7 comments

[ 0.19 ms ] story [ 24.6 ms ] thread
Typo on your middle card FYI

"Out ML algorithm will converge to a price that maximizes revenue, sales times price."

It also assumes the cost of providing service is zero....
Good catch. It's aimed at software companies at the moment although it could just take into account marginal cost in the future
Don't use anything like this - it's morally bankrupt.

If everyone who sold anything used tools like this to maximize revenue, we'd all be bled dry, even more than we are now.

Protect your customers, remember they're your top stakeholder!

Your goal should be to find a balance between meeting costs, providing affordable value to your customers, and generating a reasonable profit.

Otherwise, you're part of the disease.

WILDE.TWAIN.

I don't agree.

The price which maximizes revenue isn't always the highest price.

This API could equally suggest a lower price, bringing in more customers who get more utility of your service.

In a good market with lots of competition, the best strategy for the seller also leads to the best result for the buyer.

This sounds like the multi-armed bandit problem...

No need for fancy ML here - it's a thoroughly understood mathematical problem...

Unless of course you're using parameters of the user (browser, purchase history, interests, favourite pet, etc.)...?

You're right although multi-armed bandit problem doesn't address all the problems in the real world. E.x. adding different slot machines (prices) intelligently, or discounting past data without using magic parameters