While price is simply a component of the offer, it has a powerful influence on purchase conversions. Because of this, it’s important to test your price separate from your other offer elements.
The advantage of price testing on the Web is that you have several desirable conditions:
- Large numbers of new prospects who have not been exposed to your company before
- Ability to easily modify pricing displayed to a particular visitor
- Ability to hold all other factors constant
- End-to-end tracking and immediate recording of the sale
However, a host of possible issues make price testing potentially problematic even on the Web:
- Changing supply-demand imbalances
- Product obsolescence
- Degree of commoditization
- Reactions of your competitors
Because of this shifting environment, it’s important to regularly retest your price to detect any significant changes.
True price testing is still something of a “Wild West” situation in practice. Hopefully the following summary will give you a framework for how to approach it.
Price Testing Basics
Unlike other page elements that involve distinct choices (i.e., testing a green button against a blue button), price is a continuous variable. Continuous variables can theoretically be set to an infinite number of different values. Price can be varied over a wide range in one-cent increments. There is no way to properly guess the exact value that will give you the highest profits.
A typical price/profit curve for a single product looks like an inverted “U.” At the lower end, your profits will be zero because the price of the product equals your cost to produce and deliver it. At this point, it doesn’t matter how many people buy – there is no profit to be made. At the higher end, your profits will also approach zero because the price will be too high, and you will not have any customers.
Setting the wrong price can have disastrous consequences for the success of your product or service. Your goal should be to set the price at (or near) the profit sweet spot at the top of the curve. Following are several common approaches to finding the right price.
Spot testing treats price as a discrete variable. The advantage to this approach is that it works with your existing (discrete variable) landing page optimization tools. However, by testing price as a discrete variable (e.g., you test three distinct prices – your current price, a specific lower price, and a specific higher price), you’re only getting information about the exact prices you choose to test. You’ll know which one of the tested prices is better, but you won’t know if any of them are at the exact profit optimum point. In other words, you may very well be leaving money on the table.
If your only alternative is not to test price at all, then you should use spot testing – a little bit of something is better than a whole lot of nothing.
Walking the Price Curve
Many companies conduct informal price testing by “walking” their price/profit curve line. They change the price and measure the resulting performance. If it improves, they incrementally change the price again in the same direction (either raising it more or lowering it more depending on the circumstances). Eventually they will overshoot the top of the curve and experience a decline in profits. At that point, they back up to the previous price and lock that in as their winner.
This approach has significant drawbacks. First, it can be very time-consuming. Depending on the size of the price change increment that you choose, you may have to run several back-to-back tests. The lost opportunity cost of being at suboptimal pricing for the length of these tests can be significant. Second, there is no clear way to decide on how to calculate your price change increment. Some companies use a fixed amount, others use a percentage of the current price. Regardless of the approach, if you choose incorrectly, you will either require many tests (as mentioned above), or not find the top of the profit curve because your increment is too large. Third, pricing changes are done sequentially. Once a change is made, everyone sees the new price.
For all of these reasons, sequential sampling should be used as a last resort. You don’t know what other outside factors have impacted price across all of your tests (seasonality, traffic changes, or external events such as competitor price changes or company announcements).
Price Elasticity Modeling
It’s possible to build a model of the predicted sales conversion rate as a function of price. Such price elasticity models are constructed using a variety of mathematical approaches, and include different assumptions at their core. But the basic idea is the same. If you can predict what percentage of people will buy your product at a given price, and you know your costs at any price point, you should be able to calculate your profit per visitor for all prices.
It’s possible that your current price is already close to the profit optimum, especially if the top of the profit curve is pretty flat across a wide range of prices. Since the model also predicts the conversion rate at a given price, you can consciously make the trade-off between higher market share and perceived exclusivity in such cases. In other words, you can choose a lower price and more customers, or a higher price and fewer customers while still maintaining a near-optimal profitability per visitor.
Price elasticity modeling also works for a product with a single upsell option. Within this configuration, you need to determine the revenue-per-visitor optimal pricing for both the base product and the upsell. The upsell can be displayed in parallel (shown side-by-side on the same page) or serially (shown on a subsequent page once someone has decided to buy the base product).