In today’s Guardian: who will address the postcode mess?
In today’s Technology Guardian, Michael Cross examines the Royal Mail’s postcode address file – such a valuable item, if it could only be used well.
Royal Mail
is coy about how much the Postcode Address File costs to maintain and how much it receives from licensing, or even whether it runs at a profit.
Licensing arrangements for the Postcode Address File are only one part of the addressing imbroglio. Even if the Royal Mail were to give the database away to all comers, addressing would still be messy.
For a start, the postcode file has big gaps. According to Barr, it holds only 60% of buildings in England – the Royal Mail is not interested in structures such as churches, which do not receive mail. Because of the purpose for which they were set up, postcodes may bear little relevance to reality – the initial component, the “post town”, relates to the nearest sorting office rather than the nearest town.
As Ed Parsons (chief technology officer of Ordnance Survey) noted on this blog, the flaws mean that there are three databases in operation – the RM PAF, the OS’s address layer, and the National Land and Property Gazetteer (I think that’s right on the latter).
And why?
This chaos arises directly from the UK government’s policy of encouraging state-owned bodies, which are usually monopolies in their fields, to treat information as an asset to be exploited commercially. Free Our Data argues that this resource should be funded by taxation and made available for free to all takers, to stimulate a vibrant knowledge economy.
Some heads really need to be knocked together, we think.
- The following posts may be related...(the database guesses):
- Postcodes: local authorities vs Royal Mail still arguing; want to sign a petition? (24 May 2007; score: 34.93%)
- Will the Post Office and Ordnance Survey ever agree about house names? (10 July 2006; score: 34.69%)
- Postcodes to be free? But which ones? (9 December 2009; score: 31.65%)
- Postcomm makes recommendations on future of Postcode Address File: it should make a profit (19 April 2007; score: 30.34%)
- Revealed: how profitable Royal Mail's Postcode Address File (PAF) really is (26 April 2007; score: 30.07%)
July 13th, 2006 at 11:29 am
I had been thinking that companies were working with out-of-date information when they include the county name “Clwyd” on letters to me, but having just checked an address on the royalmail.com site, I see *it* *too* says that the address is in Clwyd.
Someone needs to be sacked.
Clwyd was (according to Wikipedia at http://en.wikipedia.org/wiki/Clwyd ) an “administrative county, with a county council” from 1974 to *1996*
I’ve checked details on http://www.flintshire.gov.uk and http://www.wrexham.gov.uk and these certainly show that Clwyd was disbanded during the year 95/96 and the new councils have new names from 1996.
Just try a search for house name “Stoneleigh” in Wrexham and you will see lots of entries showing “Clwyd” … and they have the gall to charge for this data !?!
July 13th, 2006 at 2:49 pm
Just discovered this blog and totally agree with campaign. I can only understand the retention of publicly paid for data if there are reasons of commercial/personal confidentiality or national security. OS mapping data and postal addresses should all be freely available and funded centrally. I’m certain that the economic and technological benefits would more than recompense the nation in the long run. So much innovative public and academic research could be freed up as a result.
July 14th, 2006 at 7:54 am
Breaking News in the Address War
http://www.opsi.gov.uk/advice/psi-regulations/complaints/SO-42-8-4.pdf
July 14th, 2006 at 9:39 am
Some years ago, working for Royal Mail, I asked how much selling the PAF data raised and was given an enthusiastic couple of million. I asked what it would be worth to have every item properly postcoded and was told “Many millions”. I suggested that giving the file away free to all comers would perhaps get more items correctly postcoded and was told “But we make £2 million by selling it”.
In defence of RM I have to say that working out what PAF costs to maintain is pretty difficult because reporting new and redundant addresses is done in marginal time by posties as well as the formal systems wil LAs etc.
July 14th, 2006 at 1:05 pm
I recently tried to get the geographical location of my company’s postcode moved to the new (1995) entrance which is on the opposite side of the site to the previous entrance, because we discovered that visitors using satellite navigation systems and on-line maps were coming to the wrong place.
Royal Mail told me that because the address was changed in 1995 to the correct (new) road, then it was correct as far as they were concerned, and Ordnance Survey told me that because the postcode ‘circle’ centred on the old entrance still covers part of the site, there was nothing they could do. Ordnance Survey suggested it was the responsibility of the company’s buying postcode location data from them to check such things as entrances.
So there appears to be no means of correcting address location data yet Ordnance Survey can continue to make money selling data that is incorrect.
July 16th, 2006 at 10:32 pm
Stuart,
Drop me a email at ed.parsons@ordnancesurvey.co.uk with your companies details and I’ll make sure it gets fixed.
ed
July 17th, 2006 at 3:18 pm
For a truly effective operational address gazetteer, emergency services are required to put immense time, effort & cost in creating a custom address file using various elements of existing data sources and custom entries.
Emergency services receive a variety of calls – some from home addresses, but many from ‘non-addressable’ locations such as the middle of a street, or in an obscure car park, or in the middle of a field. Emergency services are then expected to record an incident along with the necessary location information to inform deployment if required. Sometimes deployment of resources may be required in a few minutes, so any delay in trying to establish a location can cause serious consequences. The location data is also used for back-office analysis to establish hot spot areas for pro-active deployment.
Some of the datasets that emergency services use include – PAF, Electoral Role, Business Names, non-addressable items (such as road junctions or car parks) and others. Aside from the cost of the existing Mapping Service Agreement, other datasets such as Points of Interest, that might hold relevant non-addressable information, costs the emergency services a significant sum to purchase and then an on-going cost to keep the information updated.
The National Spatial Addressing Index could have served a significant role in reducing the need to cross-reference various datasets. For whatever reason or whoever’s fault it may have been, the failure of NSAI has cascaded onto the emergency service clients who now struggle to put together a workable, effective gazetteer that delivers all their needs.
What would I like to see? One gazetteer dataset that does the job lot – addresses (to include full listings for flats and different elements of large buildings such as universities), non-addressable locations (including road segments / junctions & elements such as car parks or landmarks), Electoral Role information, business address listings and others….
Will I get to see this? Well, I could put something like this together today but at a huge cost to the organisation – and the funding is simply not there. NLPG may deliver some of the requirements in the future, but not without a lot more work and collaboration with the Local Authorities to add some of the extra elements (but who are already stretched providing the basic LLPG feed into the National Hub).
Should emergency services be in this position? Of course not, but we are regarded as simply further customers to some location based organisations without the realisation that by not supplying us with all our gazetteer needs, they may just be putting people’s lives in danger.
July 19th, 2006 at 1:28 pm
Malachi
I agree with you. However I doubt that it will ever be a realistic vision to merge all of the different gazetteeers into one product.
I can’t offer a solution to the license costs but we can provide you with a simple way of integrating multiple gazetteer types.
Contact me at steven.feldman@graphdata.co.uk if you wish