A Model for Business Process Improvement
In short Andersen’s model makes process improvement a repeatable cycle: measure performance, plan, improve, implement, and measure again. It rests on documented processes, a balanced set of measures, and periodic self-assessment that shows where to improve first.
Based on: B. Andersen, Business Process Improvement Toolbox, ch. 1–6 and 13.
On this page
Key ideas#
What a business process is#
Andersen treats a business process as a repeatable chain of related activities that uses the company’s resources to produce measurable results for an internal or external customer. The central picture is supplier → process → customer.
Drawing on the TOPP and ENAPS improvement programs, he classifies processes as: - Primary — create value for the customer directly, running from customer to supplier. - Support — enable the primary processes (finance, HR). - Development — raise the value chain to a new level of performance (product development, supplier development).
Organizing by department brings specialization but also suboptimization: each department optimizes itself rather than the end result. Andersen’s answer is process orientation, with process owners who have the authority to start change, the duty to monitor the process, and the task of forming improvement teams.
The improvement cycle#
Andersen uses Bredrup’s model from the TOPP program: a loop of four phases (performance measurement → planning → improvement → implementation) that mirrors Deming’s Plan–Do–Check–Act. There is no fixed entry point, but Andersen recommends starting with measurement.
An improvement team has four to eight members: a leader, a link to management, the process owner, people who work in the process, and ideally customers and suppliers. Andersen advises against including very senior managers; they rarely have time to get into the details and tend to defend their own departments.
Documenting processes#
There are two strategies: - Step by step: document a process only when an improvement project starts. This suits small organizations. - All at once: document 15–100 processes up front. This involves more people and gives management a map for setting priorities. Andersen prefers it.
Documentation tools range from least to most detailed: relationship map, flowchart, cross-functional flowchart, and multi-level flowchart with numbered levels (1.0, 1.1 and so on).
Measuring performance#
Andersen holds two ideas together: you cannot manage what you do not measure, and what you measure is what you get, because the measure itself changes behavior. He distinguishes: - Hard measures (direct, objective) from soft ones (indirect, through proxies). - Financial from non-financial measures. - Result measures (what was achieved), process measures (how the process runs), diagnostic measures (early warnings of what is coming) and competence measures (readiness for future change, the hardest to define).
A balanced set covers five dimensions: time, quality, cost, flexibility and environmental impact.
Self-assessment#
Self-assessment is not routine monitoring. It happens at a fixed point in time, covers the whole organization, and supports strategic decisions. To set it up, choose 10–20 key processes, link each to measures across the five dimensions, and repeat at a fixed interval, often every six months. Interpret the results with trend analysis, a spider chart against competitors, and a matrix of importance against current performance.
Choosing tools#
Andersen classifies about 40 tools by four criteria: 1. Purpose: which of six stages the tool serves (setting priorities, understanding the problem, analyzing it, generating ideas, developing improvements, implementing). 2. Scale of change: small (basic quality tools, idea generation, statistical process control), medium (simplification, work-cell analysis, quality function deployment) or radical (idealization, business process reengineering, benchmarking). 3. Resources and time: roughly follows scale; reengineering and benchmarking demand the most. 4. Source of impulse: internal for every tool except benchmarking, which brings ideas from outside.
His central message: a tool is only a means. Do not try to use as many tools as possible; use a few that fit together.
How to apply it#
- Draw a high-level map of your processes and classify each as primary, support or development.
- Assign an owner to each key process.
- Define a few measures per process across the five dimensions, including at least one diagnostic measure.
- Run a self-assessment, plot importance against current performance, and pick one or two processes to improve.
- Form a small team, document the chosen process, and choose tools for each stage.
- Implement, measure again, and feed the results into the next cycle.
Example#
The Workshop, a fictional chain of 12 bike-repair workshops, picks ten key processes, including “book repair → return bike” (primary), “order spare parts” (support) and “introduce e-bike servicing” (development). Measures for the repair process: - Time: turnaround in days. - Quality: repeat repairs within 30 days. - Cost: cost per repair. - Flexibility: share of small jobs done the same day. - Environment: parts packaging waste.
The six-monthly self-assessment shows turnaround as both important and weak, so it becomes the first improvement project. A five-person team (the process owner, a workshop manager, two mechanics and a front-desk lead) documents the process with a cross-functional flowchart.
Pitfalls#
- Measuring only results → add process and diagnostic measures so you can act before results slip.
- Putting senior managers on the team → involve them as sponsors, not members.
- Collecting tools → choose a few that fit the stage and scale of the problem.
Sources and further reading#
- B. Andersen. Business Process Improvement Toolbox. ASQ Quality Press, 1999: ch. 1–3 (process definition and classification, improvement model, documentation), ch. 4–6 (measurement, self-assessment, classification of tools), ch. 13 (using the toolbox).