Choice, happiness and spaghetti sauce | Malcolm Gladwell

Business/strategy summary (what Howard Moskowitz changed in food product development)

Malcolm Gladwell argues that Howard Moskowitz (a psychophysicist and consumer research innovator) “reinvented” how large food companies decide which products will satisfy people by replacing universal/average-focused research with segmented, data-driven design—one that accepts variability in customer preferences.

Core mindset shifts (business playbook)

• Stop searching for a single “perfect” product

  • Instead of optimizing to one average sweet spot (e.g., “perfect Pepsi”), Moskowitz reframed the goal as finding the best options for different customer segments (“perfect Pepsis”).

• Don’t rely on direct questioning / focus groups to reveal true desires

  • Consumers often can’t articulate what they actually like (they may not say “milky, weak coffee” even if they prefer it).

• Use horizontal segmentation

  • Preferences aren’t arranged in a quality “hierarchy” (e.g., “better mustard”)—they vary across groups with different tastes.

• Replace “Platonic dish” universals with option portfolios

  • Instead of one “correct” way a dish should be, companies should build multiple variants that serve different preference clusters.

Framework/process used (implicit “segmentation + clustering” method)

• Design an experiment over many product dimensions

  • Vary attributes across a wide design space (e.g., sweetness, garlic level, tartness, visible solids).

• Collect large-scale preference ratings

  • Test many variants with consumers and have them rate each.

• Analyze with clustering to find natural segments

  • Cluster responses into distinct groups (e.g., plain / spicy / extra-chunky).

• Develop a product line to win each segment

  • Reformulate and launch targeted variants, then expand the category into more choices.

---

Concrete examples & outcomes (execution details + results)

Pepsi / Diet Pepsi (aspartame dosing)

• Initial problem: Pepsi asked how much aspartame to add to hit a “sweet spot” between 8% and 12% sweetness.
• What went wrong: They expected a clean “best concentration” curve, but results came back messy.
• Key insight: They were asking the wrong question:
  • Wrong: “perfect Pepsi”
  • Right: “perfect Pepsis” (different people have different preference profiles)

(No sales numbers were provided for Pepsi in the subtitles.)

Vlasic Pickles (creating a new “zesty” direction)

• Vlasic asked for the “perfect pickle.”
• Moskowitz: “There is no perfect pickle; there are only perfect pickles.”
• Action: Instead of just improving the standard, they needed to build “zesty.”

(No specific numeric metrics were provided for this case.)

Campbell’s / Prego Spaghetti Sauce (category leadership via segmentation)

• Market context: In the early 1980s, Prego struggled against Ragù, the dominant sauce.

• Howard’s approach:

  • Created 45 varieties by varying many dimensions, including:
    • sweetness
    • garlic level
    • “tomatoey-ness”
    • tartness / sourness
    • visible solids (with “extra-chunky” as a major differentiator)

• Consumer testing process:

  • Ran large testing sessions in multiple cities.
  • Served 10 bowls per session (each session included different sauces).
  • Participants rated goodness on 0–100.
  • Repeated across months to build a large data set.

• Segmentation result (3 clusters):

  • plain
  • spicy
  • extra chunky

• Key commercialization insight:

  • The extra-chunky group was the biggest and was underserved in supermarkets at the time.

• Business outcome:

  • Prego launched an extra-chunky line that “immediately and completely took over” the business.
  • Over 10 years, the extra-chunky line made $600 million.

• Industry ripple effect:

  • Competitors expanded into many SKUs (later expanding diversity in items like vinegars/mustards/olive oils).
  • Ragù later hired Moskowitz, producing 36 varieties across 6 named types (as described in the talk).

---

Implied KPIs / targets mentioned

• Diet Pepsi sweet spot range: 8%–12% aspartame sweetness (a predefined “target window”).
• Preference measurement scale: ratings from 0 to 100.
• Growth/result figure: $600M over 10 years for Prego’s extra-chunky line.
• Market/customer distribution estimate:
  • “about one third” of people craved extra-chunky.
  • Also referenced coffee preference shares: only ~25–27% like the “dark, rich, hearty roast” people say they want.

---

Actionable recommendations (how to apply the lesson)

• If customer research seems noisy or inconsistent, suspect the question

  • Shift from “find the average best solution” to “find segment-specific best solutions.”

• Build a testing plan that spans product attributes broadly

  • Use variation across meaningful taste dimensions, not only small incremental changes.

• Cluster responses into natural groups, then design an option portfolio

  • Treat a category as serving multiple preference segments—not one universal “best.”

• Treat focus groups as incomplete signals

  • Prefer experiments + segmentation, especially when consumers can’t reliably verbalize preferences.

---

Presenter/source(s)

• Malcolm Gladwell (speaker)
• Howard Moskowitz (subject/source; psychophysicist and consultant)
• Case companies mentioned: Pepsi, Nescafé, Vlasic, Campbell’s / Prego, Ragù