Organizational Design

Reducing Overhead Without Losing Performance

Marin Niehues19 December 202411 min read

Bright office space with glass facade

Overhead falls without loss of performance when an organization first assesses tasks and approvals by their contribution to delivery and only then adjusts positions and structures. Across-the-board cuts hit friction and delivery capability alike, because they start at the cost center and ignore the effect of the work. This article describes a four-step approach with which executive management and divisional leadership remove effort where it enables no delivery.

What is overhead?

Overhead refers to the indirect costs of the support functions, meaning everything that does not directly produce output for customers. The part of it that can be reduced without loss of performance is the work that neither delivers nor enables delivery: friction work. Typical overhead costs are the staff costs of finance, HR, IT, procurement and legal, the costs of executive management and its staff functions, rent for office space used for administration, software for administrative tasks, insurance and external advice for internal purposes.

These costs are not dispensable simply because they generate no revenue. Finance or quality assurance create preconditions without which delivery does not run reliably. Reducing overhead therefore requires a framework for the type of work. The following distinction is a refinement from delivery practice that adds to the accounting concept the question of each activity's contribution to delivery:

  • Delivery work: activities that are directly part of the result for the customer, such as production or advisory work.
  • Enabling work: activities that steer or finance delivery, such as planning and accounting.
  • Friction work: activities that arise only because interfaces and decision paths are poorly designed, such as duplicate data entry or multi-stage approvals.

The framework follows a distinction that Taiichi Ohno described for the Toyota Production System: value-adding work, non-value-adding work that is necessary under present conditions, and waste.

Reduction without loss of performance targets friction work. It can be removed without anything missing for the customer. Enabling work is designed and simplified where needed, and delivery work remains protected. How organizational design and delivery management work together here is described in the article characteristics and metrics of a high-performing organization.

What is a normal overhead ratio?

A normal overhead ratio is the ratio achieved by companies of the same sector and size, and it is assessed against this comparison and against the company's own trend over several years. A software firm and a machinery manufacturer have fundamentally different cost structures, so a cross-sector benchmark is misleading.

Common measures are the overhead ratio, meaning indirect costs relative to revenue, and the ratio of indirect to direct employees. For reduction without loss of performance, the share of friction work also matters, because it shows the part that can be removed without any effect on the customer. A warning sign is an overhead ratio that grows faster than revenue over three to five years. In that case the administration of the business is growing faster than the business itself.

A four-step approach

The approach follows a fixed sequence, because each step provides the basis for deciding on the next.

  • Measure: tasks per role are recorded and time shares estimated by sampling. Approval chains and recurring reports are inventoried. The result is a map of the work.
  • Classify: managers assign each activity to one of the three types of work, using the test questions in the box below. The assignment is cross-checked, because every unit tends to classify its own work favorably.
  • Remove at the cause: for each instance of friction work the cause is identified and eliminated, for example by raising approval thresholds and simplifying handovers.
  • Safeguard: before starting, delivery metrics are defined that are reviewed monthly, together with a fallback rule that determines at what deviation a measure is reversed.

The third step decides success. Cutting positions before the cause of friction has been eliminated shifts the work onto the remaining people, and the friction returns as overload in the delivery teams. Capacity therefore only counts as freed once the cause has been eliminated.

Overhead value analysis and zero-based approaches: the established methods

Overhead value analysis is a method for reducing the indirect costs of support functions, developed in the 1970s and derived from classic value analysis. Value analysis itself goes back to Lawrence D. Miles, who developed it at General Electric in the 1940s. Each support unit lists its services, sets their costs against them and has them rated by the recipients. Under an ambitious savings target, the units develop proposals that are graded by effect and risk and then approved.

The approach described here adopts from overhead value analysis the rating of each service by its recipients. It differs in two respects. It starts with the type of work and the cause of friction, and the saving follows from that. And it safeguards every measure through delivery metrics before the saving is booked. An even more fundamental approach is the zero-based organization, which justifies every structure from scratch.

Where overhead typically arises

In practice, friction work is concentrated in a few places. Approvals come first: a decision passes through several levels although the last level never changes it. The result is a longer lead time with unchanged decision quality. Reports are the second: the same data is prepared in several formats, and no recipient derives a decision from it. Interfaces between units are the third, because every handover generates coordination as long as responsibilities remain open.

Part of this friction work can be removed with AI, above all in reporting and document processing. The sequence is decisive: first the workflow is simplified, then automated. Automating a workflow full of friction accelerates the friction and entrenches it.

How much can be saved, and how long does an overhead program take?

The saving without loss of performance is capped by the measured share of friction work, and as a guide from delivery practice a program takes nine to fifteen months from measurement to booked saving. Measurement and classification take six to eight weeks. Eliminating the causes takes six to twelve months, depending on the number of approval chains and systems affected. The saving is booked once the delivery metrics then remain stable for at least three months.

The implementation rate requires the most careful estimate in the calculation. It falls with the number of causes that lie outside the program, for example in group requirements or in systems that will only be replaced later.

Safeguarding performance

A reduction counts as successful when delivery remains stable. This requires metrics with baselines before the first intervention. Lead time and on-time delivery show whether the organization continues to deliver as quickly and reliably as before. Quality serves as a cross-check, because savings first become noticeable where reviews are dropped.

The saving is only booked once the metrics remain stable over several months. Until then it counts as a hypothesis. This prevents success reports that only reveal themselves as overload in the next planning round.

How Niehues Consulting delivers overhead programs

Niehues Consulting supports organizations in reducing overhead while preserving delivery capability. Marin Niehues builds organizations that deliver: he leads initiatives as delivery manager or product owner and designs structures and decision paths so that the organization works reliably in day-to-day business as well as in transformation. Projects range from VC-funded scale-ups to DAX 40 groups, with assignments in 20+ countries on six continents.

The approach follows the Niehues Delivery Model. In the Understand phase, the map of the work with classification and metrics is produced, guided by the review question: what is the baseline, and how will success be measured? In the Build phase, the causes of friction are eliminated and the future roles, decision paths and interfaces are defined, where useful with AI prototypes that prove their benefit in the real workflow. In the Roll out phase, the new workflows are embedded so that the relief lasts, with a metric that shows after 90 days whether delivery is stable. The entry point is a free 30-minute initial consultation via the contact page.

Frequently asked questions

What are overhead costs?

Overhead costs are the indirect costs of support functions that do not directly produce output for customers. Examples are the staff costs of finance, HR and IT, rent for administrative space, administrative software and insurance.

What is a normal overhead ratio?

A normal ratio is the one achieved by companies of the same sector and size. A warning sign is an overhead ratio that grows faster than revenue over three to five years. The ratio of indirect to direct employees additionally shows how administration is developing relative to delivery.

How can overhead be measured?

Common measures are the overhead ratio and the ratio of indirect to direct employees. For reduction without loss of performance, the share of friction work is added, determined by recording tasks per role and estimating time shares by sampling.

What is overhead value analysis?

Overhead value analysis is a method for reducing the indirect costs of support functions in which each unit lists its services with their costs and has them rated by the recipients. Under an ambitious savings target, proposals emerge that are approved according to effect and risk.

How long does an overhead reduction program take, and how much can it save?

As a guide from delivery practice, a program takes nine to fifteen months from measurement to booked saving. The saving without loss of performance is capped by the measured share of friction work and is calculated with the implementation rate, as the calculation in this article shows.

How do you reduce overhead without performance falling?

Overhead falls without loss of performance when tasks and approvals are assessed by their contribution to delivery before any decision on positions. Friction work is eliminated at its cause, for example by raising approval thresholds or merging reports. Metrics such as lead time and on-time delivery, combined with a fallback rule, show whether delivery remains stable.

Which areas typically have the highest overhead?

Friction work is concentrated in approvals, reports and interfaces between units, regardless of the department. Units with many handovers and multi-stage decision chains are particularly affected. The map of the work shows for each organization where the shares are largest.

Can AI reduce overhead?

Yes, above all in reporting and document processing. The precondition is that the workflow is simplified before automation, because otherwise AI accelerates the friction.

Let’s talk about your initiative.

In 30 minutes we clarify where your organization stands and which step will pay off most now.

Request a consultation

30 minutes · free · no obligation