Removing layers can shorten a decision path, but only if ownership, authority and controls are redesigned with it.
Reuters reported on September 2 that Uber plans to cut about 3,300 employees, roughly 10% of its workforce, as part of an organisational overhaul. The company said the changes would flatten management layers, combine some teams and reduce organisational complexity that had accumulated during rapid growth. Reuters also reported that Uber intends to reduce the number of employees positioned seven or more reporting layers below the CEO by 20% and cut the number of teams with only one or two direct reports by nearly half (Reuters, 2026).
The announcement is important for process improvement because it frames organisational structure as a flow problem. Uber’s own message described growth as having created more coordination and fragmented ownership, while the stated objective of the new structure is clearer ownership and faster decisions. Whether the restructuring actually produces those outcomes cannot yet be known; the news describes an announced change, not a measured result. That uncertainty makes the case more useful, not less, because it separates the design logic from the later evidence that should be required (Uber, 2026).
Removing management layers is not automatically Lean, Kaizen or process improvement. Fewer roles can reduce handoffs and waiting, but they can also overload remaining managers, weaken controls or simply move the queue somewhere else. The improvement opportunity lies in redesigning the decision process: who decides, what information they need, which decisions require escalation, how quickly they should be made and how the organisation will know whether the new method is better.
Why Organisational Layers Become a Process Problem
Coordination can grow faster than value
As organisations grow, they add roles for valid reasons: specialist expertise, local management, risk control, cross-functional coordination and leadership capacity. Over time, however, a structure designed for one scale of business can accumulate extra interfaces. A proposal may need to be prepared by one team, reviewed by another, aligned with a third, escalated through several leaders and then returned for clarification. Each participant may add legitimate input, yet the combined process can become slower than the decision requires.
The process-improvement lens is different from a simple headcount lens. It asks how much elapsed time comes from active analysis and how much comes from waiting, scheduling, handoff and rework. It also asks how often a decision moves upward because the lower level lacks authority, how many people review the same information, and whether a control changes the decision often enough to justify its cost. These questions reveal managerial work as a process that can be measured and redesigned.
Fragmented ownership creates hidden rework
When several teams partially own the same outcome, work often returns in loops. One group prepares the commercial case, another owns technology, another controls risk and another manages the customer impact. If nobody owns the complete decision, each team may optimise its own requirement while the overall case moves back and forth. The visible symptom is more meetings; the underlying issue is an incomplete decision process.
Clear ownership does not mean one person ignores specialist input. It means one role is accountable for reaching the defined outcome and understands which inputs are advisory, which are mandatory controls and which decisions can be delegated. The process should also define what constitutes a complete submission. A large share of managerial delay comes from work arriving without the evidence required for the next decision, forcing clarification and resubmission.
Removing roles can move the bottleneck rather than remove it
Flattening can shorten the formal reporting chain while increasing the number of decisions landing on each remaining manager. If authority does not move downward at the same time, the organisation may create a wider but more congested approval point. The same risk appears when teams are combined without redesigning prioritisation. More work arrives at one place, but the rules for sequencing it remain unclear.
That is why spans of control should be considered together with work type, decision volume and variability. A manager overseeing highly standardised work can often support a larger span than one overseeing specialised, high-risk or rapidly changing work. Process Improvement Training helps make this distinction explicit by focusing on demand, capacity, queues, decision rules and failure modes rather than assuming that one organisation-chart ratio is universally efficient.
Redesign Decision Flow Before Removing Structure
Map the approval path and measure managerial waiting
Start with a small number of important decisions rather than the organisation chart. Choose decisions that materially affect customers, revenue, cost, delivery, people or risk. Map the trigger, information required, people consulted, approval points, queues, escalation paths and final action. Then measure total elapsed time and separate active work from waiting. The result often shows that a decision taking five days contains only a few hours of actual analysis.
The map should capture rework as well. Count how often submissions are returned, which information is missing, where opinions conflict and how often decisions are escalated beyond the level originally intended. This is more actionable than a general complaint about bureaucracy. It identifies specific delays that can be removed, simplified or controlled.
Set decision rights by risk and reversibility
A faster organisation needs a clear rule for where authority sits. Low-risk, reversible decisions can usually be made closer to the work if guardrails are clear. High-consequence or difficult-to-reverse decisions may still require specialist or senior review. The mistake is applying the same approval depth to both. A good process uses the lightest control that can reliably manage the real risk.
Decision rights should be paired with thresholds and escalation conditions. A frontline or team leader might approve within a defined cost, customer-impact or risk range, while exceptions move to a specialist. The information packet should be standardised so reviewers receive the evidence they need without repeated clarification. This reduces waiting without asking people to make decisions beyond their competence or authority.
Measure queue time, rework and escalation
Decision speed is easy to claim and harder to prove. The baseline should include elapsed time from request to decision, time spent waiting at each approval point, number of handoffs, percentage returned for more information, percentage escalated, decision reversal or defect rate, and the downstream outcome. A faster decision is not an improvement if it produces more corrections, risk events or customer problems.
A faster decision process defines normal authority, exception routes, escalation conditions and verification rather than simply removing approvals.
After a structural change, review these measures by decision type. Some paths may improve immediately while others become overloaded. If one senior role becomes a new bottleneck, the data should show it. The response may be to delegate authority, improve the submission standard, change meeting cadence or remove a redundant review. This is continuous improvement applied to management work rather than a one-time reorganisation.
STEP Bootcamp is built around the same practical sequence: define a worthwhile problem, analyse the current process, design an improvement, prove value and build controls. If you want to strengthen those skills, the 28-day programme provides live Process Improvement Training with applied assignments and personalised feedback.
Build a Leaner System That Can Hold the Gain
Change spans of control with the operating rhythm
A flatter structure changes how information reaches managers. If one manager now supports more people or a larger domain, the old meeting pattern may no longer work. The organisation may need clearer daily or weekly operating reviews, more visible performance measures, stronger delegation and defined escalation windows. Otherwise the manager spends the saved structural capacity recreating coordination through ad hoc meetings and messages.
The operating rhythm should match the speed of the work. High-frequency operational issues need a short cycle for visibility and escalation. Strategic or capital decisions can use a longer cadence but should still have clear preparation standards and decision dates. The aim is not more meetings. It is a reliable mechanism that moves information to the right decision level at the right time.
Standardise exceptions and escalation
Lean processes are often misunderstood as processes with fewer controls. Strong processes have clear controls that are proportionate and predictable. An exception should trigger a defined route rather than a search for the most senior person available. Repeated exceptions should be categorised so the organisation can distinguish unusual cases from a process that is routinely failing to handle normal demand.
This is also where role clarity protects speed. Employees should know what they can decide, when they must consult a specialist and when they must stop and escalate. When these boundaries are vague, people protect themselves by asking for more approvals. Clarity can therefore remove coordination without removing necessary expertise.
Apply Kaizen to managerial work
The most valuable lesson from Uber’s announcement is not that every large organisation should remove management layers. It is that growth can create coordination work that should periodically be challenged. Once the new structure is operating, the organisation should continue to examine decision queues, duplicated reviews, recurring escalations and meetings that no longer change outcomes. Small changes to authority, information standards, cadence and ownership can remove friction before another large restructuring becomes necessary.
That is a natural use of Kaizen. Teams improve the management process itself, using evidence from everyday work rather than waiting for a crisis. Continuous Improvement Training helps professionals see these opportunities as process issues: waiting, handoffs, unclear standards, overprocessing, defects and capacity constraints can exist in office and managerial work just as clearly as they exist on a factory floor.
Uber’s restructuring will eventually need to be judged by measured outcomes, not by the neatness of the new organisation chart. If decisions become faster, ownership becomes clearer and service or growth outcomes improve without creating new control failures, the design will have evidence behind it. If bottlenecks simply move, the process will need another improvement cycle. Professionals who want to build that evidence-based capability can explore STEP Bootcamp’s Process Improvement Training, Continuous Improvement Training and Kaizen Training, which focuses on applying the method rather than memorising terminology.
References
Reuters. (2026, September 2; updated September 3). Uber to lay off 10% of staff in biggest cuts since COVID. View source Uber. (2026, September 2). Building a simpler, faster Uber. View source
Leave a Reply