How much does Google shape your taste preferences?
Writing and Analysis by Michelle Lee, Numhom Techalapanarasme
Design by Eeza Sheren
Data by Yi Kuang Loh, Winston Wu
Every single day, 387,210 people around the world search for 'tom yum' on Google.
Yet, the results they see can be very different. The ingredients, taste, even the meaning of 'tom yum' differ - simply based on the country they’re in.
If you’re a home cook, Google is the first place you turn when looking up a new recipe. But depending on your location, you’ll get different results. That’s because your top results are the most-clicked upon recipes by others around you.
With huge volumes of amateur chefs turning to Google to learn how to cook new dishes, search has a profound impact on how people cook. By looking at the most clicked-upon recipes, we can confidently predict how people from a specific market cook a dish.
Our team took the most popular tom yum Google results from Thailand, the home of the famous soup, and compared them with a version from the fusion embracing foodscape of Australia.
Our findings reveal cultural differences and patterns that could help F&B brands or even home chefs find the winning flavour formulas for local palates.
So how do the dishes differ, and do you prefer Australian tom yum or Thai tom yum?
As the undisputed home of tom yum, recipes from Thailand provide an authentic view of this dish. In contrast, Australia’s thriving foodscape and increasing interest in Asian fusion food shows us the extent to which a recipe can evolve in different countries.
We started by taking the top 10 clicked-upon results for “tom yum recipe” for both Australia and Thailand. The recipes were clustered based on common ingredients to identify and remove outliers.
Once the outliers were removed, we split the ingredients into 4 different groups (Spices, Vegetables, Meat, Soup) and took the average quantity of these ingredients.
Finally, we took each base ingredient and found the closest additional ingredient based on shared flavor compounds, using FlavorDB.
The final recipes
Water (500ml) and chili padis (to taste) were omitted from the diagrams as the quantities were the same across both recipes.
Notable differences between the Thai and Australian recipes included the addition of cherry tomatoes, coconut milk, and tom yum paste in the Australian version.
The taste test
In order to gauge the difference in taste between recipes, our resident Thai strategist, Numhom, tasted each of the recipes. Here are her thoughts:
“This dish smelled a little more fragrant than the other two, I think it may be because of the coconut milk. However, for a creamy tom yum it didn’t quite meet my expectations, and I think a higher milk:water ratio may have helped. Also I’m more familiar with plain cooking milk (e.g. Carnation brand) in tomyums, so I may be a bit biased here. Taste wise I thought it was okay, I actually really liked the tomatoes but wished the dish could be a lot spicier and more sour! Instead I found it too sweet. The Australian recipe also called for garlic, which we crushed and added. This isn’t a usual tom yum ingredient, but I couldn’t really taste the difference it made.”
“This recipe was a little underwhelming at first, as it didn’t smell as fragrant as the Australian recipe. That said, I felt the flavour was more balanced even if it didn’t quite hit the sweet/sour/spice sweet spot. I enjoyed eating this beyond the first sip, and ultimately found it was my favourite of the three. The Thai average also happened to be clear tom yum (no milk!) which was quite surprising to me as the image of tom yum in my head is the creamy, orange soup with spicy goodness. It was also very difficult to resist taste testing this recipe, as I think tom yum soups shine best when you can adjust proportions based on freshness of ingredients (esp limes, as each varies in sourness). We also added chopped saw-tooth coriander to the thai recipe which the Australian recipe didn’t call for. This gives the dish a little extra greeness that makes it fun to eat. As outliers, we also didn’t add tamarind paste and coconut shoots that only appeared in one recipe from our dataset - a shame as they would’ve been great additions to the dish!”
Food science shares a lot in common with Human Centered Data Science: The data points that exist are highly influenced by cultural context, and to create great dishes with this data requires an understanding of the cultural factors that determine consumer taste.
We use open data and the principles of HCDS to help multinational F&B businesses to adapt their dishes to local markets, navigating local market taste expectations.
We also work with their R&D teams to help identify new and unexpected combinations of flavours and ingredients to create innovation inspiration for their teams of chefs. We have helped these international teams overcome cultural biases around of ingredient pairing, flavour combinations and how these can be brought to life in marketing hooks that generate buzz around their brands.
What unexpected ingredients do you add to your tom yum? Want to collaborate with us on more data-driven food explorations? Send us an email at firstname.lastname@example.org!