Colin Barry

Aardvark (LTV, Wednesday, Week 1)


Was Aardvark a "lean startup?"
Yes: hypothesis-driven, frequent smoke tests, high frequency of release, lots of prototyping (Mechanical Turk)
No: prioritized user acquisition, NOT hypothesis-driven (vision/persistence vs. arrogance/refusal to take feedback?)

Is Aardvark really viral? NO!
-- Not a game.
-- Not clear social value for users to share (no ego gain, like Quora or Klout)
-- No inherent incentives to share.
-- No social extensions/integration with social platforms.

Perennial lean startup challenge:
-- Founders pick a business model they like.
-- Hypothesis testing: Hard to differentiate between legitimate affirmation and false positives.
-- Hard to keep testing core "big assumptions"

"Talent hits a target no one else can hit; genius hits a target no one else can see." -- Schopenhauer (via Jeff Bussgang)

Jon Chim on Aardvark:
-- "Startups live hand to mouth." => Can't possibly test everything; stage investment accordingly
-- Aardvark spent 30-40% of engineering effort on enabling A/B testing and metric collection => You need to design everything around making tests possible

-- Testing: "We would get four users every week off the street (via Craigslist post). We had users before we knew what we were even testing."

-- Hired engineers who didn't know how to code. => It causes trouble when you look at a problem and start thinking, "I know how to solve this..." and then immediately start building.

-- It's hard to separate timing/implementation from the idea in hypothesis-testing. Especially in network-effects businesses...
Did your test fail because the idea was bad, or because you didn't market well enough/have enough users for the product to be valuable when you ran the test?