Resources · Optimizing for Power Law
Optimizing for Power Law, Lifetime Value
Have you ever watched every number on the dashboard go green while the business fell short? This story walks through why, one scene at a time, with Bob, 100 fish and a net, when the fish only grow into their value later. The catch goes through five pools and is sold out of the last one, weeks after it was caught, so nothing at the dock says what a cast is worth. Counting fish, the dashboard overstates the business three times over. A signal the fish carries from the catch turns it around, and the last scene shows how to apply it.
- 1Meet Bob: after thousands of casts Bob knows the average fish weighs about 10 pounds, and the 5 biggest of 100 fish carry 59% of the weight
- 2Cast the net: one cast is luck, the average is steady, and that steady average is what Bob plans around
- 3Building a profitable business: the catch goes through five pools and 1 in 4 fish reach the last one, so Bob fishes a net four times the size; the same 25 fish, 250 pounds and $100 a cast, sold weeks later
- 4The investors: the same 36 cents on the dollar, with the money out for weeks before it comes back
- 5Add bait: 10 more fish a cast, 2.5 of them sold, $125 a cast once grown, and the plan: four times the fish sold, every fish still earning $4.55
- 6What the dashboard can't see: counting fish at any pool, the dashboard reports $1,707 a cast after 20 tests; once the fish grow, the catch earns $542, short of the $639 baseline
- 7The signal at the dock: priced by tail color at the catch, the Value/signal team makes $1,188 a cast against the Lower CAC team’s $542: 2.2 times the contribution on 73% of the money
- 8Applying Signal: the fish picks its signal before it’s caught and the tag is tracked to lifetime value; once signal and value line up, optimize on the signal
The story runs in your browser. Cast a few hundred more, drag any test, and watch the averages settle.
ABPRO_
See the three lines on your own book
ABPRO_ ties every lead to the revenue its account went on to produce, tiers the book, and tracks whether each change lands above baseline, on it or below it.
Book an ABPRO_ Demo