The power of postcodes

LibDem MP Lynne Featherstone has an idea. She tells Liberal Conspiracy the one IT project she’d like to see from government would be (if I can paraphrase) an email-bouncing facility, where you’d send an email (for example) to (sic), and it would automatically get forwarded to the relevant coppers. She rightly notes that sites such as WriteToThem go most of the way towards this concept… and indeed, it’s surely the sort of project that’s right up MySociety‘s street (sorry).

Personally, I think Lynne has the right idea, but takes it to the wrong conclusion. As IT projects go, what she describes is relatively straightforward. The headaches would come in terms of (a) requiring the email recipients to keep it all up to date; and (b) the extra work generated. Reading and writing emails takes time. It would be much more efficient, in most cases, to encourage self-service via the web.

The bit Lynne gets 100% right is the power of the postcode. The UK has one of the planet’s more granular postcoding systems, with each of the nation’s 1.8m individual postcodes covering on average 15 houses. In IT terms, that’s a remarkably accurate piece of geocoding data – which virtually every adult in the country knows off by heart. You can stop people in the street, ask them, and they know it. That’s a truly awesome asset. (Which is why Ireland is now adopting a similar system, despite Post Office claims they don’t need it.)

But ask any statistician about postcodes, and they’ll glare at you – citing two problems.

  • Postcode boundaries were originally designed for postal use, and don’t match the boundaries of other statistical or political geographies. I can vouch for this: they don’t even differentiate neatly between England and Wales. But as the introduction of Royal Mail’s Mailsort demonstrates, the postal purpose of postcodes isn’t what it once was.
  • Postcodes change. True, but… Royal Mail issues a ‘postcode update‘ every six months. Their website explains that there’s only been one significant change, affecting only Cambridge, in the last 3 years – a lifetime in IT terms. Hey, it’s not as if they’re recoding the entire nation every other week.

I’ve never seen either of these problems as insurmountable. And I’d argue that the amazing potential stemming from universal awareness of postcodes outweighs the hassle factors.

Postcodes are the country’s greatest example of the Power Of Information. I believe we would unlock significant power if we enshrined postcodes as our key national geography, asking Royal Mail to bequeath them to the nation. All statistical and political geography should be aligned with postcodes, with a commitment not to change them for 10 years, perhaps coinciding with the Census cycle. I don’t care if there are marginally more meaningful statistical boundaries; a flawed system we all understand beats a perfect system nobody understands. Oh, and it’s cheaper too.

With improved accessibility to meaningful local data would come improved accountability. A single online search would reveal who is responsible for what in your local area; and would link to appropriate data showing whether or not they were meeting their responsibilities.

The data is all out there, free of charge in almost all cases – but the chaos of conflicting geographies makes it almost impossible to work with. I don’t believe that’s a defensible position. Power to the postcodes!

Update, 8 July: There’s now a Commons Early Day Motion on freeing postcodes, attracting decent levels of support from Labour MPs. See this post for details…

18 thoughts on “The power of postcodes”

  1. Don’t listen to the statisticians, they know nothing in this instance.

    Having founded and what was once the most popular local site in the UK (, now lying in relative ruins), and entirely post-coded based, I can tell you that the number of postcodes that actually do split boundaries is a fraction of 1% (far less than the number of people who don’t know correctly, say, which constituency they live in, or don’t know they’re own postcode correctly.

    Similarly, yes, postcodes change, but again, the annual drift is miniscule. The correct way to handle this is to remember all old postcodes (not hard, given that the post office supply mapping tables), something again you have to do mostly not for statistical reasons, but because the PO generally fails to communicate those changes effectively to the public.

    The statisticians who say these things are idiots, and should be ignored along with the cartographers who say that only 1m resolution maps are ‘good enough’ for general mapping – in fact postcodes have some great properties derived from their ‘inaccuracy’ – for instance postcode density is an excellent indicator for population density, and a postcode is something you can share without giving up your privacy, although you have to be careful:


  2. Not discounting its obvious usefulness, but the biggest drawback to relying on post codes for geocoding is that you only tend to know your own by heart. The power of local data lies in that grey area between just your house and your whole county.

    Now if only we could teach people their locality codes… (yes, they do exist)

  3. It may be more difficult to process e-mail than web transactions, but e-mail may be easier for users (esp. those using mobile devices). Surely we should be loading the difficulty onto the receiving agency, away from the individual member of the pubic?

    Also building-number+postcode is almost always unique.

    What we really need, though, is ubiquity of accurate and reliable GPS in mobile devices, with very strong “only send (saved or current) location with my permission” security.

  4. @Etienne Just to clarify what I was saying… I didn’t mean the postcode data was free of charge; hence the remark about Royal Mail handing it over to the nation. What I meant was, the statistical & factual data we could present using postcodes is free of charge in almost all cases. I know this to be the case, having worked at National Statistics for 2+ years.

  5. Standing outside a bank where access to the ETMs had been made hazardous because the block paving had been damaged, I called my local council’s call centre from my mobile, selected ‘highways’ from the menu, and the callcentreperson greeted me by asking for my post code. No explanation – and I countered by saying that I was standing outside a bank on such and such a road and I don’t know its postcode but I can give you the postal address number of a nearby building. My report was taken and was lost – phoning again a week later I was told that there was no record of my report (there is more to that story but its not directly relevant here). The public sector has a lot to learn about management of anything broadly classed as public administration.

  6. Just to share a post from Tom Steinberg to the E-Democracy Exchange mailing list: ‘we’re certainly not forking out the cash for something
    that it is a crime isn’t free anyway (ie the locations of our own
    electoral boundaries).’

    If we adopted postcodes as the key geographic building block, we’d know that Constituency X was made up of postcode areas A, B, C. So if we had the postcode boundaries, we’d automatically have the constituency boundaries. You can see the implications of what I’m talking about…

  7. Which is the sort of reason why we started the Free Our Data campaign in Guardian Technology. Been going two years now, making progress. Where’ve you been Simon?

  8. Hey, be fair Charles. It’s far from the first time I’ve talked about this, although it’s maybe been a while since I wrote it. But with Lynne’s piece appearing at Liberal Conspiracy, it seemed like a good time to bump it all up again.

  9. Just some parallel information for info, behind each postcode lies an additional 5 characters (delivery point suffix) that Royal Mail use to sort mail to letterbox level. So there is more granularity there than you might think.

    That said, I know some folks at Royal Mail who would take the line that ‘they own the letterbox’, although i’d like to see them prove it.

    The ‘postcode’ industry runs at many millions per year from direct marketing to insurance pricing; but if the underlying data were freed up I think we’d see the value grow enormously.


Comments are closed.