£100 poorer

Today we put down the deposit on a van.

It’s a 2012, long wheelbase, high top Mercedes Sprinter panel van, with 98,000 miles on the clock.


Is it a good deal?

I’m spending a lot of time on the Autotrader website, looking at near-identical pictures of Mercedes Sprinters. When I’m doing this, I’m trying to answer one of two related questions:

  • Is this van good value for the condition it’s in?
  • How does it compare in value with other vans I’ve seen that look similar?

But there are so many of them. They’re all white, they’re all in much the same condition. After a while they all blur into one another.

I love to nerd out with systems and technology, and I’m always interested in defining methods for solving problems, rather than ad hoc approaches. And this problem is a great candidate for such an approach. That ‘blurriness’ is a factor of the quantity of data, and its homogeneity. However, those factors also favour a data analysis approach – albeit a pretty casual one.

There are many considerations to weigh up when looking for a van for a campervan conversion. But we’ve fixed many of the variables:

  • Model: Mercedes Sprinter
  • Long wheelbase
  • High top
  • NCV (post-2006) variant
  • From a dealer, not a private seller
  • Under 100,000 miles on the clock

There are lots of Sprinters on the market with these specs, so it’s easy to compare them across the other factors that we’re considering:

  • Price
  • Mileage
  • Age
  • Condition

Condition seems to be mostly a factor of age and mileage, so that gives us three variables to compare. So I made a spreadsheet to record some basic information about each van:

  • URL or identifying label
  • Registration plate, location, or other comments
  • Mileage
  • Year
  • Price (inc. VAT if applicable)

(Interestingly, VAT-free vans don’t seem to be much better value. It looks like the sellers are recovering the 20% ‘discount’.)


I wanted to give each van a score to indicate how ‘good’ it is – a factor of how new, and how few miles are on the clock. In order to compare these very different values I needed to normalise them to a standard range. Fortunately, there’s a formula for that:

zi = (
   xi − min(x)
   ) ÷ (

Bear in mind when doing this calculation that higher numbers are better for year of manufacture, but lower numbers are better for mileage. So ‘max’ and ‘min’ could be more usefully substituted for ‘best’ and ‘worst’

That gives each van a score of 0 – 1 for year, and 0 – 1 for mileage. Combining (adding) these gives an overall score for each van.

But those factors aren’t equally important, so I also included a variable weighting modifier. For example, if we think that mileage is twice as important as age, we can multiply the mileage score by 2 before adding the two values together. Then I multiplied the result by 100 to make the numbers easier to read.

The final ‘how good is this van’ quality score is calculated like this:

Quality = 100 * (
   (Mileage Weighting * Normalised Mileage Score)
   (Year Weighting * Normalised Year Score)

Using a weighting of 1 for the age and 2 for the mileage, this gives every van a score of 0 – 300.

Plotting a trend

Putting this quality score on a scatter chart mapped against asking price gives a picture of how all the vans vary in value for money.

Adding a trend line allows us to see which vans are good value for money across the price range. Everything below the line is comparatively cheap, and can be easily contrasted with other vans that are of the same ‘quality’ or the same price.

It’s also easy to set a price limit (draw a horizontal line across the chart at the maximum price we willing to pay) and identify all the good value vans below this limit.


The more data we can collect the better. It helps to reduce the impact of outliers on the overall picture – outlier data can skew the average and make the trend line misleading. For example, private sellers are often much cheaper, but not necessarily where we want to buy from. Adding lots of cheap vans from private sellers pulls the trend line down, making the other vans seem much worse value (which may be accurate if you’re happy to buy from a private seller). Conversely vans from Mercedes dealerships come at a premium, and push the trend line up.

On the flipside, the chart makes it very easy to see outliers, and then prompt further research to explain them.

Perfection is the enemy of good

I’ve found this to be a useful tool to help make sense of all this data. But I’m conscious that fiddling around with spreadsheets and scatter charts could just be my way of stalling; of avoiding making a decision – a commitment.

To some extent this approach is about managing uncertainty and the discomfort that comes with it. We want to buy a good van, at a good price, and it’s uncomfortable negotiating that purchase process when we’re not experts in the field. But you can’t remove all risk, you can’t avoid uncertainty in life (that’s called ‘being dead’). At some point we have to let go.

xkcd: Survivorship Bias

Coincidentally, also via Hacker News, this apt comic from xkcd: Survivorship Bias.

I assume the poster’s intent (on Hacker News) was to make a comment about inspirational startup speeches. But the same could apply to the ‘drop everything and follow your dreams’ vanlife brigade.

Vanlife in the New Yorker

Appearing the in the New Yorker is a notable event in itself. This is some kind of media attention milestone, like a mention on The Late Show. Also notable that I discovered it on Hacker News, the unofficial developer / startup community message board.

The piece somewhat begs the question – it defines vanlife as having a certain demographic, and then slams those living it for conforming to that definition:

There is an undeniable aesthetic and demographic conformity in the vanlife world. Nearly all of the most popular accounts belong to young, attractive, white, heterosexual couples. “There’s the pretty van girl and the woodsy van guy,” Smith said. “That’s what people want to see.” At times, the vanlife community seems full of millennials living out a leftover baby-boomer fantasy: the Volkswagens, the neo-hippie fashions, the retro gender dynamics.

If you look for that, you’ll find it. But it’s just lazy to say that this is the vanlife aesthetic and demographic.

Photography by Jeff Minton for The New Yorker
Scrolling through King and Smith’s Instagram feed in chronological order, you can see the couple become better at tailoring the images to what their followers want. “They want to see Emily in a bikini, they want to see a sunflare, they want to see the van,” Smith said.

But the connection with social media – the uneasy blurring of life and lifestyle – is more interesting:

They decided to use their extra day in Ventura to take a photograph for one of their newest sponsors, “Outsiders,” … Smith had a particular image in mind: King sprawled in the back of the van, reading a book about Ayurveda with Penny nestled next to her, and an “Outsiders” decal featured prominently on her laptop. As Smith shot from the front seat, King tried a few different positions—knees bent; legs propped up against the window—and pretended to read the book. “Sometimes it’s more spontaneous,” she said apologetically.

“It’s about storytelling, and when you’re telling a story it’s not always spontaneous,” Smith said. “Lift your head up a little bit more, look like you’re reading.”

King positioned Penny at her feet, but the dog kept moving, distracted by grebes bobbing on the waves. Smith grew frustrated by the strong contrast between the dim van interior and the bright ocean beyond. King attempted to placate him. “Corey, this is O.K., this is O.K., this is fun,” she said.

After more than half an hour, Smith got a shot he was satisfied with. The next day, as he drove in the rain to Los Padres National Forest, King sat in the back and fixed the overexposed ocean in Photoshop. The post, when it went up, looked cozy and relaxed. King added a long caption, about how living in the van had made her reconsider what “work” actually means. “I no longer define work by money, instead seeing it as our focused action collectively creating our world,” she wrote. “Currently my work is storytelling and aligning with companies supporting our lifestyle and Earth.”

“Such a beautiful lifestyle,” one commenter wrote. “This looks like heaven,” another said.


Read the article here: #Vanlife, The Bohemian Social-Media Movement

Read the hand-wringing arguments about authenticity and self-commodification on Hacker News.

Wood selection for framing and cladding: do the math

I’m evaluating different materials for cladding the walls of the van, and for building frames and panels for furniture (bed, kitchen units, etc.)

Commercial outfitters here in the UK use materials I find unappealing, whereas DIY convertors often stick to what they can find in the out-of-town DIY store. I’d like to weigh up – literally – the pros and cons of a few approaches that are in scope for me.

Absolute weights

Weight is a key factor, not least because of the payload limit of the van. But also efficiency, handling and sizing. I consulted a few sources to find out the weight of different wood types. The excellent Collins Complete Woodworker’s Manual gives ‘average dried weight’ for a range of hard and soft woods. (This is just a selection of woods that are available to me, and might be appropriate.)

Soft woods

  • Sitka and Norway spruce — 450 kg/m3
  • Pines (White, Ponderosa, Yellow) — 420 – 480 kg/m3

To verify, I weighed some 44mm square batons I bought from Wickes (the same product is also available at B&Q). It’s usually labelled as ‘whitewood’ or  spruce. It worked out at 423 kg/m3.

Hard woods

  • European beech — 720 kg/m3
  • American white ash — 670 kg/m3
  • American white oak — 770 kg/m3


My local timber merchant also stocks lovely BB Grade Birch Plywood, and they gave me the specs for a couple of thicknesses:

  • 12 mm 8×4 sheet — 25.25kg (707 kg/m3)
  • 9 mm 8×4 sheet — 19.50kg (728 kg/m3)
  • 6 mm — extrapolated: 749 kg/m3
  • 18 mm — extrapolated: 665 kg/m3

These are the woods and panel materials that are available to me (discounting the crap I don’t want to touch).

Wall cladding

Many DIYers seem to go for either cheap plywood faced with car felt, or wooden cladding boards, made from spruce or maybe pine.

I like a wooden finish, but those cladding boards are either thin or heavy. Here’s a typical review of some 8mm spruce cladding (just labelled ‘softwood’) from a DIY store:

At least 20% of these poor quality timbers where so warped, bent or sub standard that they could not be used for the project. They are only finish sanded on one side witch means that you cant switch boards around to hide knots and holes where knots where.

Large chunks where often taken out of the tounges or groves making them incredibly hard to fit together. Every single piece is labled with an incredibly sticky label that is impossible to peel off on the face side of the board making for a enfuriating few hours of peeling and scrubbing.

This chimes with my experience too. You could go for thicker boards, but then you’re looking at some serious weight. To clad an 8′ x 4′ area with 14 mm softwood cladding would weigh about 18.8 kg. The same area in 6 mm birch ply (which has a lovely finish) would weigh about 13.4 kg.

There are other factors of course, (how well does the cladding conform to the curves of the van, does it need to bear weight or hold a screw), but the weight and aesthetics certainly point to the thin 6 mm ply as a good option.

Furniture framing

I’ve made lots of furniture out of stud timbers and that 44mm section whitewood. It’s fine, and in certain orientations and thicknesses, adequately strong.

But I’m interested in using something that cuts more cleanly, is stronger and has some flex for the construction of cabinet and bed frames. Ash looks promising here. I need to do some practical experiments – and more research, but the weight difference doesn’t look too bad, especially when you consider you could get away with smaller cross-sections of wood for many applications in the van.

What to look for when buying a Sprinter

Some useful advice on common failings of the Mercedes Sprinter:

One of the Sprinter’s most common faults is problems with the diesel engine’s high pressure injection system. It’s called the ‘black death’ though it’s not as bad as is sounds if it’s caught early enough. The seals around the injectors go, causing a chuffing from under the bonnet.

Put the heater and fan on in the cab and if there’s a faint smell of what should be coming out of the tailpipe then they’ll need looking at. A black build up around the injectors also highlights the issue. Depending on how bad it is, you could be looking at anywhere from £150 to £500 fix.

Check the propshaft on any prospective purchase, they tend to fail at about 100,000 miles though they’re not too expensive provided you don’t source it from a dealer where they’ll charge around £800. For all other parts we’d suggest genuine only, particularly brakes.

The oil filter within the engine of 2007 Sprinters can become blocked and this causes the engine oil to be expelled through the dipstick hole (even though the dipstick is in place) leaving the engine to seize.

As with any vehicle though they’re not perfect, one of the most visible problems being their tendency to rust. Bear puts this down to abuse, and the amount of operators running them without ply linings inside, though lots also attribute the Sprinter’s pockmarks to thin, water-based paints. Best advice regarding the bodywork is if it’s scratched or dinged, get some touch-up on it quickly, Mercedes-Benz’s warranty very particular in its cover regarding surface rust.



Bed ideas

Common formats include two facing bench seats with a removable table in between, or a permanent raised bed in the back of the van with lots of storage underneath. I’m more interested in a slide out bed design that only needs one bench to work (plenty of seating for 2 people) and lets you maintain a walk through floor plan.