The 100 Metres Problem

On one Australian transport contract portfolio, the difference between compliant and in default came down to 100 metres!

The contract measured whether a service departed a stop on time. The tracking system recorded two points, the first one adjacent to the stop, and one on an outer boundary that could sit up to 100 metres away. For years the inner point was used. Then the calculation was changed to the outer one. The same vehicles, on the same trips, moved from late to on time, and six figures of penalties were reversed or never applied.

Nothing about the service changed. Not one passenger waited a shorter time.

That story is usually told as a governance failure. It is more useful as a diagnosis, because it explains something that puzzles a lot of executives on why a contract can have twenty-two performance measures, a monthly reporting cycle, a penalty regime and a dedicated contract management team, and still deliver performance that gets worse every year.

The problem is rarely the data. It is who owns it.

In that same portfolio, worth roughly $8 billion across nine contracts and six providers, twenty-two measures sat in the contract. Providers were responsible for measuring and reporting eighteen of them. Every figure they reported was derived from the client's own systems.

Read that twice, because it is the whole problem. The provider carries the reporting obligation and the financial consequence, but has no independent source against which to test the number. The client holds the data but has neither the capacity nor, in practice, the procedures to interrogate what its own systems produce. So the number is authoritative to nobody and binding on everybody.

What follows is predictable most of the time. Over one seventeen-month window, providers met the punctuality target in fifteen per cent of reported months. Complaints per hundred thousand boardings roughly doubled over two years. More than eight in ten penalties traced to services not run, or not run on time. Over three hundred claims for relief were lodged on a single network, some outside the ten business day window because the cause of the incident was not yet known, and some approved anyway without a documented delegation to do so.

None of that is dishonesty. It is what happens when a measurement system is asked to carry commercial weight it was never stress-tested to hold.

The costs nobody puts in the business case

Penalties are the visible cost, and they are the smallest one.

The larger cost is time. When the number is contestable, the monthly performance meeting becomes a reconciliation exercise. Senior people on both sides spend the month establishing what happened rather than deciding what to do about it. Contract management vacancies meant single managers were carrying two contract regions instead of one, and the team was fully occupied with monthly reporting. The strategic work, the trend analysis, the root cause work, the service redesign, simply did not happen, and everything think the operator is doing this.

The second is the fix that gets blocked. If a provider wanted to vary a timetable to improve punctuality, and its punctuality was below target, the client would generally not fund the change for the first two years. The provider could not improve the number without the variation, and could not get the variation because of the number. Both parties knew it. Neither had a mechanism to resolve it.

The third appears in no ledger. Complaint topics were never systematically analysed, so nobody could direct anybody to fix a recurring cause. Services not operating, vehicles missing stops, and late arrivals made up roughly half of all complaints. The information needed to fix the service existed the whole time. Nobody had the capacity to read it.

Why performance uplift programmes underdeliver

Most uplift programmes on these contracts start on the operational side, like a rosters, depots, maintenance, workforce availability. That work is often necessary and it is rarely sufficient, because it assumes the performance picture is reliable enough to act on.

On the contracts we see, it usually is not. The measures are contestable, the causes are unanalysed, and the decision rights over relief, waivers and variations are undocumented, which means every improvement proposal enters a negotiation instead of a process.

You cannot improve what you cannot measure credibly. You cannot direct a fix when the cause has never been established. And you cannot hold a provider to account for a number that neither party can independently defend.

That is the argument for assurance, and it is a commercial argument rather than a compliance one. Assurance is usually pictured as the audit at the end. On performance-based contracts it works better as the thing that makes uplift possible in the first place. In the portfolio above, a capability review had already identified the contract management gaps, at a cost of around $700,000. The recommendations were not implemented. Three years later an independent review found the same gaps, alongside performance below target in the large majority of months. The assurance capability was the first thing under-resourced and the only thing that would have caught any of it.

What good looks like

Three things separate the contracts that improve from the ones that argue.

Provenance. One agreed source of truth, with both parties able to interrogate it. That means documented lineage from sensor to report, a defined process for challenging a figure, and at least one measure where the provider can bring independent evidence. Without this, every performance conversation starts with a dispute about the premise.

Definitions. Measurement rules stress-tested before they start costing money. Walk each measure through three or four real scenarios and ask what the number does. The 100 metres case is not exotic. It is what happens when a definition is drafted once, agreed commercially, and never tested operationally. This is cheap to do at contract award and expensive to do in year three.

Decision rights. Who can grant relief, waive a penalty, extend a time bar, approve a remediation plan or fund a variation, at what value, and on what evidence. Where these are undocumented, decisions still get made, but they get made informally, and they are indefensible when someone eventually looks.

Five questions worth asking this week

1. For each measure, can you name the system the number comes from and the last time its calculation was tested?

2. If the provider disputed a figure tomorrow, what is the process, and how long does it take?

3. Who has authority to waive a penalty or extend a time bar, and is it written down?

4. When was the last time anyone analysed the causes of customer complaints, rather than counting them?

5. Are your contracted service levels still based on the demand assumptions used when the contract was signed?

If you cannot answer three of those five, the performance regime is likely generating disputes rather than improvement, and no amount of operational effort will close the gap.

How Dolanto works on this

• We run the Data Quality report to understand the state of play.

• We test performance regimes against the way they actually behave, tracing each measure back to its source system and its contractual definition.

• We map decision rights across relief, waivers, time bars, remediation plans and variations, and identify where authority is being exercised without a documented basis.

• We turn assurance findings into a prioritised uplift plan, so the evidence leads to a decision rather than to a report.

Smarter Assurance. Stronger Infrastructure.

---

Follow the series

We publish one of these a week, on the contract mechanics that quietly decide whether a major project performs. Follow Dolanto on LinkedIn, or subscribe to our website, and we will send each piece as it publishes.

Figures in this article are drawn from publicly reported independent performance reviews of Australian performance-based service contracts. No client information is used.