The Modern Bar Cart Podcast: Episode 310

"Data-Driven Drinks with Dr. Kevin Peterson"
Original Airdate: March 5, 2026
Host: Eric Kozlik
Guest: Dr. Kevin Peterson

Episode Overview

In this engaging and revealing conversation, host Eric Kozlik reunites with Dr. Kevin Peterson—returning guest, acclaimed technical mixologist, and author of "Cocktail Theory"—to explore Kevin’s latest project: mining and analyzing years of customer cocktail rating data from his innovative Detroit bar, Castalia. Their discussion centers on how data can help decode the complexity of drink preferences and improve the guest experience, moving beyond the optimization of iconic cocktails to the thorny challenge of drinker-cocktail matchmaking.

Featured Cocktail: The Perfect Spicy Margarita (from Cocktail Theory)

• Recipe Overview (01:00 - 02:00)

  • 2 oz blanco tequila
  • 3/4 oz curaçao (Pierre Ferrand recommended)
  • 1/2 oz fresh lime juice
  • 1 ml jalapeño tincture
  • 1 ml serrano tincture
  • 1/2 ml habanero tincture

• Preparation Tips

  • Shake all ingredients with 4–5 ice cubes for 12 seconds.
  • Strain into a snifter glass, garnish with dehydrated lime.
  • Dr. Peterson’s approach: Layering three different chili tinctures for a time-evolving, spatially dynamic heat.

Notable Quote:
“Flavors are not isolated single point sensations. They have both spatial locations...and time evolutions. I used three tinctures to create an evolving time and space sensation of heat, as if a wave were traveling through your mouth.”
—Dr. Kevin Peterson [02:00-03:00, quoting "Cocktail Theory"]

Key Discussion Points & Insights

1. Reflecting on Castalia and Sfumato’s Legacy

[05:39-08:15]

• Dr. Peterson recounts his journey from engineer to bar owner, explaining Castalia’s unique model as a perfumery by day and a cocktail bar by night.
• The emotional process of winding down after nearly a decade: “Bars are where a lot of important crossroads in life are decided...doing justice to everything that happened there became kind of a weighty proposition.”
  —Dr. Peterson [06:48]

2. The Shift to Data-Driven Mixology

[08:15-09:33]

• Kevin began collecting detailed guest feedback and drink ratings, amassing “a massive stash of data.”
• He chose analog (paper) data collection to preserve the screenless, immersive bar experience:
  “A lot of information gets conveyed in the weird little drawings people make...You can sort of see the handwriting degrade from the first drink to the last drink.”
  —Dr. Peterson [08:29]

3. From Cocktail Theory to Theories of Taste

[10:41-14:48]

• "Cocktail Theory" focused on optimizing classic drinks by systematically varying variables like ratios, temperature, and dilution.
• Key insight: Making a perfect version of a cocktail doesn't make it the perfect drink for every person.
  • People’s palates and preferences vary wildly; thus, the current challenge is matchmaking guests with their “optimal” drink—now a multidimensional, subjective problem.

Notable Quote:
“Even a perfectly made gin and tonic isn’t the perfect drink for somebody that doesn’t like gin and tonics.”
—Dr. Kevin Peterson [12:04]

4. The Complexity of Palates and Preferences

[14:48-19:34]

• There are archetypes (bitter/boozy, citrusy/light, etc.), but real-world data reveals high individual variability.

• Collecting data on guest responses to individual tastes/textures did not yield robust predictions; overall drink impression depends on nonlinear interactions between components.

• “My initial approach was...to break a cocktail down to its components. But...it's crazy. You have to look at everything at the same time. You can't look at them one by one.”
  —Dr. Kevin Peterson [15:51]

• Even small ingredients (e.g., absinthe, in a Sazerac) can make or break a drink for someone.

• Bartending is already a process of practical problem-solving; Kevin now seeks to formalize, measure, and potentially improve it.

5. Toward Practical Solutions: Better Menu Writing & Dialing In Guest Preferences

[19:34-22:16]

• Practical insights include using more descriptive menu language, especially for highly polarizing traits like bitterness and spiciness.

• Bitterness and spiciness are particularly tricky—no level will please everyone.

  • Spicy drinks: “So many times as I'm reading through the comments...Way too spicy. The spice is right on. Couldn't taste the spice. I wish there was a little more spice. And it's like, guys, you got literally the same liquid. Like, like, this is not a me problem.”
    —Dr. Kevin Peterson [23:45]

• Suggestion: Enable guests to customize their experience, e.g., with spiced rims or modular garnishes as in a Batanga.

  • “It's not about like the chef, how the chef likes the steak, it's the customer.”
    —Dr. Peterson [25:50]

6. The Notion of Perceptual Thresholds

[25:55-29:28]

• Guests experience flavor on non-linear, often asymmetric hedonic curves. Example: Some crave extreme spice, some are turned off by it.
• Sensory “thresholds” apply to both flavor and pleasure; exceeding them can make a drink instantly unenjoyable.

Notable Quote:
“You cross that threshold and now all of a sudden, it sucks...It’s easy to think of spiciness in those terms where...you cross that line where you're like, okay, this hurts. Not into this anymore.”
—Dr. Kevin Peterson [28:18]

• Subjective flavor descriptions complicate data: Guests ascribe perceived notes that may not be present, underscoring the complexity of aroma and taste memory.

7. Average vs. Transcendent: What Should Bartenders Optimize For?

[29:28-34:35]

• The tension between “most liked by most people” and “transcendent, memorable experiences.”

• Eric recalls his first taste of green Chartreuse as a “transcendent moment...like seeing a new color.”
  —Eric Kozlik [30:36]

• Guest surveys rarely yield technical descriptions; most guests describe desired feelings, not ingredients or specs.

  • “Less than 5% of people actually described a cocktail. They described the feelings that they would have from drinking that cocktail.”
    —Dr. Kevin Peterson [32:57]

8. The Future: Bartender Choice, Data, and Guest Profiling

[34:35-37:42]

• Could guest profiles, generated by algorithms or pre-service surveys, help bartenders deliver more tailored experiences?
• There’s already a “mental algorithm” at play in bartender’s choice programs.
• Formalizing guest preference data in advance could support deeper personalization and efficiency—especially for unique or time-intensive requests.

Notable Moment:
Eric and Kevin discuss logistical challenges for “bartender’s choice” service, and dream of a future where more guest data allows advance tweaks, ordering, or prep for niche requests.

[Quote]
“A lot of bartenders have an algorithm in their head, whether they call it that or not...This is just a more formalized way of doing this.”
—Dr. Kevin Peterson [35:26]

9. Archetypes and Personality

[37:42-39:52]

• Eric notes D.C.'s Copycat and Astoria bars as examples of menu structures balancing set offerings with bartender's choice.
• Further food for thought: Can finding cocktail archetypes and matching them to human personality indices (like the Big 5) yield new insights?

Notable Quotes and Moments

• On the magic of the bar experience:
  “Bars are where a lot of important crossroads in life are decided and decisions get made and celebrations happen.”
  —Dr. Kevin Peterson [06:48]

• On palate complexity:
  “Nobody's quite got the same combination of preferences...My initial approach was going to be to break it down...but it's crazy. You have to look at everything at the same time.”
  —Dr. Kevin Peterson [14:48, 15:51]

• On the subjectivity of flavor:
  “For the same drink, people will say, ‘Oh, I love the spruce note’...‘The lychee was amazing’...‘The pineapple note was incredible’...Well, half of these things aren’t even in the drink, but that's the impression you got.”
  —Dr. Kevin Peterson [28:39]

• On guest experience vs. technical perfection:
  “It's not about...how the chef likes the steak, it's the customer.”
  —Dr. Kevin Peterson [25:50]

• On drink "archetypes" and optimizing for joy:
  “Every once in a while we just absolutely hit the nail on the head.”
  —Dr. Kevin Peterson [19:12]

Timestamps for Major Segments

• Intro & Featured Cocktail: 00:21–05:11
• Kevin Introduces His New Project: 05:11–08:15
• Analog vs. Digital Data Collection: 08:15–09:33
• From "Cocktail Theory" to Taste Theory: 09:33–14:48
• Palate Diversity and Data Surprises: 14:48–19:34
• Practical Tips: Menu Writing & Spice: 19:34–25:49
• Thresholds & Subjective Experience: 25:55–29:28
• Average vs. Transcendent: What to Optimize: 29:28–34:35
• Data-Driven Bartending/Guest Profiling: 34:35–39:52
• Closing Remarks: 39:52–41:28

Summary Takeaways

• Optimizing for individuals is far more complex and rewarding than optimizing drinks in the abstract.
• Analog, pen-and-paper data collection at the bar uncovers subtle, subjective details often lost in digital surveys.
• Explicit menu language and ways of allowing guests to customize are practical ways to honor individual thresholds and maximize joy.
• Algorithms and guest profiles could, if thoughtfully implemented, help bars deliver truly bespoke service without compromising ambiance.
• Ultimately, the feeling, context, and guest experience outweigh ingredient lists and technical “perfection.”

This episode blends deep nerdiness and emotional insight, setting the stage for part two’s deeper dive (not included here). Listeners are left with a sense of the enormous potential—and challenge—in marrying data analysis with hospitality artistry.

For more episodes and community discussion, join the Modern Bar Cart Discord or connect with Eric via email.