Launching or optimising a new market

When you'd do this

Opening in a new shopping centre entirely, adding a language inside a country you already serve, or going back to fix a branch that was only ever translated once and never revisited since.

What you need before starting

• An actual business reason, and a rough sense of what to expect
• Someone who genuinely uses shops in that language day to day, not only someone who can translate words
• Which languages that particular country's branch actually supports
• Your current shopfront, exported, from your home country
• What rivals' branches in that country actually look like
• Agreement on which tier this country gets: full attention, or simply correct and unembarrassing

What to actually do

First, decide whether you need a whole new branch, or just another language inside one you already have. Every country's branch supports a default language plus a handful of others, and this varies a lot: some support many extra languages, some barely any. Check whether the cheapest win is simply adding a language inside a country you already operate in, rather than opening somewhere entirely new. Often it is.

Then find out what shoppers in that country see today, before you touch anything. Where no language matches, the centre falls back to the country's own default, or to your home language. Find out which one that actually is for this market. It's frequently your home-language shopfront showing up untouched, and that's your real starting point.

Then research your filing words in the local language, entirely from scratch. Not "translate the list and see how it reads". Start from rivals' branches in that country, what local shoppers actually write in reviews, the directory's own suggestions with the country properly set, and your own complaints from shoppers in that market.

For a language written in a very different script, count what your filing-word field actually holds before you plan around any assumed limit. Don't assume it behaves the same as it does in your home language.

Then write the shopfront separately for each centre. On Apple's side: your sign, the line beneath it, filing words, leaflet, all set per country, and all locked to a refit. On Google's side: your sign, pitch line and leaflet, all indexed, all visible, and written differently from Apple's side entirely, this isn't a straight copy across centres either.

Then translate at least your first few window photos, as a minimum. The captions do most of the actual work. Swapping the caption text over your existing pictures captures most of the value without a full reshoot. For a market read right to left, the whole order of the sequence reverses too.

Then, if the country justifies it, check uncertain words with a paid billboard first. A billboard running in that country tells you which words real local people actually type, and which of them convert, before you spend your limited filing-word space finding out the hard way.

Then open, and write down the date. Apple's side ships with your refit; Google's side goes through its own inspection, plan for around a week.

Then set yourself a date to come back and check. A few weeks out. A new country's first shopfront is a guess, not a finished piece of work.

What "done" looks like

• The country's supported languages are actually known, and the choice was made on purpose
• Filing words researched in the local language, not translated
• Both centres written up separately, from one shared brief
• At least the first few window photos translated
• A starting point exported and dated
• The filing-word field's actual local-script capacity checked and written down
• A follow-up date sitting in the calendar
• The country's tier written down somewhere

What tends to go wrong

Translating the filing-word field instead of researching it fresh. Produces perfectly correct words that nobody actually types.

Treating one language as one market. The same language spoken in three different countries is three different vocabularies.

Translating the words and leaving the window photos in your home language. The most visible, most obviously half-finished version of this job.

Opening in countries you won't actually keep maintaining. A stale, abandoned branch is worse than no branch at all.

Assuming a non-Latin filing-word field holds exactly as much as the Latin one does. It's genuinely unresolved, go and check it yourself.

Leaving your home-language shopfront wrong. It's the fallback for the entire world, not just for home shoppers.