When an inventor/businessman has translated an idea into a product. One of the most valuable questions is: “What should my price be?”.
There are many ways to answer this question:
- Price competitively relative to competition.
- Maximize growth with zero margin.
- Set a random margin.
- Price / Test / RePrice and repeat.
- Email/Coupon based polling.
Each of these methods has flaws in that the price is never set to methodically maximize profit given all the available information. By definition, only the optimal price can do this.
I will discuss each of these methods in future posts.
The optimal price is a function of only two variables:
- Price-Sensitivity: The rate at which customers leave as price is raised. This is elaborated below.
- Marginal Cost: The cost paid by the firm to deliver the additional item to the customer. This can usually be calculated by carefully tabulating costs. This should typically include:
- Cost of Goods sold
- shipping and handling: Shipping should include the costs of raw materials such as packaging, postage, and the cost of time to package goods.
- Taxes: This must include the sales tax and any tariffs.
Notice that the fixed costs don’t enter the calculation of the optimal price because it doesn’t change with price or quantity.
The figure shows a case of underpricing. When price is raised, the green shaded region of profit gained from price increase is larger for the existing customer base than the blue shaded region of profit sacrificed per lost customer.
The converse is true in a case of overpricing. When price is lowered, a larger gain is realized from increasing customer base than the reduction in profit from the existing customer base.
At the optimal price, the profit sacrificed matches the profit gained for small price and volume changes.
What is the fastest, cheapest and easiest way to estimate the optimal price?
The fastest and easiest way involves automation to stay continuously at the optimal price. An algorithmic approach to maximize profit is called for. The algorithm would follow these steps:
- Gather information: about customer price sensitivity through controlled experiments around an initial guess about the price to reduce the impact of noise. These experiments would have to be designed to minimize the loss resulting from pricing sub-optimally given all the available information.
- Build a model that forecasts the optimal price. The most basic model only considers price as a factor in consumer choice. This is also the model that can be built with the least amount of data.
- Update the guess about the optimal price and start over with step 1.
A 1% error in pricing causes a magnified effect on profit because revenue is several times operating profit. A 1% increase in price without any reduction in demand can increase profit for Sprouts (a grocery chain) by 14% and for Amazon by 500%.
The scientific method of pricing can be used in e-commerce settings. The potential exists in a wide range of markets. A few e-commerce examples include online grocers, device manufacturers and online booking of golf tee-times.
Visualize the range of prices a business can charge.
If profit is different at all prices, there is only one price that maximizes profit.
We call this price the optimal price. Pricing differently from the optimal price implies a loss.
At prices below the optimal price, the profit per customer drops too fast to increase total profit when prices are lowered. Above the optimal price, customers leave faster than profit rises for the existing customer base as the price is increased. At the optimal price, the change in volume and the change in profit from existing customers offset each other exactly for small price changes.
For this basic model, the only variable that needs to be forecast is the rate at which consumer demand changes with price. I call this the price-sensitivity of the market.
Want to know how big the loss can be under different conditions?
Lets call the price that maximizes profit the optimal price. Any business that prices differently suffers an ignorance loss.
When I started talking to businesses with an e-commerce component, I realized that pricing takes the back seat because of a lack of appropriate tools.
I developed a calculator to allow people to calculate what their loss relative to optimal pricing may be.
I realized that people are paying a huge ignorance loss.
The challenge businesses face is to process customer purchasing behaviour to extract information about customer’s willingness to pay.