3.2 Benchmarking Types and Performance Goal Setting
Key Takeaways
- ASQ names four benchmarking types: best practices, competitive, collaborative, and breakthrough.
- Competitive benchmarking compares against direct rivals and yields relative position; collaborative benchmarking exchanges data with willing partners, often outside the industry.
- Breakthrough benchmarking looks outside the industry for a radically different process model rather than an incremental performance gap.
- A benchmark converts into a project only after it is normalized to comparable units and paired with a measured internal baseline.
- The classic benchmarking cycle is plan, collect, analyze the gap, set goals, adapt and implement, then re-measure.
What benchmarking is for
Benchmarking is the structured comparison of your process performance and practices against a reference, in order to set a defensible performance goal and to learn how the reference achieves it. The second half matters more than the first. A gap number tells you that a 4-day cycle time is possible; only the practice study tells you that it is achieved by removing a batch approval step you still run.
ASQ tests this topic at the Apply level: given a situation, choose the benchmarking type and translate the result into measures and goals.
The four types named in the Body of Knowledge
| Type | Comparison target | What it produces | Typical limitation |
|---|---|---|---|
| Best practices | Recognized leaders in a specific practice, wherever they are | A documented method known to work | Practice may not transfer to your constraints |
| Competitive | Direct competitors in the same market | Relative market position on cost, quality, delivery | Data is hard to obtain and often estimated |
| Collaborative | Willing partners exchanging data, often through a consortium or industry group | Verified, normalized, comparable metrics | Restricted to what partners agree to share |
| Breakthrough | Radically different process models, usually outside your industry | Step-change redesign ideas | High adaptation cost; often needs DFSS rather than DMAIC |
Some organizations also use internal benchmarking -- comparison between plants, shifts, or business units of the same company. It is the cheapest and fastest source of comparable data because definitions and systems already match, and it is often the right first step before paying for external studies. Its ceiling is that it can only reveal the best you already do somewhere.
Choosing a type
- Need to know whether you are losing orders on price or lead time? Competitive.
- Need a proven method for a specific practice such as changeover reduction? Best practices.
- Need trustworthy, normalized numbers across a peer group? Collaborative.
- Current process architecture is at its limit and incremental gains are exhausted? Breakthrough.
The classic breakthrough example is the pit-stop study: hospitals redesigned intensive-care handovers by studying motorsport pit crews, not other hospitals. Nothing about a competitive benchmark would have produced that design.
The benchmarking execution cycle
- Plan. Identify the process to benchmark, the specific metrics, and the operational definitions. Decide the type and the partners.
- Collect. Gather internal baseline data first. Then gather partner data through published sources, consortium reports, site visits, or structured interviews.
- Analyze the gap. Normalize both sides to comparable units and quantify the difference. Then investigate the practice difference that explains it.
- Set goals. Convert the gap into a target with a date, and decide how much of the gap this project will close.
- Adapt and implement. Adapt rather than copy: the partner's constraints are not yours.
- Re-measure. Benchmarks age. A target set against a competitor's 2023 performance is a target against a moving reference.
Normalization: where benchmarking studies fail
Raw comparisons mislead unless both sides are expressed in the same denominators. The three normalizations that matter most:
- Scope: does the partner's "order cycle time" start at order receipt or at credit approval?
- Volume and mix: cost per unit at 10,000 units/month is not comparable to cost per unit at 400 units/month, and a high-mix line cannot be compared with a single-product line.
- Accounting basis: fully loaded labour rate versus direct rate can change a cost benchmark by more than the gap you are trying to close.
Write the operational definition for every benchmarked metric before you collect anything. If the definition cannot be matched, the metric cannot be benchmarked.
Turning a benchmark into project measures and goals
A benchmark becomes usable only through an explicit chain:
Then decide the closure fraction. Committing to close 100% of a gap against a best-in-class reference in one project is usually a charter written to fail. A common convention is to target 50% to 70% of the gap in the first project and to re-benchmark afterwards.
Worked example. A claims operation measures mean cycle time of 11.4 days. A collaborative benchmark across eight insurers, normalized to the same start and stop events, gives a median of 6.2 days and a first-quartile value of 4.1 days.
- Gap to median: $11.4 - 6.2 = 5.2$ days.
- Gap to first quartile: $11.4 - 4.1 = 7.3$ days.
- Charter goal at 60% of the median gap: $11.4 - (0.60 \times 5.2) = 8.3$ days.
The practice study then explains the gap: benchmark partners auto-adjudicate low-value claims below a threshold, while this operation routes everything to a human adjuster. That practice, not the number, is what the Improve phase acts on.
Ethics and legal limits
Competitive benchmarking must not become information exchange that restrains trade. Do not share or solicit current or future pricing, costs that reveal pricing, capacity plans, or customer allocation with a competitor. Most consortia route data through a neutral third party that returns only aggregated, anonymized results for exactly this reason. Where a benchmarking code of conduct exists, follow it: confidentiality, reciprocity, and use of data only for the stated purpose.
A hospital's incremental improvements to its intensive-care handover process have plateaued, and it studies Formula 1 pit crews to redesign the handover entirely. Which benchmarking type is this?
Eight insurers agree to submit claims cycle-time data to a neutral third party that returns anonymized, normalized results to each participant. Which benchmarking type is being used, and why is the neutral third party important?
A benchmarking study reports a partner's order cycle time as 3.1 days against the company's 9.0 days. Before setting a project goal, what must be verified first?