Postal codes feel like infrastructure – something that was always there, like street names. They are not. Most of them arrived in the 1960s and 1970s, they were introduced against real public resistance, and postal services spent years advertising them with mascots, songs and films to persuade people to write them down.
The objection was not practical. It was that being given a code felt like being turned into a number, and a surprising number of people said so in print.
Sorting Was the Problem and It Was Enormous
A letter posted in one town and addressed to another has to be read, sorted, routed, re-sorted at every stage, and delivered to one door among several thousand. Until codes existed, every one of those steps depended on a person reading a handwritten address and knowing where it was.
That required sorters with detailed local knowledge, it was slow, and it could not be mechanised at all, because a machine in 1960 could not read handwriting and could not have interpreted an address even if it could.
Mail volumes were rising steeply, labour was the dominant cost, and the only way to increase throughput was to employ more sorters. A code solved it by moving the geographic knowledge out of the sorter’s head and onto the envelope, where a machine could eventually read it.
Like our content? Follow us for more.
What the Parts Mean
Systems differ but the logic is consistent: the code narrows from large to small, left to right.
The first characters identify a broad area – a region, a postal district, a city. The next narrow it to a sector or a delivery office. The final characters identify something very small: a street, part of a street, one side of a street, a single large building, or in some systems an individual delivery point.
That hierarchy is what makes mechanisation possible. A machine at the first sorting stage only needs to read the opening characters to put the letter on the right trunk route, and the later characters are only read when the letter reaches the office that needs them. Each stage reads only as much as it needs.
It is also why a full code identifies a location far more precisely than most people realise. In a number of countries a complete postcode covers a handful of addresses, and in some it covers exactly one.
The Campaigns Were Extraordinary
Persuading the public to add characters to an address they had written the same way their whole lives turned out to be the hard part, and postal services treated it as a mass advertising problem.
They produced animated characters, songs, posters, television films, booklets, school materials and roadshows. Some of the mascots became genuinely well known and are remembered with affection by people who have no idea what they were for. The campaigns ran for years, not months, because adoption was gradual and because every letter without a code still had to be handled the old way.
The persuasion was necessary for a reason that is easy to miss: the system only pays off at high adoption. A sorting machine is worth installing when most of the mail can be read by it, and until then it is an expensive device standing beside the people still doing the work by hand. The advertising was not public relations – it was a prerequisite for the capital investment.
Then Everybody Else Found a Use for It
The postal service invented the code to sort letters. Almost immediately, other industries discovered that a short string identifying a small geographic area was the most useful piece of data they had ever been given.
Insurance priced risk by it. Retailers used it to plan where to open. Lenders used it in credit scoring. Market researchers built classifications of neighbourhood type around it. Emergency services dispatched by it. Delivery firms routed by it. Government allocated funding by it. Election campaigns targeted by it.
None of that was the intention and all of it followed within a generation, because the code did something nothing else did: it attached a short, unambiguous, machine-readable label to a piece of geography small enough to be meaningful.
It also created a genuine problem. Pricing insurance or credit by postcode means treating people as members of a geographic group rather than as individuals, and where those groups correlate with income or ethnicity the practice has been challenged – in some jurisdictions successfully. The term for the historical version of this, applied to lending, is redlining, and the postcode made it administratively trivial.
The Machines That Read Them Got Good Surprisingly Slowly
The code was introduced so that machines could sort the post, and for the first couple of decades the machines could barely do it.
Early mechanisation did not read anything. A human operator sat at a keyboard, read the code off the envelope, and typed it, and the machine sorted on what was typed. That is faster than sorting by hand and by local knowledge, but it still requires a person per letter, and the keying was the bottleneck.
The next step was to print the code onto the envelope as a pattern of phosphorescent bars, which a machine could read reliably at speed even though it could not read the characters. Those faint marks along the bottom of an envelope are the machine’s own annotation, applied at the first sorting office so that every later stage can read it without a person.
Optical character recognition of printed codes came next, then of handwritten ones, which is genuinely hard – handwriting varies enormously and a misread code sends a letter to the wrong city. Modern systems photograph every item, read the whole address rather than just the code, and route by what they can resolve, falling back to a human only for the small proportion they cannot.
So the sequence took about forty years: code introduced, keyed by people, translated into bars, read as print, read as handwriting, and finally the whole address read directly – at which point the code became a helpful redundancy rather than a requirement.
A Code Is Not an Address
The most common practical confusion is worth clearing up, because it causes real problems.
A postcode identifies a delivery area, not a building, and the two do not always line up. A code can cover both sides of a street, several buildings, or a block of flats with one entrance and forty addresses. Two houses with different street addresses can share a code, and a single large building can have several.
Satellite navigation systems that accept a code alone will take you to a point that is a representative location for that code – frequently a centroid, sometimes the first address in it – which is why a navigation system occasionally delivers you confidently to the wrong end of a long road, or to a field.
For a delivery, a visitor or an emergency, the code narrows the search and the house number and street name complete it. Anybody who has waited for a delivery that went to the right code and the wrong building already knows this.
The Codes Change, Which Nobody Expects
Postal codes are not permanent. They are revised when delivery volumes change, when new housing is built, when an office closes or when a sector is split because it has grown too large.
That causes a slow, persistent problem for everybody who built a system on them. Databases hold codes that no longer exist. Historical records cannot be mapped to current geography without a conversion table. Research comparing one area over time has to account for boundaries that moved underneath it.
Statistical agencies maintain lookup files specifically to handle this, and the work of matching old codes to new is a real and unglamorous part of a great deal of social and commercial research. The code was designed to sort post, and sorting post does not require the boundaries to be stable – which is a reasonable design decision that became inconvenient the moment everybody else started depending on it.
Which Is the Pattern Worth Noticing
The postcode is a clean example of something that recurs: a piece of infrastructure built for one narrow purpose, which turns out to be so useful for other purposes that it becomes load-bearing for things its designers never considered.
The same happened to the registration plate, which was introduced to identify vehicles and now underpins parking enforcement, congestion charging, insurance checking and automated policing. It happened to the national insurance or social security number, which was designed for one benefits ledger and became a general identifier. It happened to the barcode, the shipping container and the email address.
In each case the original purpose is modest, the adoption required persuasion, and the eventual dependence is far broader than anybody planned – including dependencies the designers would have objected to.
Which is why the people who resisted being reduced to a code were wrong about the post and not entirely wrong about everything else. The postal service only wanted to sort letters faster. It was the insurers, the lenders and the marketers who turned the thing into a description of you.
Like our content? Follow us for more.
