How do we get our customers to use our product the way it was intended?
When we take a product to market, we can make some guesses as to how our customers will consume it, but we’re always going to be wrong. This is fine. Our customers should be telling us what they want and we should be listening.
Well, it’s fine until they start using the product in a way we didn’t plan for. They might be gaming our system, they may be unintentionally loopholing us in a way that’s more convenient for them, or we may have just mis-predicted their behavior. In any case, when this departure from our norm starts to get expensive, we need to act before it becomes too late.
The problem is that there are all sorts of wrong ways to act, and they can be fatal to our company. Here are just a few that I’ve had personal experience with, and how we can use a peak pricing model to get much better results.
The Wrong Fix
Filling Exception Demand: Company A delivers a mobile service to customers. During their early stage, they discovered that their customers had a preference to have the service performed during business hours, so they optimized their model to capture that business.
When Company A hit the growth stage, they ran into a big problem. Demand during business hours exploded to a point where Company A had to hire more workers and purchase more equipment to serve the workday demand. Now, that’s a good problem to have. However, that same large number of workers and equipment sat expensively idle at nights and on the weekends.
Company A now faced a cold reality. The larger they grew, the more money they would lose.
Complexity Overload: Company B began life as a full-service product with some self-service steps. As their technology evolved, they built out a self-service framework product that allowed their customers to do more and more of those higher-complexity full-service tasks on their own.
Eventually, Company B started to bloat, hiring a bunch of people to develop the self-service framework and a bunch more to provide the full service as they grew. To remedy, Company B threw everything at the self-service framework, including the people resources on the full-service side.
The result was a product that was way too complex for their self-service customers and way too underperforming for their full-service customers, who didn’t want to be self-service in the first place. Company B went into crisis mode for months, making no one happy along the way.
Feature Removal: Company C launched with a feature that automatically offered their best customers the ability to reserve premium slots. This feature was critical to building a loyal customer base. However, as Company C grew, they were getting more and more no-shows in these premium slots, as their best customers snapped them all up “just in case.”
This not only wasted the premium slots, but the growing lack of availability of those slots for their regular customers caused the regular customer usage rate to drop.
To solve the problem, Company C killed the reserve offering for their best customers. Their best customers screamed and many of them bolted. Eventually, Company C had to turn the feature back on to stop the bleeding, which resulted in the same problem all over again.
Like I said, these are just a few examples. And I’m not judging because these are difficult problems to solve. Rock-and-a-hard-place kind of problems.
Solving Customer Behavior With Variations on Peak Pricing
We’re all familiar with peak pricing. In fact, we’ve probably done the math in our heads that we’re going to get dinged the extra X bucks for Uber to get us back from the pub district on a Saturday night.
But Uber didn’t invent peak pricing. Airlines, hotels, rental car agencies, any company that offers a service-turned-product usually works with peak pricing. Disney World, the happiest place on earth, started using peak pricing this year, charging different ticket prices for entry at different times of the year.
Yes, happiness is more expensive during the holidays and in the summer.
But let’s not get cynical. The fact is that the crowds at Disney World during the holidays and the summer were untenable. The alternative was turning people away at the gate, and in the meantime the folks who did get in would have a less-than-optimum experience, with large crowds, long waits, amenity shortages, etc.
Rock and a hard place.
In many cases, we can use a derivative of peak pricing to actually drive customer behavior. Here’s how we would solve each of those problems above.
Filling Exception Demand
The obvious answer to for Company A to fill their nights and weekends with customer demand would be to offer discount pricing at those times. Yeah, but if startup was that easy, we’d all be successful.
When we optimize our model to meet a certain scenario, it’s more than likely we’ve de-optimized the reverse. Thus, due to the de-optimized costs, lowering the price on nights and weekends would mean losing money on the service. To make any discount work, Company A would need to raise prices on the workday demand, which, for a host of critical reasons, was not an option.
The opposite of peak pricing is opportunity pricing, where we develop a new line of business that acts like the peak line of business, and discount that. Company A began offering opportunity pricing to wholesale customers, those who could meet workday demand numbers but didn’t have the same needs as the retail customers. Once the wholesale business was established, their retail customers immediately became peak pricing customers without having to pay higher prices.
Company B pulled the trigger too quickly on moving from full service to self service. What they needed was a smoother transition once they realized the move to a self-service model was inevitable.
Tiered pricing is a form of peak pricing that shifts demand over a certain time as opposed to during a certain time. A good example of tiered pricing came from Netflix, who used to be in the DVD-to-your-door business.
Netflix launched streaming in 2007. By 2011 they had realized that this was the model to bet the house on, so they announced a split between the two services and a price hike. The idea was right, but again, the timing was bad and the change was too abrupt. The stock cratered, but it eventually recovered and rocketed.
Building on the Netflix example, what Company B should have done was gradually raise their full-service prices to fund expediting the development of the self-service offering. This would have provided three benefits.
- They would have gently forced some of their full-service customers to self-service.
- They would have forced out some of their full-service customers who would have ultimately been unprofitable as the market dynamics changed.
- They would have retained some of their full-service customers who would have paid the premium to never have to touch a keyboard.
Company C needed to keep alive the ability for their best customers to reserve premium slots, while at the same time curbing no-shows and making more of their premium slots available to regular customers. They solved all three problems with one peak pricing model.
Peak pricing is actually a subtle form of penalty pricing. We’re penalizing customers for using our product in a way we didn’t prepare for. Again, let’s not be cynical here and lay blame, but let’s find another way to penalize the customer without killing the feature.
In this case, Company C implemented a $5 penalty for no-shows. The penalty is very small in comparison to the product economics, less than 10% of the value of the slot. The low number is key, because we want the customer to take the risk that “something may come up” which might prevent them from taking the slot, but we also want to curb the abusive behavior of reserving a whole bunch of slots just to have them available.
In fact, Company C never has to enforce the penalty. The best way to avoid having to penalize customers is to notify customers about the penalty before they transact. These are Company C’s best customers. They don’t want to piss them off, just change their behavior.
These are just three examples I’ve been involved with, but to underscore the power of peak pricing on customer behavior, I’ll leave you with this:
Yesterday, Uber announced that they will start deactivating riders with low ratings. Even though that’s a sledgehammer tactic and kind of shocking, I totally understand the move. But I can’t help but wonder if it’s a gateway to another type of peak pricing to adjust customer behavior, only this time adjusting personal behavior.
Penalty pricing for being a jerk. I wonder if that would work.