Colin Barry

Triangulate (LTV, Monday, Week 2)


It can be really hard to test some hypotheses around demand in B2B startups
-- Will eHarmony or buy an algorithm? They'll always SAY they're interested...
-- Explicitly list out hypotheses; figure out a test to validate or reject. Good ideas\

Can we either fake or mechanical-turk this complex feature to figure out whether people will buy it?
Can we gauge interest by making people do something costly (hand over their e-mail address)?

It is really hard to develop an MVP for products with strong network effects (i.e. dating websites, two-sided markets, etc.).

Given that your ability to run experiments drives the pace of your development, it's hard to move fast without a bunch of customers. How many A/B tests can you do with 100 users?

You can hide crappy users to boost perceived user quality (hi Quora).

You get one round of good PR when you raise money. Use it wisely.

Sunil Nagaraj on Facebook platform:
-- By the time you hear about a new feature on Facebook and you think you're early in exploiting it, you are already too late.
-- The only people who seem to win on Facebook apps are those who read Terms of Service incredibly closely --- and then do exploit them by toeing the line of what is permissible.