Which vans hold up best, and which only start well?

All 32 vans ranked

Ranked on the three-to-six-year band, where every model has tens of thousands of tests behind it and where most used van purchases happen. The two columns to the right show what becomes of that position later.

Share of tests ending in a failure, by age band. Change is the difference in percentage points between the first and last columns. † marks a twelve-plus figure resting on fewer than twenty thousand tests, where the model is too recent for the band to be reliable — see the note below. Models with fewer than twenty thousand tests in total are excluded. Source: PriceMyRepair analysis of the DVSA MOT testing dataset.

Fifteen points between the top and the bottom

The Amarok fails 15.94% of tests in this band and the Primastar 31.63% — a spread of 15.69 percentage points between two vehicles of identical age. That is more than twice the difference between vans and cars as a whole, which sits at 7.30 points at the first test.

Which van you buy matters considerably more than whether you buy a van. And, as the last column hints, the position a model holds in this table is not the one it will hold in six years.

Who declines fastest

Plot how well a van starts against how much its failure rate rises by twelve years, and the points fall along a clear downward line. The models with the best early record deteriorate the most.

The relationship is strong by the standards of this dataset

A correlation of −0.65 across thirty-two models is a substantial relationship. For comparison, mileage at the first test correlates with failure rate at only 0.333, which we set out separately — half the strength, on the variable most people would expect to matter most.

The fitted line has a slope of −0.38. In plain terms, every extra percentage point of failure in the early years is associated with roughly four-tenths of a point less deterioration later.

The two ends

The Mitsubishi L200 starts sixth-best at 20.60% and adds 12.81 points by twelve years, more than any other model. The Nissan Primastar starts worst at 31.63% and adds only 4.39, less than any other. They finish 2.4 points apart, having begun 11 points apart.

The pickup pattern

Five of the six best-starting vehicles in the ranking are pickups. Four of the six fastest-declining ones are the same vehicles. No other body type produces that combination.

† marks a twelve-plus figure resting on fewer than twenty thousand tests. The Amarok and D-Max are recent enough that their oldest band is thin; the Ranger, L200, Hilux and Navara each carry between two hundred thousand and three hundred and twenty thousand tests in that band. Source: PriceMyRepair analysis of the DVSA MOT testing dataset.

Best start, steepest fall

Averaged across the six, pickups add 11.45 points to their failure rate between the three-to-six band and twelve years. The eight large panel vans — Boxer, Crafter, Ducato, Master, Movano, Relay, Sprinter and Transit — add 7.21. That is a difference of over four points in how much ground each group gives up.

The effect is not driven by the two models with thin data. The L200, Ranger, Hilux and Navara all sit above eleven points on hundreds of thousands of tests each, and between them they account for the four steepest declines among vehicles with solid volumes.

Starting position does not survive

In the three-to-six band, pickups occupy five of the top six places. Past twelve years the L200 sits at 33.41% and the Hilux at 33.04%, which puts them below the Ducato at 27.17%, the Boxer at 28.23% and the Crafter at 30.68% — three large panel vans that started well behind them.

A buyer choosing a three-year-old pickup on the strength of its MOT record is choosing correctly for that age. A buyer choosing a twelve-year-old one on the same reasoning is working from a table that no longer applies.

The vans that hold up

At the other end of the decline chart sit the models that change least between six and twelve years. One of them is there for a reason that undermines the whole measure, and two of them are the genuinely interesting cases.

The six smallest increases in failure rate between the three-to-six band and twelve years. All six carry more than a hundred thousand tests in the oldest band. Source: PriceMyRepair analysis of the DVSA MOT testing dataset.

Two vans manage both

The Fiat Ducato starts eighth of thirty-two at 21.17% and declines less than every model except the Primastar and the Boxer. By twelve years it records 27.17%, the lowest figure of any van in the oldest band — better than the Amarok, which started six points ahead of it.

The Peugeot Boxer does much the same thing, starting twelfth at 22.75% and finishing second at 28.23%. These are the only two models on the list that combine a competitive start with a shallow decline, and on the twelve-plus figures they are the two best vans in Britain.

A complication worth naming

The Ducato, the Boxer and the Citroën Relay are the same vehicle. They come off a shared production line and differ mainly in badge, trim and dealer network. If build quality drove these figures, all three would land in the same place.

They do not. The Relay starts at 28.56% against 21.17% for the Ducato, and finishes at 36.60% against 27.17% — a gap of nine and a half points at twelve years between vehicles that are mechanically the same. Whatever produces the Ducato’s record, it is not something built into the vehicle.

Why the table flattens

Every model’s path through the three age bands, drawn on one chart. The lines start spread out and converge, which is the same finding as the decline chart seen from a different angle.

Two measures, same direction

The gap between best and worst falls from 15.69 points to 11.05. That measure depends on two models at the extremes, so it is worth checking against something more robust: the standard deviation across all thirty-two models falls from 3.77 to 2.87 over the same span.

Both point the same way. Vans become more like one another as they age, and the model badge tells you progressively less about what to expect from the test.

What survives the convergence

Eleven percentage points at twelve years is still a large difference — the Ducato at 27.17% against the Transit at 37.27% is the difference between one in four failing and better than one in three. Choosing the model still matters on an old van. It simply matters less than the three-to-six figures would lead you to expect.

The practical version: a ranking built on nearly-new vehicles should not be used to buy a twelve-year-old one, and a ranking built on old vehicles tells you little about what a new one will do.

Using this when buying

The single most useful thing on this page is not the ranking. It is that the ranking changes, which means the table you should read depends entirely on how old the van in front of you is.

Reading a specific van rather than a model average

Everything on this page describes populations. A Transit with a clean record across six tests is a better bet than a Ducato that has failed three times, whatever the model averages say. The full test history for any registration is public, and our history check pulls passes, failures, advisories and odometer readings into one view without a sign-up.

Two things are worth looking for specifically. Repeated advisories in the same category across consecutive tests mean deferred work rather than a one-off, and we set out what happens to those in the van advisory index. Odometer readings that jump or reverse mean something else entirely, and the mileage checker shows the whole sequence.

Context from the other figures

Mileage is the variable most buyers weigh hardest, and across these thirty-two models it explains far less than expected — we cover that in average van mileage, where the highest-mileage model in Britain also has one of the better test records. If you are weighing a van against a car rather than one van against another, the comparison against cars covers that directly, and the equivalent car ranking sits alongside it.

Whatever you buy, the running costs follow the same categories that decide the test. Suspension, brakes and structure account for most van failures, and prices for all of them are indexed under repair costs. Service intervals and test fees are in the service guide and the MOT cost guide, and what the MOT actually inspects explains what a tester is looking for.

We earn a commission if you book through this link, at no cost to you. It does not affect which garages you are shown or what they charge. Our funding page explains how the site pays for itself. You can also check any registration free, with no sign-up.

Methodology and citation

One public dataset, one measure, and seven things this ranking cannot do. Two of them change how the table should be read, and one is the strongest objection anyone could raise against it.

The dataset

The DVSA publishes the complete record of MOT tests as open data. This analysis uses the most recent full snapshot: 829,654,015 tests across 137,840,356 vehicles.

We use tests carried out by DVSA only, excluding those recorded by the DVA in Northern Ireland and by the DVLA, and only outcomes recorded as a pass or a fail. Vehicles with an implausible recorded age are excluded. Thirty-two commercial models carrying more than twenty thousand tests each are reported; anything below that threshold is left out.

The measure

Failure rate is the share of tests in an age band that ended in a failure. Bands are three to six years, seven to eleven, and twelve and over. Change is the difference in percentage points between the first and last of those.

The ranking is ordered on the three-to-six band rather than on an average across all ages, for the reason set out below.

What this ranking cannot do

The objection we would raise ourselves

The Fiat Ducato, Peugeot Boxer and Citroën Relay are the same vehicle from the same production line. If this ranking measured how well a van is built, all three would land together. They do not: at twelve years the Ducato records 27.17% and the Relay 36.60%, nine and a half points apart.

That single comparison shows the ranking is measuring the population of vehicles wearing each badge — who buys them, what work they do, how they are maintained — at least as much as anything engineered into them. We think that is still worth publishing, because a buyer is choosing from that same population. But nobody should read this table as a verdict on manufacturing.

Why there is no survival rate here

A durability figure could be produced by dividing vehicles still being tested by all vehicles ever recorded. We do not publish it, because it mostly measures how long a model has been on sale: a recent arrival scores well and a discontinued model scores badly for reasons unconnected to engineering. Doing it properly means following registration-year cohorts, which we have not yet done.

Free to cite

Every figure here may be reproduced in any publication with attribution. No permission is needed and nothing sits behind a form. If a different cut would be more useful — by region, by fuel, by body type, by single year of age rather than band — we will produce it on request.

We are not affiliated with the DVSA, the DVLA or any manufacturer, dealer or garage group. The analysis, the classifications and any errors in them are ours alone. Corrections are welcome and will be published on this page rather than made silently.