Things that were so deeply basic and often sentimental (“Mom’s meatballs with brown gravy and mashed potatoes, Dad’s rice pudding–the real, Basque version”) made the first cut as well, with the presumption that we would have to either produce the best version of that dish possible or at the most original interpretation. These were dishes that felt like real things people wanted to eat and we weren’t interested in doing things that seemed like novelty for novelty’s sake.

The first cut yielded 32 dishes, the second cut brought it down to 18, and one evening dinner discussion with Derrick and Soo over pizza and beer (pizza: mentioned 24 times, beer: 6 times) brought our list down to 8. We originally had this overly elaborate scheme to invite each dish’s owner to the event and have them maybe say a word or two about why they chose what they chose. But what if they couldn’t make it? Would we comp them? Would they get a plus one? And how would we make a coherent menu that flowed from this crowdsourced pastiche? Need a soup course? Too bad! No one wants soup on their deathbed! The idea was quickly nixed.

So what’s another way? While we were scanning through the list one by one, we started to see repeats in what people suggested and trends started to appear more clearly each time we read through the list. Mac and Cheese was frequently mentioned, so we started to pull all the different interpretations of this dish (11 in total) aside. Halfway through manually copy/pasting the Nth version of Chocolate, I slouched away from the computer and decided this was way too tedious to sustain.

Then I started to think about how we could use everyone’s answers in some way to create a menu that represented the most popular items on the list. More a meal that represented the trends, not the specific dishes. I recalled on what the news media would do after a State of the Union speech, showing a word count of certain keywords such as “taxes” or “war.” It was a vaguely quantitative way to quickly extract what the president’s priorities were in that large body of text. I thought it could work for us.

Searching for a tool that could help me do this led me to Textalyser. Textalyser is a web service where you can enter free text into an input box and have it spit back a full breakdown of the most frequent words in your text. I took our mailing list, performed a simple text export and pasted it into the Textalyser box. A few seconds  later, it spit out a rank ordered list of most frequently used words, filtering out certain prepositions, punctuation, and other noise. The list broke out the top one word answers, two word answers, and three word answers, and it even gave the option to go even deeper.

Here is the top 20:

  1. Cheese
  2. Chocolate
  3. Chicken
  4. Fresh
  5. Beef
  6. Pork
  7. Fried
  8. Lobster
  9. Wine
  10. Salad
  11. Mom
  12. Red
  13. Sauce
  14. Roasted
  15. Potatoes
  16. Bread
  17. Rice
  18. Lamb
  19. Ice Cream
  20. Butter

There are two items that we edited out of the list. “Something” and “anything” originally showed up in the top 20. Many people took a general route in describing their last meal, punting and saying something like “something that my mom made” or “anything with chocolate.” I did do another Textalyser analysis on the top word pairs that started with “something” and “anything,” hoping that I’d find a significant number of users rallying around one ingredient, like “something with bananas,” but this wasn’t the case. We even danced around the idea of going really conceptual and designing a dish that was called, “something” and “anything,” but God did that sound horribly pretentious. With that, we struck those two words from the top 20 and promoted everything else up accordingly.

A lot of items on list were expected, but we were initially a bit surprised by the top 3. Chocolate, ok. But cheese and chicken? Why was bacon so low? What about Foie Gras? No love for Truffles? You hear foodies drone on and on about how bacon is essentially a protein based orgasm, how Foie is the Internationally Recognized Symbol for “oh, no you didn’t,” and how Truffles taste like Unicorn Tears. You’d think these would rank higher, like, in the top 3? Nope.

To be more specific, cheese showed up mostly as “Mac and Cheese,” and simply “Bread and Cheese.” Chicken was most frequently associated with a rustic ethnic dish that matched that person’s last name (Food: Fatteh, Surname: Abdalla), a simple thing that their mother used to make (Mom’s Meatloaf), or both (Food: “Mom’s chicken stew, her thosai with her coconut chutney the way her mom used to make it”, Surname: Teoh). They were all basic, homey dishes that you could envision any mother or father making for these people when they were small.

We were learning that people wanted simple comfort for their last meal, not unchecked lavishness. A small faction of people wanted to have things that they had never eaten before, which was puzzling, but for the most part it was the old favorites that made it to the list. This was reassuring on some level, proof that this supperclub experiment of ours didn’t just attract wanton thrill seekers and name dropping, self-conscious socialites. Real people with authentic connections to food, no douche-bags.

From the top 20, we started writing recipes and testing dishes. In the next post, I’ll talk about what we made.

All photos by David Christiansen.