 45% of the training program's total benefits to those managerial spillovers. So the lesson is never that managers equal distance. Hierarchy is a purchase, and the question is always whether each layer earns its price.

JMD Drift, one rational hire at a time

No company starts distant. A founder sells, builds, ships, and hears every complaint personally. Then the founder hires salespeople. Salespeople need coordination, so there is a sales manager. Managers need forecasts, forecasts need operations, operations needs systems, systems need administrators, and soon every function produces reports for every other function. Each hire is individually rational. The cumulative result can be a company where a growing share of the work has other internal work as its only customer.

Call that JMD Drift: separation from market information and consequence growing faster than the value the added specialization creates. You can feel drift from the inside. Metrics turn into proxies, then proxies of proxies, until teams measure meetings held and tickets closed because outcome attribution got too hard. Internal demand starts feeling like real demand.

Finance requests a report, the report feeds an analysis, the analysis feeds a committee, and the committee requests another report. Everyone is busy. Every task is necessary to someone else's task.

The question JMD forces is the one drifting organizations are structurally bad at asking: where does this chain of work finally terminate in something the outside world can judge? Distance that creates specialization, risk protection, or scale is worth every layer. Distance that merely creates more distance is organizational debt.

Market Bridges beat reorgs

The good news: reducing effective distance almost never requires a reorg, because structure and information travel on separate tracks. You can keep the specialization and shorten the path. Call the mechanisms Market Bridges:

• Raw usage data on every engineer's screen
• Selected customer calls open to anyone who wants to listen
• Win and loss analysis that leaves the sales team
• Churn interviews shared beyond customer success
• Post-project outcome reports that actually reach the people who did the project

This is an old finding wearing new clothes. Michael Tushman and Thomas Scanlan showed in the Academy of Management Journal in 1981 that information crosses organizational boundaries through people who are well connected on both sides of them. And in 2025, a field experiment across 144 teams published in the Journal of Organizational Behavior found that deliberately injecting the customer's perspective into team process improved how teams worked with information and, downstream, customer adoption. The more functionally diverse the team, the more it helped.

So the number that matters is not structural distance, which specialization fixes in place, but effective distance: what remains after the bridges are counted. An engineer eight layers deep with live telemetry, churn narratives, and a monthly customer call can be effectively closer to the market than a salesperson who talks to customers daily but never learns what happened after the contract was signed.

How to measure JMD

It runs without a consultant, and it works because JMD is structural: you are scoring how the work is wired, not how anyone feels about it.

Step one: score each role on the four dimensions

For any role, rate agreement with a short set of structural statements on a 1 to 5 scale, scored separately for each dimension:

• Signal: people in this role regularly receive information based directly on customer behavior, and market changes can reach them without passing through many internal interpreters.
• Consequence: changes in the quality of this role's work can affect an external outcome without a long chain of unrelated decisions, and the external cost of poor performance is identifiable.
• Feedback: people in this role normally learn what happened after their work reached customers, in a form they can interpret and act on.
• Time: external consequences become observable soon enough to shape the next round of work.

Invert the scores so higher means more distant, and keep the four numbers separate. The dimensions do not reflect some hidden trait; they create the distance, the way the parts of a machine create its output. A role with clean inbound signal and a severed feedback loop needs a different repair than a role drowning in interpreters, and one blended number would hide which wire is cut.

Have three people score each role independently: the person in it, their manager, and an adjacent coworker. The ratings should broadly agree if the construct is as structural as it claims. Where they diverge, you have learned something anyway: the company does not know how that job is actually wired.

Step two: roll it up into a company number

Now aggregate. For a first pass, give each role one score by averaging its four dimension numbers, equal-weighted; the right weights are an empirical question the framework leaves open.

The headcount-weighted average of role scores then gives you the simplest version of Company Market Distance, or CMD: one gauge of whether your people, on average, are connected enough to the market and the customer. The payroll-weighted version is often more honest, because it shows where the money sits rather than where the people sit. Track it the way you track any health number: on a cadence, against your own baseline.

There is no universal red line, but the alarms are specific. A company average that climbs quarter after quarter as headcount grows is drift, caught in the act. And three companion numbers tell you whether a high average is a problem or a choice: the share of payroll sitting in high-distance roles, the percentage of roles that ever receive interpretable feedback from the outside world, and the share of work whose immediate customer is another internal role.

When the average is high and rising, feedback closure is falling, and internal consumption is growing, the issue is not one bad team. The organization is detaching.

Two honest warnings. Measure the normal path, not the shortest one: a CEO who talks to one customer a year does not have low distance just because a direct route technically exists. And the average alone can lie. Two companies can post the identical mean while one wires raw customer information into every function and the other passes summaries of summaries up and down the chart. The number starts the conversation. The wiring diagram finishes it.

Step three: ask the six questions

With scores in hand, run the qualitative pass. Take any role, especially the distant ones, and ask:

1. What market information does this job actually need, and how many hands touch that information before it arrives?
2. What external consequence does the work influence, and how many contingent steps sit in between?
3. Does the person ever find out what happened? How long does that take?
4. Who consumes the output? If the answer is another internal role, where does the chain finally end?
5. What event in the outside world could prove this work useful, or useless?
6. If the role is distant, what does the distance buy? Could you compress the information gap without touching the specialization?

Notice the review never asks who to cut. Distance alone predicts vulnerability badly, because distant roles include the invisible backbone of the company. The dangerous quadrant is specific: high distance, low strategic weight, easily substituted, and no feedback loop that could ever prove otherwise. That is where restructuring risk actually concentrates, and it is also where the least examined work tends to accumulate.

AI just changed the price of distance

One more reason this matters right now. A large share of historical distance exists because moving information used to require people: collecting, summarizing, reformatting, routing, reconciling, reporting. Whole layers exist mainly to carry meaning between other layers. AI makes exactly that work cheap.

Galbraith's theory predicted that organizations would restructure when information-processing technology changed, and there are early signs it is happening: the Ewens and Giroud paper reports organizational flattening at firms following AI adoption.

The prediction JMD makes is sharper than the usual line that AI will flatten companies. It is that AI applies the most pressure to distance whose only job was information intermediation. Distance built on scarce judgment, accountability, expertise, relationships, or risk protection keeps its justification. Distance built on human message-passing is losing its reason to exist.

Distance you can defend

Every company will always contain distance, and it should. The failure mode is not distance. It is detachment: the point where the connection between distant work and external reality becomes too weak to carry information, consequence, or justification in either direction. A research scientist can be distant and deeply connected. A dashboard-fed executive can sit at the top of the chart and be completely detached.

So skip the flattening crusade and ask a sharper question of every layer, role, and recurring report in the building: what does this unit of distance buy us? The org chart will not tell you which layers pay their way. Measuring JMD will.