getlamppost.app

local business, amplified

How I Accidentally Built Vancouver's Worst Multilingual Website (And What I Learned)

Running a small business in Vancouver taught me one thing: most of our customers don't speak English as their first language. I own a small coffee shop near UBC, and every day I watch international students and tourists struggle to understand our menu, our neighborhood recommendations, and basically everything about getting around this city.

After three years of pointing at menus and using Google Translate to help customers, I decided to do something about it. I was going to create a website that actually served Vancouver's diverse community in their own languages. Not Google Translate garbage, but real content that made sense.

I was wrong about almost everything. But I learned enough to eventually build something that works, and more importantly, I figured out a system that any local business owner can use without spending thousands on developers or translation services.

Why Vancouver Businesses Need Multilingual Content

Let me give you some context that changed how I think about my business. Vancouver has an incredibly diverse population. In Richmond, visible minorities make up the majority. UBC has students from over 150 countries. YVR processes millions of international visitors every year.

These aren't just statistics. These are customers walking past your business every day because they can't figure out what you sell, where you're located, or why they should choose you over the place next door that has pictures on their window.

I started tracking this at my coffee shop. Before I had any multilingual content, many potential customers would walk in, look confused at our menu board, and leave without ordering. They weren't being rude. They just couldn't figure out what we offered quickly enough to feel comfortable staying.

The fix seemed obvious: translate everything. But professional translation is expensive. A simple menu description can cost a significant amount per language. A basic website? We're talking thousands of dollars, and that's before you factor in keeping everything updated when you change your offerings.

So I decided to figure out how to do this myself using AI tools. The first attempt was a disaster. The second attempt was slightly less of a disaster. By the fourth try, I had something that actually worked.

My First Failed Attempt

I started with what seemed like the simplest approach: ask ChatGPT to translate my existing website into Mandarin, Korean, and Spanish. Copy and paste the results into new web pages. Done in an afternoon, right?

Wrong. The AI translated everything word-for-word, which sounds right until you realize that "medium roast coffee" becomes something that literally translates to "middle-sized burning coffee liquid" in Mandarin. Our signature drink, the "Westside Wake-Up," became "Western Death Revival" in Korean.

But the real problem was that I had no way to know any of this was happening. I don't speak Mandarin or Korean, so I published pages that probably scared more customers away than they attracted.

The website looked professional enough. Clean layout, proper fonts, everything formatted correctly. It wasn't until a regular customer (a grad student from Beijing) told me that my Chinese menu was "confusing and a little scary" that I realized I'd completely missed the mark.

What I Should Have Done Instead

Here's what I learned after talking to actual multilingual customers and doing this the right way:

First, you can't just translate content. You have to adapt it. A coffee shop menu isn't just a list of drinks and prices. It's communicating your brand, your neighborhood, what makes you different. Direct translation kills all of that.

Second, you need feedback loops. You can't publish content in languages you don't speak without having native speakers review it. This sounds expensive, but I found ways to make it work on a small business budget.

Third, start small and specific. Don't try to translate your entire website on day one. Pick one language, pick one specific thing (like your menu or your location/hours page), and get that right before moving to the next piece.

The Step-by-Step Process That Actually Works

After six months of trial and error, here's the system I use now. It costs me a reasonable amount per month in tools and maybe four hours of work to maintain content in three languages.

Step 1: Pick Your First Language

Look at your actual customer base. Don't guess based on neighborhood demographics. Track it for two weeks. Ask customers during checkout, or just pay attention to what languages you hear in your shop.

For my coffee shop, I assumed it would be Mandarin because we're near UBC. Turned out Spanish was the bigger opportunity. Lots of families in the neighborhood, plus international students from Latin America who were being underserved.

Step 2: Start with Your Most Important Page

Don't start with your homepage. Start with whatever page answers the question "What do you sell and how much does it cost?" For restaurants, it's the menu. For retail, it's your product listings. For services, it's your service descriptions and pricing.

Write this content specifically for your target language audience. Don't translate your existing content. Sit down and think: if I was writing this for Spanish speakers in Vancouver, what would I say?

Step 3: Use AI as a Writing Assistant, Not a Translator

Here's how I actually use ChatGPT now. Instead of asking it to translate, I give it context and ask it to create:

"I run a coffee shop near UBC in Vancouver. My customers are mostly international students and local families. I want to describe our breakfast menu to Spanish-speaking customers in a way that feels welcoming and explains items they might not be familiar with. Here's what we serve: [list items]. Write this as a natural Spanish menu with brief, friendly descriptions."

This gives me content that's written for Spanish speakers, not translated for them. Big difference in how natural it sounds.

Step 4: Get It Reviewed Before Publishing

This was my biggest breakthrough. You don't need to hire professional translators. You need native speakers to tell you if your content makes sense.

I found reviewers three ways: posting on UBC Facebook groups offering free coffee for 15 minutes of feedback, asking regular customers who speak the language if they'd help in exchange for store credit, and using language exchange meetup groups where people are happy to help small businesses.

Most people want 15-20 minutes to read through a menu or service description and tell you if it makes sense. In exchange, I offer store credit. Way cheaper than professional translation, and I get feedback from people who are actually my target customers.

Step 5: Test and Measure

Put up your new language content and track what happens. I made simple signs in Spanish saying "Menu disponible en español" and watched how many people asked for it. I also tracked sales during times when my Spanish-speaking customers typically came in.

Within two months of adding Spanish content, my weekday morning revenue increased noticeably. Not huge, but enough to see the difference. More importantly, I started seeing the same faces regularly instead of one-time visitors who looked confused and never came back.

Tools I Actually Use

Here's my current tech setup. Total monthly cost is quite reasonable.

ChatGPT Plus: For creating initial content. I use GPT-4 because it's significantly better at understanding context and cultural nuances than the free version.

Squarespace or WordPress: For hosting the website. Both have decent multilingual plugins now. Squarespace is easier if you're not technical. WordPress is more flexible if you want to customize.

Google Fonts (free): Make sure your fonts support the characters you need. Some languages require specific font files that aren't included in standard web fonts.

Grammarly Business: They added basic Spanish and other language checking. Not perfect, but catches obvious errors.

Common Mistakes I Made So You Don't Have To

Publishing without testing: Always test your content on actual devices your customers use. What looks good on your laptop might be unreadable on a phone with different language settings.

Forgetting about cultural context: Direct translation misses cultural references, local slang, and expectations. A "small coffee" in Vancouver is different from a "small coffee" in Mexico City.

Not updating multilingual content: When you change your English menu, you need to update all language versions. I forgot this for three months and was advertising different prices in different languages.

Overcomplicating the design: Keep it simple. Focus on clear, readable content over fancy design elements that might not work properly with different character sets.

Results After One Year

My coffee shop now serves customers in English, Spanish, and Mandarin. A significant portion of our daily customers use the non-English materials regularly. Revenue increased meaningfully year-over-year, but more importantly, we became a neighborhood gathering place instead of just another coffee shop.

The Spanish-speaking families who come in on weekend mornings now know our staff by name. The Mandarin-speaking grad students have a corner table they treat like their regular study spot. These are

Get new posts

Subscribe in your language

New posts delivered to your inbox. Unsubscribe anytime.

Receive in:

This blog is hosted on Hostinger — fast, affordable hosting I use for all my domains. If you're building something, it's worth a look.
Connect Pinterest
🌐 Read Your Way Fluent $1

Enjoying the language switcher?

Learn any language faster with the widget you are already using

Get the eBook — $1