4 ms·

So I know next to nothing about this, and I've never shipped a product of my own, but why can't you just do something like: 1. Price it at price $X, record num

So I know next to nothing about this, and I've never shipped a product of my own, but why can't you just do something like: 1. Price it at price $X, record number of buys/month over a few months. 2. Price it at price $Y, record number of buys/month over a few months. 3. Now with the assumption that you have a linear demand curve, you can extrapolate the buys at any price, which would roughly follow the equation B = (X-Y)/(P1-P2) x P, where B is the number of buys and P is the price in dollars. Obviously this equation will only work (if at all) for values in a certain range. 4. You want to maximize B x P, which is the total money made per month. That is, assuming costs are always the same regardless of the number of sales.

SiVal 14y ago
A good question with an unfortunate answer: the demand curve is a curve, not a line, and it's a curve that changes shape when measured. If you test at $100 and then at $50, what you're testing isn't the demand at $50, it's the demand for a $100 product that has been discounted for some reason to $50. Some will wonder why it apparently wasn't worth $100. They'll be less likely to buy, even for $50. Others will wonder if the $50 is temporary and jump on the deal. Which group will be bigger? How would you know? If you restore the price to $100, you'll be testing the demand for a product that is now better known than when it was originally $100, so it should sell more than before, but many of those new people will think of it as a $50 product with a jacked-up price. They'll regret not having bought it when it was $50 and not want to feel like suckers paying $100 now. So your demand at $100 won't equal your previous demand at $100, and you won't be sure what the "real" demand is, even at the numbers you've tested, much less at numbers you haven't. Even if a genie granted your wish and gave you today's demand curve, it would soon change shape the way prices in the stock market move around. If you could predict the future shape of your demand curve, you could just go into fulltime investing in the stock market and make a killing. I hope this helps make it clear why you can't make it clear.
roel_v 14y ago
The two samples aren't independent, and you can't really isolate other effects (e.g., maybe people buy more in summer for the product at hand?) Secondly, demand curves in practise aren't that fluid. Plus you want to capture consumer surplus. And and and... Spolsky wrote a post on it once, it has a good and accessible introduction to the concepts, it's called 'Camels and rubber duckies'.
wikwocket 14y ago
Demand curves sadly are notoriously nonlinear: -If priced low, more will buy... unless it's TOO low, then the product is junk -If price high, fewer will buy... unless it is VERY high, then the product must be a luxury item, hand-crafted out of artisan materials! And demand will vary with seasons, holidays, work schedules, school schedules, the stock market, the phase of the moon, the price of tea in China, etc. Fortunately you can get around time-based fluctuations with A/B testing. There are theories about how to learn the demand curve (see Van Westendorp's price sensitivity meter), but they are theories only.

SiVal 14y ago

A good question with an unfortunate answer: the demand curve is a curve, not a line, and it's a curve that changes shape when measured. If you test at $100 and then at $50, what you're testing isn't the demand at $50, it's the demand for a $100 product that has been discounted for some reason to $50. Some will wonder why it apparently wasn't worth $100. They'll be less likely to buy, even for $50. Others will wonder if the $50 is temporary and jump on the deal. Which group will be bigger? How would you know? If you restore the price to $100, you'll be testing the demand for a product that is now better known than when it was originally $100, so it should sell more than before, but many of those new people will think of it as a $50 product with a jacked-up price. They'll regret not having bought it when it was $50 and not want to feel like suckers paying $100 now. So your demand at $100 won't equal your previous demand at $100, and you won't be sure what the "real" demand is, even at the numbers you've tested, much less at numbers you haven't. Even if a genie granted your wish and gave you today's demand curve, it would soon change shape the way prices in the stock market move around. If you could predict the future shape of your demand curve, you could just go into fulltime investing in the stock market and make a killing. I hope this helps make it clear why you can't make it clear.

roel_v 14y ago

The two samples aren't independent, and you can't really isolate other effects (e.g., maybe people buy more in summer for the product at hand?) Secondly, demand curves in practise aren't that fluid. Plus you want to capture consumer surplus. And and and... Spolsky wrote a post on it once, it has a good and accessible introduction to the concepts, it's called 'Camels and rubber duckies'.

wikwocket 14y ago

Demand curves sadly are notoriously nonlinear: -If priced low, more will buy... unless it's TOO low, then the product is junk -If price high, fewer will buy... unless it is VERY high, then the product must be a luxury item, hand-crafted out of artisan materials! And demand will vary with seasons, holidays, work schedules, school schedules, the stock market, the phase of the moon, the price of tea in China, etc. Fortunately you can get around time-based fluctuations with A/B testing. There are theories about how to learn the demand curve (see Van Westendorp's price sensitivity meter), but they are theories only.