Targets derived from the purpose, not from the baseline

Once a baseline is known, the easiest target to write is the baseline improved by a round number: five points better, ten per cent cheaper, a fifth less water. Targets like that look ambitious and are easy to remember, but they have no connection to what the business needs. Some will be too weak, so the project can meet them and still fail its purpose. Others will be too strict, so a good option is rejected for missing a number nobody needed. A target should answer one question: what level must this measure reach, by when, for the fundamental objectives to be met?

This chapter derives the laundry’s four targets from that question, using the purpose and the value at stake set out in Chapter 2: the tender’s delivery requirement, the margin at the new price, and the tender’s scoring thresholds for water and energy. It then fixes the period over which each target is judged and the date by which it must be met, separates the minimum acceptable level from a stretch, checks the wording against SMART, and ends with the first set of targets the laundry wrote, which were round numbers and wrong in three different directions.

Where each target comes from

This is an illustration. All amounts are Australian dollars excluding GST. The purpose, from Chapters 1 and 2, is to keep the hospital contract through the tender that closes at month 12, at the expected price of A$1.76/kg, while holding the margin, and to score fully on water and energy. Each part of that sentence fixes one target.

• OTIF comes from the tender’s delivery requirement: 98% over the trailing 26 weeks.
• Cost per clean kilogram comes from the new price and the margin the owners want to keep.
• Water and energy intensity come from the tender’s scoring thresholds.

None of the four comes from the baseline. The baselines in Chapter 4 say how far each measure must move; they do not say where it must end up.

OTIF: at most 93 failures in the evidence window

The hospital group will assess OTIF over the 26 weeks before the tender closes. With the tender closing at month 12, those are the 26 weeks of months 6 to 11, the evidence window. At 180 deliveries a week the window holds 4,680 deliveries. A 98% requirement allows 2% of them to fail: 4,680 × 0.02 = 93.6. Failures come in whole numbers, so the limit is 93: 93 failures give (4,680 − 93) ÷ 4,680 = 98.01%, which meets the requirement, and 94 give 97.99%, which does not. The limit is rounded down because rounding up would set a target the tender would fail.

Ninety-three failures in 26 weeks is an average of 3.6 a week, against 15.8 a week in the baseline. Expressed as a count, the target is easier to manage than as a percentage. Everyone in the laundry can understand a budget of 93 failures for the half-year, and everyone can see how much of it has been used.

A count also makes clear how tight the target is. A delivery process whose true failure rate is exactly 2% would expect 93.6 failures in the window, and the binomial distribution gives it only about a 50% chance of finishing at 93 or fewer. Running at exactly the requirement means even odds of failing it. That is why the laundry treats 98% as the minimum acceptable level and sets a stretch of 99%, at most 46 failures in the window (4,680 × 0.01 = 46.8, rounded down). A process designed for better than 98% leaves room for the bad weeks Chapter 4 showed every stable process will have.

Cost: the ceiling from the price and the margin

The cost target comes from arithmetic on the purpose. After the tender, the laundry’s volume will still be 55% hospital linen at the new price and 45% hotel linen at A$2.05/kg, so its average revenue per clean kilogram will be the blend of the two. The owners’ second fundamental objective is to stay profitable, and they set the level of profit as today’s margin: 11.8% on revenue, which they round to 12% and want to hold after the price cut. The cost per kilogram that leaves a 12% margin is the blended price multiplied by 0.88.

Figure 1The cost ceiling derived from the price and the margin

blended price per kg = 0.55 × hospital price + 0.45 × hotel price

cost ceiling per kg = blended price × (1 − required margin)

0.55 and 0.45

Hospital and hotel shares of clean kilograms

hospital price

A$1.76/kg, the expected tender price

hotel price

A$2.05/kg, unchanged

required margin

12%, today's 11.8% held after the price cut

1. Blended price: 0.55 × 1.76 + 0.45 × 2.05 = A$1.8905/kg
2. Ceiling: 1.8905 × 0.88 = A$1.6636/kg
3. Target: A$1.66/kg, 3.0% below the baseline A$1.711; A$158,504 a year at 3,120,000 kg

The target is A$1.66/kg or less. At the baseline volume of 3,120,000 kg it requires the annual cost to fall from A$5,337,704 to no more than A$5,179,200, a reduction of A$158,504 a year. The reduction is 3.0%, smaller than many people would guess for a project prompted by cost pressure, and that is useful information in itself. The cost problem is real but modest; the delivery problem is large.

The ceiling includes the capital charge at 8% a year on any new plant and building, because the definition in Chapter 3 includes it. A project that spends heavily on capital must save enough in operating cost to pay that charge as well as meet the ceiling.

Water and energy: the tender’s thresholds

The hospital group’s tender awards sustainability points on water and energy intensity, reported monthly with the method stated. Full points are given at 10.0 L/kg of water or less and 5.0 MJ/kg of energy or less, assessed on the average over the reporting period; higher intensities earn partial points on a sliding scale. These thresholds are fictional, like every number in the case, but the structure is common in tenders that score environmental performance.

The targets are set at the thresholds: water ≤ 10.0 L/kg and energy ≤ 5.0 MJ/kg. From the baselines of 14.0 L/kg and 6.00 MJ/kg, they are reductions of 28.6% and 16.7%. Going below the thresholds earns nothing more in the tender. It still saves money, but that saving is already counted in the cost target, so there is no reason to set the water and energy targets lower than the purpose requires. The target is the threshold, not a round improvement on the baseline that might fall either side of it.

Energy needs one more piece of definition. Chapter 4 showed that the season moves energy intensity by about 0.45 MJ/kg between summer and winter. A target judged on a single month would be easy to meet in January and hard in July. The target is therefore judged, as the tender judges it, on the reporting-period average, with each month’s figure published alongside it; and for the laundry’s own comparison with the baseline, each month is compared with the baseline for the same season.

Minimum acceptable and stretch

Each target has a minimum acceptable level, the value at which the purpose is just met. Only OTIF has a stretch. For water and energy, the tender gives no further points below the threshold. For cost, any saving below the ceiling is simply more margin, welcome but not something the project must achieve. For OTIF, the 50% chance of failure at exactly 98% makes a margin of safety part of the purpose itself.

Figure 2OTIF: baseline, present process and target band

Baseline weeks, lowest to highest86.7% to 94.4%

Baseline, 26 weeks91.2%

Control limits of the present process84.9% to 97.5%

Draft target, baseline plus 5 points96.2%

Target band, minimum to stretch98.0% to 99.0%at most 93 to 46 failures in 26 weeks

84889296100

Periods and target dates

A target without a period and a date cannot be judged. The laundry’s dates come from the tender. Taking month 0 as the July in which the project is approved:

• the evidence window is months 6 to 11, the 26 weeks from January to June, and the OTIF target is judged on that window, with the result submitted with the tender;
• water and energy are reported monthly through the same window and judged on its average, with the method stated;
• cost is judged on a rolling 12 months, with each month of the window annualised as an early check;
• the tender closes at month 12, and the current contract ends at month 18.

Figure 3The dates the targets are set against

1. Month 0ApprovalJuly
2. By month 5Every change in place and settledthe last month before the evidence starts
3. Months 6 to 11Evidence window26 weeks, January to June, 4,680 deliveries, at most 93 failures
4. Month 12Tender closesOTIF, water and energy evidence submitted
5. Month 18Current contract endsnew contract starts at the tender price

The dates impose a constraint that no option can escape. A change that is not working by the end of month 5 cannot contribute to the evidence the hospital group will judge. It may still be worth making for its own sake, but it cannot help win this tender. Chapter 8 applies that test to each option and finds that it decides more than any other single number.

SMART as a writing check

The best-known checklist for objectives began with George Doran’s 1981 article in Management Review. In Doran’s wording, an objective should be specific, measurable, assignable, realistic and time-related. Later versions changed some of the words, most often to achievable, relevant and time-bound, and the acronym is now used loosely. It is a useful check on how a target is written, not a method for choosing its value: a round-number target can be perfectly SMART and still wrong.

UK appraisal guidance puts SMART objectives at the start of option work. HM Treasury’s Green Book (2022) asks for SMART objectives to be set before a longlist of options is drawn up, so that options are generated and filtered against them. The NSW Treasury Benefit Management Plan template, which accompanies the TPG24-31 guide that Chapter 4 used for the baseline, records a target and a target date for each benefit beside its baseline measure. Both point the same way as this chapter: targets are fixed before options are compared, and they carry a date.

Checked against Doran’s words, the laundry’s targets pass. They are specific: each names the measure as defined in Chapter 3. They are measurable: each has a number and a period. They are assignable: each will have one owner, named by role in Chapter 7. They are realistic in the only sense that matters here, because each is the level the purpose requires, and Chapter 8 shows that at least one option meets all four. And they are time-related: each is judged over a stated period that ends on a stated date.

The error: targets from the baseline

This is an illustration. The first set of targets, drafted before the purpose had been worked through, was written as improvements on the baseline: “OTIF +5 points”, “cost −10%”, “water −20%” and, for energy, “reduce where possible”. Each sounded ambitious and none was derived from anything.

Figure 4Draft targets against targets derived from the purpose

	Draft, from the baseline	Derived, from the purpose	Effect of the draft
OTIF	+5 points: 96.2%	≥ 98% over months 6 to 11	Allows 178 failures in 26 weeks against 93; the tender is lost
Cost per clean kg	−10%: A$1.540	≤ A$1.66	A$0.120 stricter than needed; rejects options that meet the purpose
Water	−20%: 11.2 L/kg	≤ 10.0 L/kg	Misses the full-points threshold
Energy	Reduce where possible	≤ 5.0 MJ/kg	No level and no date; cannot be judged

The faults were found when the operations manager put the same question to each draft: what happens if the project hits exactly this? An OTIF of 96.2% allows 178 failures in the evidence window, nearly twice the 93 the tender permits; hitting the target exactly would lose the contract. A cost of A$1.540/kg is A$0.120 below the A$1.66 the purpose requires; hitting it would be welcome, but requiring it would reject any option between A$1.540 and A$1.66, and Chapter 8 shows that the option that meets every other target sits in exactly that interval. Water at 11.2 L/kg misses the full-points threshold by 1.2 L/kg; hitting it would lose tender points the laundry could have had. And “reduce where possible” for energy could be met by any saving at all, so it set no requirement.

The correction was to replace all four with the derived targets and to record, beside each, where its value came from: the tender clause, the price and margin calculation, or the scoring threshold. If the source changes, the target changes with it, and nobody needs to argue about whether a round number is ambitious enough.

Targets as decision rules

Derived targets do more than describe success. Because each one is the level the purpose requires, each can be used as a pass or fail test on every option and, later, on the project as it runs. Stopping rules and kill criteria written before the evidence in Stage Gates and Stopping Rules by the Numbers shows how criteria like these become stopping rules at a gate, fixed before the evidence arrives so that nobody can move them afterwards. Stages and gates: organising the front end in Front-End Planning by the Numbers places target setting where it belongs, among the deliverables of the front end, before the project is committed.

The four targets now complete the laundry’s register: OTIF at least 98% over months 6 to 11, at most 93 failures; cost at most A$1.66 per clean kilogram; water at most 10.0 L/kg and energy at most 5.0 MJ/kg on the reporting-period average. Options derived from the outcome, and why the choice changed scores every option against all four. Holding acceptance and benefit tracking to the outcome follows the count of failures through the evidence window month by month.

Exercise: a lower hospital price

Late in the tender process, the hospital group signals that it expects A$1.72/kg rather than A$1.76. The hotel price, the volume shares and the required margin of 12% are unchanged. What is the new cost ceiling, how much annual saving does it require at baseline volume, and would the option Chapter 8 recommends, at A$1.625/kg, still pass?

Worked answer. The blended price becomes 0.55 × 1.72 + 0.45 × 2.05 = A$1.8685/kg. The ceiling is 1.8685 × 0.88 = A$1.6443/kg. Against the baseline of A$5,337,704, the annual cost at baseline volume must fall to no more than 1.6443 × 3,120,000 = A$5,130,216, a saving of about A$207,488 a year instead of A$158,504. The recommended option’s A$1.625/kg would still pass, by about A$0.019/kg. The margin of safety is thinner, which is a reason to watch the price signal closely, but the choice of option would not change. Because the target was derived from the price, the new value took one line of arithmetic to find; a round-number target would have given no guidance at all.