Blog > Postcrossing behind-the-scenes: address validation
As promised, the first post on our behind-the-scenes series is about address validation. This is the process through which we check that all addresses in Postcrossing are (to the best of our knowledge) correct and complete.
"Red alert! Red alert! I think that one over there is missing the postal code!"
Addresses are one of the most important things in Postcrossing. Wrong or incomplete addresses would result in postcards being delayed or not delivered, and general unhappiness - which we would like to avoid.
There is specialized software to process addresses of course, but it's either dodgy, specific to a country or just crazy expensive for our size. So, we chose to do this task ourselves, splitting the job between Paulo, Vicki and I. Each one of us picked a set of countries according to our strengths and learnt the UPU address format of those countries by heart. Most countries have their own format - and some are much simpler than others. We use this knowledge to go over thousands of addresses per week, spotting invalid ones, or just rearranging the order of the elements on the address. Most of the time, addresses will be read by automated sorting machines, which expect them to be written in a certain way. So if every address is formatted as UPU recommends, everything will go faster!
What makes an address invalid? Lots of things:
- Parts might be missing: name, postal code, country name, etc.
- The address might be written exclusively in the local script (Chinese, Thai, Cyrillic), making it hard to write for people who don't understand that alphabet or those without a printer.
- The address might be completely translated into English - which is usually a bad idea. For non-Latin scripts, transliteration works better and faster in our experience.
- The address might not match the account's location/country.
A few postal operators also provide a list of all valid postal codes for their country, giving us the ability to automatically detect typos or invalid postal codes.
"Let's just add the country name here at the bottom... I think it'll help!"
The whole process of validating addresses has been streamlined over the years, but it still takes quite a bit of time, since every single new address needs to be checked and validated. We've been doing it for so long that by now we can tell at a glance whether an address is valid, incomplete or just in a strange order. We might not find all the problems - but we do spot lots of them!
Plus, we've learnt many nice things in the process too! For instance, did you know that an address in China should be written from the biggest component (province) to the smallest one (adressee) if written in Chinese characters, but the other way around if written in Latin characters? Or that there are no street names in Japan? Or that every house and building has its unique postal code in Singapore? It's amazing!
I think that's it for addresses, so I'll leave you with one last tip: make sure that you write the address on your postcard exactly as it is shown - it's already optimized for that country!
I hope you liked this first post on our behind-the-scenes series - we'll get cracking on the next one! :)