# Why choosing which roads to fix is hard

*For somebody who has to defend the output and has not spent a career in pavement management. It is the only page here that is not about using the software.*

## The problem underneath

A sealed road is a structure, not a surface. Under the bit you can see there is a base, and under that a subgrade, and most of what decides how long the road lasts happens down there where nobody looks. By the time a failure reaches the top it has usually been developing for years.

That is the first hard thing: **what you can see is the last thing to happen.** Condition scoring exists because looking at the surface is the only affordable way to ask a whole network how it is doing — and everyone in the field knows it is a proxy for a structural question it cannot answer directly.

The second hard thing is the money, and it is not simply that there is too little. Roads deteriorate slowly and then quickly. A road resealed at the right moment costs a fraction of the same road rebuilt three years later, so the expensive mistake is almost never *fixing the wrong road* — it is **leaving the right one a year too long.** A council that spends its whole budget on the worst roads it has can be doing the most expensive possible thing with the money.

So the useful question is not "what is worst". It is "what is about to get much more expensive", and that is a question about rates of change, which is why a single survey is worth so much less than two.

## What people do instead

Every council already answers this question somehow. Four ways are common, and each is a reasonable answer for somebody — including, for some councils, a better answer than this one.

- **A spreadsheet.** Right for a small network where one person has driven every road and holds it in their head. It is fast, it is free, and it works. What it cannot do is survive that person leaving, and it cannot show a committee why one road beat another.
- **An experienced inspector's judgement.** The most accurate single input anybody in this field has, and better than any model on any individual road. It does not scale to a whole network every year, and it cannot be defended to an auditor who was not in the car.
- **A consultant's report.** Real rigour, a proper method, and a document you can put in front of anybody. It is a point in time, it is expensive to repeat, and the council ends up owning the conclusion without owning the method that produced it.
- **A state or regional authority's system.** Consistent across councils, which is exactly what a state needs. It is built around the questions the state asks, and those are not always the questions a council has to answer locally.

What they have in common is that none of them gets cheaper the second year. That is the only claim this product makes against them, and it is a claim about the second year rather than the first.

---

[All 13 steps](/how-it-works) · [What PaveKeep cannot do yet](/what-it-cannot-do-yet) · [What can be shown, not just said](/proof)

---

Source: https://pavekeep.com.au/how-it-works/why-this-is-hard · PaveKeep · derived content