App makes toast of driver stereotypes

The Intelligent Marmalade service promises to protect good drivers from high premiums. Cherry Reynard road-tests the claims

Cherry Reynard, with daughter Hannah, made many trips to discover her driving could earn her a10 per cent discount on her premiums (Jason Alden)

Everyone thinks they are a good driver. Any evidence to the contrary is dismissed as the fault of the other idiot/ hedge/bollard. Now there is an independent test to prove it. Insurance group Provisional Marmalade has launched an app – Intelligent Marmalade – which links to the GPS system on an iPhone or iPad to monitor driving and give a score.

This is the latest in telemetrics, which records the location, speed and braking performance of a vehicle, not unlike a black box in an aircraft. Supporters have suggested that it is the "future of driving" with driver scores feeding back into insurance premiums, rewarding good drivers at the expense of the bad. Insurers say that savings of 10-15 per cent are not uncommon in trials. Critics have found the concept too "big brother", supported by a natural mistrust of insurance providers.

Car insurance in its current guise provides only a crude assessment of driver skill. It is backward-looking, takes little account of how much is driven and is based on peer group rather than individual behaviour. This is bad luck for those 19-year-old, responsible male drivers and good news for reckless grannies.

With that in mind, a more accurate and nuanced assessment of skill would seem useful. The Intelligent Marmalade app ambitiously describes itself as follows: "Driver behaviour is the future of insurance, allowing good drivers to prove they are good drivers and pay lower premiums. This app allows you to road-test the future."

Road-test the future? This sounds seductively space-age. So, I decided to put my driving to the test.

The app is relatively straightforward. It is currently available for iPhone and iPad with 3G capability. Drivers sign up, press "start" at the beginning of a journey and "stop" at the end. They receive a score out of 100, based on the ride's "smoothness". They are penalised for red and purple accelerating or cornering manoeuvres. Should they wish to show off, they can share it with their friends.

My own excursion into the future encountered some early problems having neither an iPhone nor a car. Nevertheless, these were not insurmountable obstacles in the interests of journalistic endeavour, so I set off in a rented Toyota Aygo with my brother-in-law's iPhone.

However, there were some problems with this set-up – the Aygo doesn't lend itself to dangerous driving. It might have gone over 70mph on the motorway, but I wasn't sure I wanted to try. However, this might have lent some authenticity to the experiment because the app is – for the time being at least – aimed at 18 to 25- year-olds, who presumably don't generally drive Volvo Estates.

Which leads me to the next problem: No matter how much I would like to be, I am not aged 18-25. No one has asked me for ID when buying alcohol for 20 years and I'd probably kiss them if they did. My girl-racer days, such as they ever were, are well and truly behind me. The app is designed to isolate those 18 to 25-year-old drivers who are not driving home at 90 mph along country roads, having had several pints, and offer them cheaper car insurance. It is not designed to separate middle-aged drivers who drive a bit erratically on the school run.

Also, I don't drive very much. So I had to – ahem – if not quite invent journeys, then at least take a few unscheduled trips. My mother was undoubtedly delighted to see me and her granddaughters unexpectedly mid-week. Waitrose may be equally glad of the unnecessary excursions to the shops. I drew the line at heading off in the middle of the night – apparently this is a sure-fire way to rack up a low score.

So, it was an imperfect experiment. Nevertheless, it yielded some interesting results. The system is set up to encourage "smooth" driving with no sudden acceleration or sudden braking. As it says: "The key to improving your score is to drive smoothly; it is as simple as that. Try to avoid harsh acceleration and sudden changes of direction. Leave a large gap between you and the car in front; give yourself plenty of time to slow down and avoid heavy breaking unless you have to in an emergency."

On my first run I did pretty well; A trip to the shops of some 20 minutes or so generated a score of 75 per cent, with no red or purple breaking or accelerating manoeuvres. This was equivalent to a drop of 10 per cent, or £75 a year, in my insurance premiums, though given that I was 15 or so years older than the competition, this did not feel much of an achievement. Indeed, the group specifically says that experienced drivers should not expect to see a similar dip in premiums.

My next run was a trip from London to Hertfordshire to see my mother. This included some 20 minutes of motorway driving. My overall score was higher, shifting up to 78 per cent; presumably because the system had been exposed to more of my immaculate driving. However, I racked up four purple cornering manoeuvres (I know, it sounds like a Carry On film).

With hindsight, these purple manoeuvres were relatively easy to spot. I found myself in the wrong lane turning off a roundabout and had to cross two lanes. I knew it wouldn't show up well at the time. However, just in case I was in any doubt, the app shows you exactly where these manoeuvres have taken place (the St Albans turning for the M1 in my case).

My remaining journeys did not change my score very much. Admittedly, these were generally within London and the opportunities for reckless manoeuvres relatively limited. My theoretical discount of 10 per cent remained the same.

I did occasionally find myself tempted to see exactly what would prompt a "red" flag and what that might do to my score, but the app does have a few inbuilt defences against those who would try to abuse it. It will not record a score of lower than 56 per cent to prevent people trying to achieve a competitively low score, for example.

The app has limitations, though these are mainly the usual round of connectivity issues. It worked much better on an iPhone than an iPad, for example. It occasionally abruptly ended a journey when it moves out of signal for example. However, this is presumably not aiming to be perfect science, but to provide sufficient indication of driving ability for insurers to make a judgement.

It didn't turn up any big surprises, but then, I knew I was a good driver.