The engine behind Hopology: How automation delivers the freshest and most exclusive beer

Bij Hopology draait alles om één ding: de jacht op het meest exclusieve en verse speciaalbier dat je in Europa kunt vinden. Maar hoe zorg je ervoor dat een supervers blik dat net in de VS is afgevuld, zo snel mogelijk in onze webshop staat, compleet met een goede omschrijving en alle juiste info? Zeker als je dit naast een fulltimebaan doet, is tijd je grootste vijand.

Vroeger was dit een avondvullend programma. Producten handmatig toevoegen, specificaties overtikken, saaie productteksten schrijven... Tijd die we niet konden besteden aan het belangrijkste: het sourcen van het volgende unieke bier voor jullie.

That had to change. As a hobby project, I decided to automate the entire process. The result is a product pipeline that I'm quite proud of, and it's at the heart of why we're so fast and efficient.

Step 1: The Source - A Simple Google Sheet

It all starts with purchasing. As soon as a new batch of beer is confirmed, we fill out a simple Google Sheet. It's nothing more than this:

Column A: The beer's Untappd URL

Column B: The number of cans (the stock)

Column C: The sales price

Column D (optional): Country of origin (if not the US)

Our 'command center': a simple spreadsheet.

Step 2: The Data Collector - Python & Selenium

Once the list is complete, I run a Python script in a (free) Google Colab notebook. This script is our tireless data assistant. It uses the Selenium library to visit every Untappd URL in the list, just like a real user.

The script 'reads' the web page and retrieves all the crucial basic information:

1. Name of the beer and the brewery
2. The exact beer style (e.g. "IPA - Triple New England / Hazy"
3. Alcohol percentage (ABV)
4. The average rating and the number of ratings

Simultaneously, the script also scans the web for an indication of potential competitors' prices. Essential market information, instantly available.

Step 3: The Creative Engine – The Gemini 1.5 Flash API

This is where the magic happens. The raw data from step 2 is useful, but not very enticing. So, we send that information directly to Google's Gemini 1.5 Flash model via an API call.

The AI has been given a very specific set of instructions by us, as if it were a new marketing employee:

• Write as Hopology: Generate unique, catchy Dutch product descriptions that emphasize exclusivity and freshness.
• Create the perfect page: Structure the description in HTML, complete with a <h3>Proefnotities</h3> heading and a list of the key flavors.
• Optimize for Google (SEO): Write a perfect SEO title and description, so experts can find us more easily.
• Populate the metafields: Choose the appropriate beer variety (e.g., "India pale ale (IPA)") and beer style tags (e.g., Hazy; Sweet; Full ) from a predefined list. It even creates the shorthand style abbreviation ( TIPA ) for our custom field.

Step 4: The final result – A perfect CSV

Within minutes, the script produces a complete list of perfectly formatted product data. All fields are exactly as Shopify expects them. All that's left is to upload this CSV.

The result? What used to take hours of manual labor is now reduced to maintaining a spreadsheet and taking a picture of the can.

This automation is our secret weapon. It gives us the speed to get the freshest "hype beers" online before anyone else can even think about them. It allows us to spend more time on what really matters: hunting for the next great beer for you.

Working with passion is one thing, but working smarter makes all the difference. Cheers!
Follow Yari Bouwman on LinkedIn for more data and AI applications!