How Elan Lee built Exploding Kittens into a 100-game empire

Executive overview

Most creative teams waste time on open-ended brainstorming that produces nothing. Elan Lee, CEO of Exploding Kittens, runs his design process on constraints, rapid rejection, and a single-question playtesting filter.

Games should make players entertaining — not the other way around. That principle drives every design decision, from the core gameplay loop to the one-liner that sales must nail before a game ships.

The fastest path to great ideas is killing bad ones without ego.

Kickstarter launch strategy

• Built audience before launch by leveraging co-founder Matt Inman's decade-long following on The Oatmeal
• Used Kickstarter's preview page to share early access with influential people
• Deliberately inserted errors (typos, broken images) so reviewers had something to "catch" — turning passive observers into invested advocates
• Result: 219,000 backers, still the most-backed Kickstarter campaign in history

The super fan formula

• Games should make players entertaining, not entertain players passively
• First game: players discover relationships and social dynamics through the game's toolset
• Subsequent games: players develop personal strategy and a sense of mastery
• Mastery feeling — self-discovered, not handed to them — drives word-of-mouth

Core gameplay loops

• A core gameplay loop is a simple interaction repeated from start to finish, escalating in tension with each cycle
• Exploding Kittens draws from Russian roulette: one "kill" card hidden in a shrinking deck, odds worsening each turn
• Every other card in the deck modifies the loop (peek, deflect, steal) — enhancements, not distractions
• New gameplay loops are still being discovered; the medium is young enough that no definitive taxonomy exists

Creative constraints in brainstorming

• Open-ended "come up with a game" sessions produce nothing after three days
• Constraint method: buy 20 random objects from three different stores, combine any two, design a game around them
• Key audience constraints: known demographic, target game length (20-minute sweet spot), one-sentence describability
• If a game can't be pitched in a single memorable sentence, it doesn't get developed

Quarterly design retreats

• Three days, same physical location, no phones, no outside obligations
• Everyone sleeps and eats together to sustain creative momentum
• Process: generate and kill ideas rapidly — thousands per day, not two or three
• Improv's "yes, and" rule is explicitly rejected; "that sucks, next" is the operating mode
• Ideas need a champion to survive: one person arguing for more time is enough to keep it alive; silence kills it
• Decision-making is collective — no single authority in the room

Protecting psychological safety while rejecting ideas fast

• Leader goes first: Elan publicly accepts immediate destruction of his own ideas
• No ego modelled visibly in round one; by day three new designers trust the process
• Pride shifts from individual ideas to the five or six strong concepts that survive

Sales and marketing as final filter

• On day three, marketing and sales join the retreat to hear pitches
• Their test: can they write a one-liner? Can they visualise 15 seconds of compelling social content?
• A great game with no sellable hook is abandoned without sentiment
• Surprising reversals happen: a "math game" the designers doubted became a top prospect after marketing stripped out 90% of the math
• Half-baked ideas are shown deliberately; marketing sometimes sees potential the designers missed

Knowing when a game is done

• 400 volunteer families in the "Kitty Test Pilots" program receive pre-release games and return video recordings of play sessions
• Old approach: 50-question feedback forms — useless
• Current approach: one question — "Do you want to play again?"
• Release threshold: 99% of responses are "yes"

Game naming

• Names must be evocative, ownable, and legally clear — most candidates fail on at least one criterion
• Process: designers generate names → art team discards and regenerates with imagery in mind → Matt Inman often delivers the final answer
• Let's Hit Each Other With Fake Swords: long, cumbersome, fits barely on the box — but you instantly know how to play
• About 20% of released names are acknowledged mistakes; the rule is to make new mistakes, not repeat old ones