Sep 13, 2024, 10:00 AM ETFireside Chat, Interview

Mike Hudack: How Facebook, Monzo and Deliveroo Build Great Products | E1201

The core art of product is understanding user goals and helping achieve them with minimal work.
Optimal product team size is 6–8 people, consisting primarily of engineers.
Ideal team composition includes:
- One data scientist (if possible).
- One designer.
- One product manager (optional).
Teams should be aligned with outcome-based goals (e.g., increasing user satisfaction or sales) rather than output-based "ship" goals.
Founders should adopt a "theory of the world" regarding user needs before writing code, though this theory is often wrong upon real-world testing.
Product leaders must distinguish between being 20% wrong (where a product can be iterated on) and 100% wrong (where the project should be killed).
Leaders should encourage admitting when a product is failing; investors respect teams that pivot or shut down early with remaining capital intact.

Page Insights Success:
- Rebuilt the confused, slow Page Insights product with a small team of 8–9 people.
- The rapid, high-quality rebuild shifted organizational culture, proving the Ads org could ship great software quickly.
- This established a new quality bar for the entire organization.
Audience Insights Failure:
- Built a "nice-to-have" product allowing marketers to explore population-level data.
- The project was technically difficult to build and ultimately deemed low-value.
- Key Lesson: "Nice-to-haves" often have underestimated costs and lower value returns; prioritize features that create meaningful value for users and shareholders.
Organizational Culture:
- Facebook was characterized by a flat organization where founders (like Mark Zuckerberg) participated in "founder mode," reviewing code diffs and product details regardless of seniority.
- The company prioritized employee time over minor costs, exemplified by vending machines providing free laptops and high-quality headphones.

Simplicity vs. Features:
- Product art lies in simplifying the interface to match user goals (e.g., helping advertisers sell products rather than managing complex targeting).
- While data guides science, product intuition and aesthetics ("art") remain crucial for psychological effects that cannot be immediately proven via metrics.
The "Long Road to the Start Line":
- Product teams should spend significant time developing a hypothesis before shipping V1.
- The goal is to bridge the gap between a 50% wrong theory and success rather than giving up early due to initial negative feedback.
Distribution Strategy:
- Distribution is critical; a great product without distribution is valueless.
- Viral growth is preferred over paid marketing if the product is truly delightful, but requires a tight Initial Customer Profile (ICP).
- Verticalization: Targeting specific communities (e.g., Filipino expats transferring money) yields higher conversion and virality than broad, generic targeting.
- Network Effects: Products like WhatsApp succeed by solving specific pain points (expensive international texting) that trigger organic, community-wide adoption.

Deliveroo vs. Facebook:
- Deliveroo operates in a real-time logistics environment (e.g., rain, rush hours) requiring constant demand modulation and urgency, unlike the slower iteration cycles of social media.
- Key Learning: The most critical product work was often in the backend delivery algorithms and dispatch systems, not the consumer-facing app.
- Lateness Model: A machine learning model was built to accurately estimate delivery times, shifting the strategy from promising unrealistic 15-minute windows to honest 45-minute estimates.
- Selection Strategy: A trip to Madrid revealed that users would tolerate soggy food from far distances if the restaurant was popular (e.g., Five Guys), disproving the assumption that delivery radius must strictly limit food quality.
Monzo (Banking Context):
- Regulation: Banking constraints require a shift from "move fast" to ensuring stability and trust, prohibiting tests on pricing or rapid global expansion without local adaptation.
- Monzo Flex: A lending product integrating directly into the bank account to make Monzo a user's primary bank.
- Strategy vs. Competitors: Chose to avoid crypto products to protect the brand's position as a trusted bank, prioritizing core banking features over "feature bloat."
- Global Expansion: Banking is highly regional; successful expansion requires local teams to adapt core product logic to local financial behaviors (e.g., salary cycles, mortgage systems).

Founder Mode:
- "Founder mode" (operating in details, skipping levels) is natural and beneficial in early-stage companies but can be detrimental in massive organizations if not adapted.
- Founders who behave badly will use "founder mode" as an excuse; the label itself does not cause bad behavior.
Handling Morale & Expectations:
- Leaders should set low expectations before launches to prevent team deflation if immediate goals aren't met.
- Reframing work to highlight its global impact (e.g., ads helping small businesses grow, Facebook sharing maintaining loose ties) boosts team purpose.
- Product teams need a "win" (like Page Insights) to reset morale and a clear reason to work hard.
Work-Life Integration:
- Family life can enhance product execution by providing a mental reset and fostering intellectual honesty.
- Distinguish between "grinding" (obsessive dashboard checking) and "focused work" (strategic decision-making); the former leads to burnout.

Hiring CPOs: Best done very late in a company's lifecycle; founders should remain product-minded as long as possible.
Promotion Strategy: Product Managers should focus on solving the company's most critical unmet needs rather than chasing titles.
Angel Investing: Invest in people capable of building from zero to one, not just experienced managers from large corporations.
Worst Product: A multi-tiered filtering encrypted proxy built as a teenager; technically impressive but not viable.
Monzo's Recent Success: Impressed by the investment product and family features (kid accounts) for expanding the demographic reach.
AI Outlook: Acknowledges current hype cycles for AI/LLMs but expects a trough before reaching a plateau of utility.
Product Promotion: Great PMs combine deep intuition, great design taste, engineering respect, and intellectual honesty about what they do not know.