Almost every restructuring is sold internally as a financial decision and remembered by everyone who lived through it as a human one. That gap, between the spreadsheet that justified it and the trust it spent, is where most cost transformations quietly fail.

The quick version

  • Layoffs frequently miss their own target. The best longitudinal study found no consistent evidence that workforce cuts improved financial performance versus firms that didn't cut (Cascio).
  • The people who stay are the real cost line. "Survivor syndrome", drops in trust, commitment and productivity, is driven less by the cut itself than by whether it was seen as fair (Brockner).
  • Process beats severity. How you decide, communicate and treat people predicts the outcome more than how many you remove. Start from a stated philosophy, not a number (Sucher & Gupta).
  • Cut roles, not just people. Responsible restructuring redesigns the work first and treats people as assets to redeploy; pure headcount-cutting tends to grow the cost back.

The idea in depth: cutting cost is not the same as cutting people

The case for redundancy looks airtight on a slide. Labour is the biggest controllable line; remove some of it and earnings rise. The reason it so often disappoints is that the slide models the saving and ignores the second-order costs, and those costs land on the part of the business you kept.

Start with the uncomfortable headline finding. In a longitudinal study of S&P 500 firms, Wayne Cascio and colleagues compared companies that made significant employment cuts with those that didn't, and reported finding no significant, consistent evidence that downsizing led to improved financial performance, measured by return on assets, including industry-adjusted ROA. Cascio's framing in "Strategies for Responsible Restructuring" (Academy of Management Perspectives, 2005) draws the distinction that matters: employment downsizing treats people as costs to be cut, while responsible restructuring treats them as assets to be redeployed, and it is the second approach, not the first, that tends to recover. Which is why the conversation that begins "we need to remove N heads" is the wrong conversation. Restart it at "which work should we stop, simplify or move, and what does that imply for roles?" Cost transformation that begins with the org chart instead of the work tends to cut muscle, then quietly rehire it within a year or so.

An honest limitation. Cascio's work is correlational, not a controlled experiment, distressed firms are more likely both to cut and to underperform, so some of the gap reflects the trouble they were already in, not the cut itself. The defensible reading is narrower but still damning for the reflex: cutting people is not a reliable route to better financial performance, and anyone promising that it is should be asked for their evidence.

Why the survivors decide whether it worked

Here is the part the cost model never shows. The savings are immediate; the damage is deferred and lands on the people who remain. Joel Brockner's research on "survivor syndrome", summarised in "Managing the Effects of Layoffs on Survivors" (California Management Review, 1992), found that the people who keep their jobs often respond with reduced commitment, motivation and productivity, the opposite of the relieved gratitude managers assume. His central insight is that survivors' reactions are driven less by the layoff itself than by the fairness of it: whether the process was even-handed (procedural justice) and whether people were treated with dignity along the way (interactional justice).

That reframes the whole exercise. A restructuring is not over when the leavers leave; it is a message to everyone watching about how this organisation behaves when it is under pressure. Treat leavers shabbily, vague reasons, a security escort, a Friday-afternoon email, and the survivors learn exactly what loyalty is worth here, then update their own behaviour accordingly. So manage the layoff for the audience that stays: explain the logic and the criteria honestly, give people room to ask questions, and let the workforce see departing colleagues treated decently. You are not being kind at the expense of the business; protecting the survivors' trust is the business case.

flowchart TD
  D(["Decision to cut cost"]) --> A(["Headcount-first
remove N people"])
  D --> B(["Work-first
redesign the work,
then the roles"])
  A --> A1(["Immediate saving
on the slide"])
  A1 --> A2(["Survivor trust falls,
knowledge walks,
rehiring creeps back"])
  A2 --> A3(["Cost returns
goal often missed"])
  B --> B1(["Slower, deliberate
fewer, defensible cuts"])
  B1 --> B2(["Survivors see it as fair
capacity preserved"])
  B2 --> B3(["Saving holds
structure actually changes"])
Two routes from the same decision. The headcount-first path books a saving and often loses it; the work-first path is slower but holds. Leaders Loop

An honest limitation. Fairness mitigates the harm; it does not erase it. A well-run redundancy still ends jobs, and no amount of process turns that into a good day for the person carrying a box to their car. The justice research tells you how to avoid compounding the loss with a sense of betrayal, not how to make people glad it happened. Leaders who oversell "we did it humanely" to their remaining staff usually find that survivors trust the actions, not the adjective.

The execution moves that actually decide it

If process matters more than severity, then the craft of restructuring is mostly in the sequence of decisions before anyone is told. Sandra Sucher and Shalene Gupta, in "Layoffs That Don't Break Your Company" (Harvard Business Review, May–June 2018), argue that the most successful workforce transitions start from a stated philosophy, an explicit set of commitments about how the company will behave, rather than from a target number. From that philosophy flow two disciplines: genuinely exploring alternatives to redundancy (reduced hours, redeployment, retraining, a hiring freeze, voluntary exits) before reaching for it, and treating the people you do let go as ambassadors who will talk about how they were handled for years.

Concretely, that means a different order of operations. Decide the philosophy and the selection criteria first and write them down. Pressure-test the alternatives honestly, not as a box-ticking ritual, but as real options that might change the number. Then, only once the criteria are set, identify the roles. And when you communicate, tell people the truth about why, give them the support you would want (notice, severance, references, outplacement), and brief the managers who deliver the news so the worst conversation of someone's year is not also clumsy.

flowchart LR
  P(["1. Philosophy
what we commit to"]) --> C(["2. Criteria
fair, role-based"])
  C --> ALT(["3. Alternatives
hours, redeploy,
retrain, voluntary"])
  ALT --> R(["4. Roles
decide what stops"])
  R --> COM(["5. Communicate
why, with dignity"])
  COM --> S(["6. Support survivors
rebuild the work"])
Restructuring as a sequence, not an announcement. The number comes near the end, after the philosophy, the criteria and a real look at the alternatives. Leaders Loop

The cut is the easy part. Whether it works is decided by the order in which you make the decisions, and by the people you keep.

One legal guardrail, not legal advice: redundancy is heavily regulated and the rules differ sharply by country. In the United States, for example, the federal WARN Act requires larger employers to give 60 days' notice of a qualifying plant closing or mass layoff; the UK, the EU and Australia each have their own collective-consultation and notice regimes. Before you finalise anything, check your jurisdiction and take qualified employment-law and HR advice, getting the process legally wrong can cost more than the redundancy was meant to save.

A worked example: the 15% that became a rehire

Picture a 600-person software business told by its board to take 15% out of operating cost within a quarter. (Illustrative scenario and figures, not a real company.) The fast path is obvious: 15% of cost is roughly 90 roles, so HR is asked to model a 90-person reduction by team, weighted to the most expensive functions.

The headcount-first version. Ninety people go in a single morning, skewed toward senior engineers because they cost most. The quarter's numbers look good. Within six months, three things surface: two products have no one who understands their codebase, so contractors are hired back at a premium; the remaining engineers, watching their most respected colleagues escorted out, start interviewing elsewhere, and voluntary attrition climbs; and a vague, defensive all-hands has left everyone assuming another round is coming. The booked saving is real on paper and largely gone in practice by year-end. This is precisely the pattern Cascio's data would predict.

The work-first version. Leadership starts with a philosophy ("we will be honest about why, fair in how we choose, and generous in how we part") and clear criteria. A real look at alternatives finds genuine savings: pausing two low-conviction projects, a hiring freeze already in train, and a voluntary-exit offer that 25 people take. That shrinks the involuntary number to around 45, chosen by role redundancy rather than salary. Survivors hear a straight account of which work the company is stopping and why, see leavers given notice, severance and references, and the affected products are deliberately handed over before anyone departs. The saving is smaller in the first quarter and far more durable, because the structure genuinely changed and the capacity to deliver survived.

The difference between the two is not compassion as a luxury. It is the same Brockner and Sucher mechanism: the second company spent less trust, kept more knowledge, and didn't rehire its own cut. Same board mandate, very different year.

Frequently asked questions

Don't layoffs at least save money in the short term?

They book a saving immediately, yes, that's why they're so tempting under quarterly pressure. The question is whether the saving holds. Once you net off severance, lost productivity from survivor disengagement, knowledge that walks out the door, and the rehiring that often follows within a year or two, the durable saving is frequently far smaller than the model promised. Cascio's longitudinal work is the reason to be sceptical of the clean number: cutting people is not a reliable path to better financial performance.

What exactly is "survivor syndrome"?

It's the well-documented dip in morale, trust, commitment and productivity among the employees who keep their jobs after a layoff, alongside guilt, anxiety about being next, and a heavier workload. Brockner's research found it is driven less by the cut itself than by whether people judged the process fair and humane. The practical implication is that managing the layoff well for the survivors, clear reasons, fair criteria, dignified treatment of leavers, is what protects the productivity you were trying to gain.

How do we make redundancies feel fair?

Fairness has two parts the research separates. Procedural fairness is about the process, consistent, role-based criteria applied evenly, not who is liked or who is loudest. Interactional fairness is about treatment, telling people the real reason, giving notice and support, and letting them keep their dignity. You can't make a redundancy welcome, but you can stop it being experienced as a betrayal, and that distinction is most of what survivors remember.

Should we cut deep once or trim repeatedly?

Repeated small rounds are usually worse for the survivors than one well-explained reset, because each round renews the fear that another is coming and keeps the whole organisation looking over its shoulder instead of working. That argues for doing the analysis properly once, cutting to a structure you actually believe in, and then saying clearly that this is the change, rather than salami-slicing. The caveat is honesty: don't promise "this is the last round" unless you mean it, because a broken promise costs more trust than the second round itself.

Is restructuring always about redundancy?

No, and conflating the two is part of the problem. Restructuring is changing how the work is organised; redundancy is one possible consequence. Plenty of genuine cost transformation comes from stopping low-value work, simplifying processes, renegotiating contracts, redeploying people to higher-value roles, or removing layers of management, none of which require the indiscriminate headcount cut that does the most damage. Treat redundancy as the last lever, not the first.

Related in the Toolkit

A restructuring is a change programme with unusually high stakes, so the wider change craft applies directly: the choreography in change models (Kotter, ADKAR, Lewin) and the survivor-trust problem at the heart of communication during change are really the same discipline seen from two angles.

Where to go next