Does price optimization work for today's businesses?
Price optimization means improving the performance of the price aspect of your marketing mix (also known as the 4P's, product, price, place and promotion). Price optimization leverages data to drive more revenue or margin by adjusting the price or prices customers see and pay.
While price optimization can be done as a one-time or occasional project, we define price optimization be a continuing process, taking in new data and updating prices dynamically (though perhaps not instantly). This is because, in today's volatile markets, if price can be dynamically optimized–someone will optimize it, and any business failing to do so will eventually lose. (See: Uber v. Taxis).
Price optimization done right results in more closely matching supply and demand, enabling a company to capture more margin while also serving more customers and increasing customer happiness. Executed badly, it can result in upset customers, reduce revenue, margin, or all three.
What does price optimization look like?
Price optimization takes different forms. In all forms, a company uses forecasted demand data (usually based on historical sales behavior) to predict demand, and then matches prices to conform to expected demand. Note that in some industries, such as travel, price optimization is called "revenue management".
Common examples of price optimization include:
- Airline fares varying by time of day, type of trip, and origin-destination pair.
- Rental car rates varying by location, type of car, and time of day or week.
- Uber "surge pricing" where prices increase during times of higher demand in a zone, for example, by AT&T Park when a baseball game ends, or during evening rush hour.
- Supermarkets promoting items with different discounts to different customers based on historical purchase behavior, expected cart size, and more.
- Banks offering different mortgage, CD or savings rates based on the branch a customer visited.
- Pharmacy adjusting the margin on certain items, choosing some "loss leaders" to induce visits and others with much higher margin to increase profit.
- Apparel retailers marking down inventory as the season progresses.
- Online retailers matching the price on competitive items like paper towels while increasing margin on items only they carry.
- Hotels charging different amounts by day of week, time of booking, and what events are scheduled nearby.
- Electricity prices changing by time of day, day of week, or demand on the electrical grid.
- Basketball tickets cost more for certain seats, and certain games, depending on how the team is doing, opponent for a particular game, and which players are playing in that game.
What is required to implement price optimization?
There are several well studied areas where price optimization makes sense. Though academics believe capacity or supply constraints, and expiring inventory are required, our research shows that these are less important. For example, price optimization is quite possible when selling software, which is neither supply constrained nor expiring; banks typically have few constraints on the CDs they sell or mortgages they issue, but also actively optimize prices.
This leaves just three crucial requirements for price optimization:
- Different willingness to pay amongst customers
- Data upon which to base a reliable demand forecast
- The ability to segment prices based on willingness to pay
Let's look at each in detail.
Different willingness to pay
It would be very difficult to charge $1,000 for a first class seat and $100 for a coach seat if every single customer were only willing to spend $100 on the flight. However, becuase certain customers are willing to spend more–up to, and perhaps exceeding $1,000–we have multiple fare classes and optimized pricing for airlines.
Data to base a demand forecast
With no prior data, it is impossible to create dynamic or optimized pricing. Mathematically, without any evidence, if you arbitrarily choose a price–even if well researched–there is a very near 100% chance that you have not chosen the optimal price. How could you have, with no data?
For airlines, retailers, hotels, and any industry, a history of sales (and, preferably, behavior data) is required in order to estimate future demand. This demand estimate or forecast is a requirement for price optimization.
The ability to segment prices
Just because one can find different segments with different willingness to pay does not mean you can charge different prices to each segment.
The regulated taxi industry knew demand spiked at rush hour, with people leaving work willing to pay more to get home dry. But regulations prohibited increased prices.
In retail, Best Buy has tried to charge higher prices in-store where customers typically have higher willingness to pay (they drove to a big box store there, after all). However as most customers look up prices online on their mobile phone while in the store, knowing that willingness to pay is higher in store or even at particular stores becomes impossible to act upon. (That said, retailers need to stay price competitive now more than ever, so dynamic pricing remains important).
Even unregulated industries exhibit a stubbornness when it comes to segmentation. Concert tickets have proven historically difficult to optimize due to customer mistrust and supplier ignorance–even in spite of an active secondary market, with well documented varations in willingness to pay.
Price optimization is the science of matching the right price to the right customer. It is part of the larger marketing mix science, and can be implemented either separately or in conjunction with changes to other P's–but does not necessarily need to involve other changes.
This article is a primer on the subject. We also wrote a book, which you can get here, if you want more detail, including a list of common strategies and common pitfalls. Robert Phillips wrote the seminal textbook, which you can find Amazon here, or drop us a note with your address and we'll send you a copy.