Benchmarking and Target Setting in the Public Sector: Types, Problems and Uses

A state hospital in Kano learns that another state hospital treats similar patients for a cost per bed-day that is a fifth lower. A secondary school sees its pass rate published in a league table, ranked below the school across town. A refuse collection department finds that a neighbouring council collects a tonne for less.

None of these organisations has a profit to measure, so each asks a different question: how good are we, compared with the best, and what should we aim for next year? Benchmarking answers the first question and target setting the second.

This topic explains the types of benchmarking, what makes it succeed and what makes it fail, and how targets are set and used in public bodies, where objectives are many, hard to measure and sometimes in conflict. It draws on the earlier topic on critical success factors and key performance indicators. The next topic covers the changing role of the management accountant and contemporary issues.

Benchmarking: meaning, types and requirements

Benchmarking compares an organisation’s performance with the performance of another organisation or unit, used as the benchmark. Performance is judged better or worse than the benchmark, and the comparison shows strengths and weaknesses.

Its aim is not the comparison itself but the improvement that follows: to identify differences and then to develop new ways of doing things, ideally so that the organisation becomes better than the benchmark and not merely level with it.

Competitor analysis is not benchmarking

Competitor analysis compares a company’s results with those of its main rivals, often as a league table with the best at the top. It tells the company where it stands. Benchmarking goes further: it identifies why others perform better and develops improvements. A table of prices for rival supermarkets is competitor analysis; a study of how the cheapest one organises its stock is benchmarking.

The types

• Internal benchmarking. Similar operations inside one organisation, such as regional branches or area offices, are compared. The best branch is the benchmark for the others. The data is easy to get and no secrets are shared.
• Competitive benchmarking. Performance is compared with the most successful competitor, to find where the rival does better and how to close and then reverse the gap. Information may be hard to obtain. Some bodies supply “inter-firm comparisons” for a group of companies without revealing any single firm’s figures, though most comparison in practice is by analysts of listed companies.
• Process (or operational) benchmarking. The commonest type. One process, such as purchasing, order handling, call-centre operation or delivery, is compared with the same process in an organisation in a different industry that has a name for excellence. The partner is not a competitor, so each side can share openly. Comparison with the best can give ideas to copy or to do differently.
• Functional benchmarking. A core function, such as sales and marketing or research and development, is compared with the same function in a successful competitor, to find why it performs better.
• Strategic benchmarking. Strategies are compared with those of the most successful competitor, or of successful companies elsewhere: objectives, core competences, process capability, products, alliances and technology. A start is a survey of customers and shared suppliers.
• Product benchmarking (reverse engineering). The organisation’s products are compared with competitors’ products by obtaining and analysing them, looking at the rival’s costs, product concepts, and strengths and weaknesses in design and quality.
• Customer benchmarking. The customer is the benchmark: what customers want and expect is compared with what the organisation provides, and the gap is closed.

Benchmarking is also a tool of total quality management. The usual steps are to choose the measures that matter, choose partners, collect data, analyse the performance gaps, understand the best practice behind them, plan improvements, carry them out, monitor progress, and then repeat, because it is an ongoing process.

Requirements and problems

For benchmarking to improve competitiveness:

• select the key aspects of performance that must succeed
• make it continuous, not a one-off
• aim to be better than the benchmark and not just to catch up
• be open and honest with partners, ready to give as much as you get

Problems include:

• it may be hard to find a critical process, so the benefits are small
• hard to find the right benchmarks and best-practice partners
• hard to explain the value of the exercise and how to use the results
• hard to plan and scope
• hard to get reliable comparable information
• the cost may exceed the benefit (so keep it simple and computerise preparation)
• managers and staff may not understand its purpose, so the use must be explained and systems set up to monitor it
• comparison can create a culture of blame in which staff who look worse than a rival become defensive and resentful

Targets and benchmarks in the public sector

Public bodies and not-for-profit organisations have no profit to measure, so their performance is judged against a mix of objectives, many of them non-financial and some not quantifiable. They must work within budget limits, so they aim for value for money: economy, efficiency and effectiveness. Benchmarking is natural here, because there are many similar units to compare: schools, hospitals, local authorities, tax offices, police divisions.

• League tables. Units are ranked on selected measures. They create pride and competition and show the weak performers. They are competitor analysis, not benchmarking, unless the best performers’ methods are then studied and spread.
• Internal and external benchmarks. A body compares its units with its best unit, with similar bodies elsewhere, and with a process in a different sector, such as a hospital learning from an airline about turnaround.
• Targets. Targets turn benchmarks into goals: a pass rate, a waiting time, the days to process a claim, the cost per case.

Setting good targets

A target should be specific, measurable, achievable, relevant and time-bound. Three further questions come from the Fitzgerald and Moon work on standards. Ownership: do those who must meet it accept it as their own, which is likelier if they took part in setting it? Achievability: do they think it can be done?

A demanding target motivates more than an easy one, but an impossible one demotivates, so targets should be challenging but reachable. Equity: is the target fair between units whose local conditions differ, such as a school in a poor area and a school in a rich one? Where conditions vary, targets may need allowances or judgement and not one rule for all.

What goes wrong

• Gaming. Staff manipulate the measure: a hospital moves patients between lists to meet a waiting time, or a school enters only strong pupils for an examination.
• Neglect of the unmeasured. Effort moves to what is counted and away from what is not, such as quality of care.
• Perverse incentives. A target for speed can lower quality; a target for cost can reduce access.
• Poor comparability. Units differ in the people they serve, so rankings may be unfair.
• Short-term focus. Annual targets can drive out long-term investment.
• Cost and blame. Collecting data costs time that could serve the public, and league tables can create fear.

The remedies are a small set of measures that cover the objectives, outcome measures as well as outputs, adjustment for local conditions, audit of the data, involvement of staff in setting targets, and a focus on learning from the best and not on punishing the worst.

Layouts for benchmarking and target calculations

1. Unit cost. Total cost ÷ units of output (tonne collected, bed-day, case).
2. Gap to the best. Unit cost of the unit − unit cost of the best unit; gap % = gap ÷ best unit cost × 100.
3. Saving if the unit matched the best. Gap per unit × the unit’s output. Add for all units. The best unit has none.
4. Target number. Current rate × cohort = current number; target rate × cohort = target number; difference = extra needed.
5. Cost per extra success. Cost of the programme ÷ extra successes expected.

Money is in ₦, rounded once, at the end. Treat a benchmark saving as a target to examine, not a certainty, because units differ.

Worked examples

1. Cost per tonne.
   A 45,000,000 ÷ 6,000 = ₦7,500.
   B 50,400,000 ÷ 7,200 = ₦7,000.
   C 38,400,000 ÷ 4,800 = ₦8,000.
   D 32,500,000 ÷ 5,000 = ₦6,500, the best.
2. Saving if each matched D.
   A (7,500 − 6,500) × 6,000 = ₦6.0m.
   B (7,000 − 6,500) × 7,200 = ₦3.6m.
   C (8,000 − 6,500) × 4,800 = ₦7.2m.
   Total ₦16.8m.
3. Comment. The saving is a target to study. Before demanding it, find why D is cheaper: shorter routes, newer vehicles, a denser population, or different service levels. If the reason is geography, A, B and C cannot copy it.

1. Passes now.
   62% × 400 = 248.
2. Passes at target.
   70% × 400 = 280.
   Extra needed = 32.
3. Programme cost.
   80 × ₦45,000 = ₦3,600,000 = ₦3.6m.
4. Cost per extra pass.
   3,600,000 ÷ 32 = ₦112,500.
5. Comment. The target is reachable if the programme works. The board should also check that the school does not raise its rate by keeping weak pupils out of the examination, and that it measures learning, not only passes.

Expect variants: a different number of units, benchmarks in value for money terms, targets for several measures, or a request to say what type of benchmarking is being used.

Process benchmarking in practice, and openness

Process benchmarking compares one process with the best in any industry. A well-known case is Xerox, which wanted to improve the way it handled customer orders and studied a catalogue retailer known for excellence in order fulfilment; the two firms shared information, and Xerox used what it learned to improve its own system.

For this to work the collaboration must be open and honest, and a company should be ready to give more information to its partner than it hopes to receive. Sharing confidential information with a direct competitor is risky, so competitive benchmarking often depends on published data, while process benchmarking with a firm in an unrelated industry is easier to agree.

Common mistakes

• Calling a league table benchmarking. Benchmarking needs the step of learning how the best do it.
• Naming types without examples. Give one example for each type, in the setting of the question.
• Ignoring differences between units. A fair comparison allows for local conditions.
• Treating a saving as certain. It is a target to investigate.
• Setting targets without ownership or fairness. These decide whether staff accept them.
• Forgetting perverse effects. Gaming and neglect of the unmeasured are standard points.

Quick revision

• Benchmarking: compare with best practice, then improve to be better. Types: internal, competitive, process, functional, strategic, product, customer.
• Requirements: key aspects, continuous, aim to beat, open partners. Problems: cost, critical process, data, understanding, blame.
• Public sector: multiple, non-financial objectives; value for money; league tables are comparison, not benchmarking.
• Targets: specific, measurable, achievable, relevant, time-bound; ownership, achievability, equity.
• Risks: gaming, neglect, perverse incentives, poor comparability, short-termism.

How the marks are earned

• Describe types: a definition and an example for each, about two marks per type.
• State problems or requirements: one mark each, tied to the body described.
• Calculate: unit costs, gaps and savings with workings and a comment.
• Discuss targets: ownership, achievability, equity, risks and a conclusion.

Practice questions

These are original questions, written in the style of the Skills examination. Work each one before you open the answer.

Recall check: multiple choice

Section B style: written questions

Model answers

Spotted a mistake? Tell us and we will check it.