One of the most powerful levers a CEO has to impact profitability is pricing. Typically, a 1% price increase has more than twice the impact on profit of a 1% volume increase. However, optimizing pricing is hard enough for the CEO of an established company, let alone a startup with a new service, often for a new market.
My longtime friend Mark Dresdner has deep expertise in pricing. Mark has been a startup CEO, management consultant at Accenture, leader of revenue strategies for Starwood Hotels and a $1B aviation company, and spearheaded growth at a portfolio company of the private equity arm of Macquarie. He generously agreed to answer a few pricing questions that come up frequently for early-stage companies. (Check out his site, StratapultAdvisors.com, where you will find a number of ebooks on pricing and related topics.)
David Teten: What are the major options in pricing a service (e.g., a SaaS startup), and how should I pick among those options?
Mark Dresdner: First, thank you for the opportunity to share ideas that I hope others will find useful.
There are four typical pricing models for services that can be used separately or in combination. Let’s use the SaaS market as an illustration.
Volume based pricing allows the price to change with usage. Typically the per unit price declines as volume increase, although sometime a free offer for low volume is available to encourage trials. The volume model is good for situations where switching costs are high or operating costs grow with usage. With the email marketing system offered by MailChimp, which I see you use for Teten.com, it’s easy to get started with a free service for accounts with less than 2,000 email addresses. However, once MailChimp starts charging users they typically have their email marketing setup with templates, segmented lists, and autoresponders established, making switching painful. As a user’s email list at MailChimp grows, so does the monthly cost, although the cost per email address declines. If operating costs grow with usages,volume based pricing helps companies keep large volume users profitable.
Subscriptions charge a recurring amount. Google Apps, for example, charges businesses a monthly or yearly fee per user for services like email, integrated calendar, and storage. Subscriptions work well when the user becomes a long-term, frequent user. This way the price point doesn’t seem high, but is paid regularly so the revenue adds up over time. Given that subscriptions often allow for a high or unlimited level of usage, they are best used when the incremental cost of additional usage is small.
Packages allow users to pay for groups of services based on their needs. QuickBooks, for example, has core accounting packages, plus an additional package to support payroll. It is suitable to offer packages when different customer segments have different needs. Packages work well when a customer’s needs develop overtime, encouraging them to upgrade when they can see the value.
Freemium offerings allow customers to use the basic service for free and them pay for premium services. Companies like LinkedIn use the freemium model to encourage a volume of customers–which creates its network value–while still charging for advanced functionality to power networkers, such as recruiters. The freemium model usually is only suitable when a huge number of users are expected, because the share of those who upgrade to the paid services is often small (4% is best practice). And, the value of the premium, paid for services needs to be high for a segment of customers, while at the same time the free services need to valuable enough to attract many customers.
David Teten: Once I’ve figured out my pricing structure, what are the best ways to test prices?
Mark Dresdner: Great question. Price testing is often neglected, but the lessons learned could lead to a dramatic impact on profitability. Given that many startups sell through the internet, let’s use that model for our examples of ways to test prices.
1. A/B Testing
A/B testing involves offering different prices that are allocated randomly to customers. From a statistical perspective, this is a very clean approach because the only thing you change is the price. An easy way to do this is to offer a variety of price points on a website using an A/B testing program like VWO or Optimizely. This type of price test is often used by companies such as Dell when they release new products.
2. Offer Discounts Off a Stable List Price
By setting a relatively high list price and then showing a discount, one has the option of varying the discount and hence the end price can change without showing a frequently changing list price. This provides a flexible platform to try different prices without causing jarring changes to the overall pricing and it makes it easier to increase pricing by just removing some or all of the discount.
3. Segment Your Customers
If a company can segment its customers into groups that are very similar, it can offer different pricing for each segment–this is a manual way of simulating A/B style testing. For example, if there was a landing page off a specific blog, those visitors could be offered a different price (although, it should be lower than list price as they might make their way to the general pricing on the website). When price testing is done based on segment, one needs to be aware that different segments could have varying levels of price sensitivity, which could cause havoc on test results if the goal is to find a general market price.
David Teten: Our portfolio company Phone.com is in an extremely competitive market, telecom services for SMBs; so currently they must compete on price just as much as they do on features and applications. Good customer service and a solid offering can let Phone.com charge based on value, but that does not scale until we have a much more recognized brand. The management team has observed that their clients are all familiar with cable companies offering unlimited service, or or mobile services offering unlimited minute plans, and so on. Even though they’re in a very different market, that led them at one point to offer a product at an extremely low price point of $4.99 and without having an unlimited plan. Instead, they announced that the monthly fee above the allotted monthly minutes included in the $4.99 will be priced at $0.03 per minute, but that we will cap the monthly charge at $18.99. To our astonishment, we got many negative comments from upset customers who complained we don’t offer an unlimited plan, even though our offering was much better than an unlimited. If you went over your minutes by $7 you would end up paying $11.99 and NOT $18.99, so we actually saved the customer money and hurt our margins. Our customer service agents had to do quite a bit of back of the envelope calculations in order to convince customers that we are better than unlimited. The market was fundamentally not ready, we think, for this plan. I’m curious to hear your thoughts on customers’ reaction to the plan?
That is a fascinating case study. I believe that Phone.com made an incorrect assumption, which a lot of other companies also have made in this situation. They assumed that customers would think like analytical economists, but in the real world few do. Customers’ reactions to Phone.com’s pricing can be better understood by considering two concepts: prospect theory and the customer’s frame of reference.
Prospect theory contends that people are impacted emotionally more negatively from a loss than positively from an equal sized gain. What is your gut response to a stock investment falling 30% compared to increasing 30%? If you are like most of us, the gain is nice, but the loss is likely to create a more powerful emotional response. Because people shun losses, pricing should be positioned as gains; for example, as discounts rather than penalties. For example, it is more palatable to present an invoice of $1,000 with a $25 discount for payment within 30 days than an invoice of $975 with a penalty fee of $25 if not paid within 30 days.
In Phone.com’s situation, the customer sees that they are being charged–and therefore penalized–for each minute over their allotted monthly minutes included in the base $4.99 until they reach a limit. Based on prospect theory, customers’ emotional response will be negative.
Frame of Reference
A frame of reference for pricing is what customers are comparing a price to. For Phone.com, customers’ frame of reference, as you pointed out, is the unlimited plans that are now available. These unlimited plans in telecom are loved because the customer receives predictable bills and doesn’t feel a need to monitor minutes used. Phone.com’s plan that charges for minutes is likely to bring back memories (and nightmares for some) of another frame of reference, the historical phone company pricing plans that have a reputation for nickel and diming from usage based charges. Phone.com’s messaging could consider people’s frame of reference more directly.
Idea for Phone.com
One way Phone.com could show simpler pricing that addresses customers’ frame of reference, prospect theory and communication is as follows.
$18.99 per month. Unlimited calls. Save up to $14 when less than X minutes are used. That’s a savings of up to 75% when you aren’t a heavy user.
This approach, while still charging exactly the same amount, offers the following benefits:
- Clear, in your face amount for unlimited calls. Provides predictably.
- Savings rather than extra charges to adjust the invoice.
- Allows easy comparison to the frame of reference created by the competition.
Some sort of savings calculator could also help people compare Phone.com to the competitors, whom usually don’t offer the ability to save when call volume is low.
David Teten: ffVC’s portfolio companies Whisk and Surfair are providing subscription pricing in sectors which historically did not offer it–ground transport and air transport, respectively. When does subscription pricing make sense?
Subscription pricing is certainly gaining in popularity and can be suitable for a number of situations, particularly when a number of these situation co-exist.
Costs Are Predictable
When costs are predictable, companies can offer subscription pricing without the fear of heavy users destroying profitability. The predictability of costs can be due to the volume being predictable or the incremental costs per use being very low.
Segment of Recurring Revenue to Contribute to Fixed Costs is Desired
When a sale covers the incremental costs incurred from the sale, the profits contribute toward paying for the company’s fixed costs. Targeting a specific customer segment with a subscription service can provide a regular cash flow to help pay for fixed costs, while other non-subscription customers typically pay much more per use and deliver higher profit margins. The recurring revenue from subscriptions can also help a startup establish less volatile financial performance, as Jeff Bussgang discusses in his blog post Why Recurring Revenue is Magical.
Ongoing, Regular Use
When users believe they are going to regularly use a service, a subscription is a convenient way to bill. It is important to note that users don’t have to actually be frequent users, they just need to think that they will be. This has powered the profitability of gym memberships, because a large percentage of members don’t use the gym often, but it takes a long time for them to admit that they should cancel their gym membership. If fact, I have seen a survey showing that around two-thirds of gym members rarely use the gym.
Customers Don’t Want to Think About Incremental Fees
Often customers will pay a premium for subscription that will remove the frustration or stress of worrying about incremental per use or occurrence costs. One example is phone plans, as we mentioned before. And a different example is appliance warranty programs, where for a yearly fee appliances like your refrigerator or washing machine will be repaired, if needed.
Subscription Pricing as a Game Changer
Surfair’s subscription pricing complements their game-changing market position. They offer a monthly membership price for an individual to have unlimited access to Surfair’s flights within a network of airports on their comfortable, seven-seat general aviation jets. For customers, they have predictable pricing for a service that enhances their life–reduced travel time and enhanced productivity in the air–and they believe they will use regularly. For Surfair, they have recurring revenue covering costs that are fairly fixed. Their main costs are plane ownership (which is fixed) and fuel, which is almost the same if they have one passenger or six (so it is fixed for passengers past the first).
Whisk is using monthly subscription pricing to provide a unique product to regular taxi users. Here one can purchase an unlimited number of car rides between two locations or within an area in New York City. Customers now don’t have the pain of paying per ride and can grab a Whisk ride without doing a cost benefit analysis in their head, because they don’t have to think about the incremental cost. While Whisk gains steady revenue from a group of loyal customers, helping cover fixed costs, while still having much of the revenue come from pay per ride sales, which can often have higher profit margins. One thing I recommend Whisk watches is that it doesn’t get hurt by customers self-selecting their pricing to the point that profits decline. This would occur if the heavy users switched from pay per ride to subscription and their usage ends up higher than forecast.
David Teten: Uber has received a lot of attention for their experiments with pricing. What do you think of their approach?
You’re right, Uber is taking price testing seriously. Uber, the current leader in the car service and alternative taxi market, is testing the combination of two different pricing strategies.
1. Different Brands, With Different Price Positioning: Uber offers five brands from UberX to UberLux. The two main brands are:
- UberX is the lowest price positioned of the Uber brands and uses everyday cars. UberX is following the Southwest approach. Southwest, when it started, priced to compete with the cost of driving or traveling by bus and priced below the large carriers’ operating costs. UberX is pricing to compete with owning and using one’s own car, with the goal of having price points below taxis.
- UberBlack provides a professional experience with high-end sedans. UberBlack competes with the traditional “black car” service offered in large cities.
2. Surge Pricing
Uber’s surge pricing is following a traditional yield management approach to pricing, much like a hotel or traditional airline. When passenger demand is low relative to cars available, prices are low, and as demand increases relative to car capacity, prices increase. Applying yield management for Uber will require a good understanding the price elasticity by market, that is how demand changes with price. In addition, the company needs to feel comfortable that some customers will try competitors when Uber’s prices have surged, undermining loyalty.
There are some things to watch as these pricing tests play out. First, how well Uber can create distinct brands in people’s minds. Second, whether UberX cannibalizes business for Uber’s higher priced brands (and profit margins).Third, whether the surge pricing can keep the price points for the brands in-sync. For example, when there are limited UberX cars available, will its pricing surge to UberBlack levels.
Pricing errors are easy to make because pricing is often not a core skill of a startup team. There are a number of common pricing errors startups make.
Don’t Use Pricing as a Differentiator, They Copy the Competition: Pricing should be part of the company’s brand positioning and can often be used a differentiator in the market. Copying the competition’s pricing not only doesn’t differentiate the company, but also encourages price pressure and commoditization.
Don’t Price Based On Customer’s Perception of Value: The lens every pricing decision needs to look through is the customer’s perception of the value offered and how that differs from their alternatives. While the company has to be profitable, customers don’t care about the company’s costs. Imagine if Starbucks priced using the cost-plus model. A cup of coffee could average $0.90, not around $4.00.
Don’t Consider Emotional Aspects: Customers apply heuristics and emotions to purchases, so we can’t expect the typical buyer to behave like a procurement officer with a calculator in hand. Harley Davidson has done well to build an emotional attachment to their brand, which is reflected in their pricing. Harley Owners Group (H.O.G.) members are buying into a feeling of belonging and an inspiring lifestyle.
Don’t Price Test: It can be challenging for a startup to pick the optimal price for their product or service. One way to do this scientifically is to test different prices to see how customers react. When pricing and the purchase is online, A/B testing – showing different customers different prices—is easy to implement and can show which pricing option performs best.
David Teten: Thank you for your time and insights.
Mark Dresdner: Thank you for the opportunity to share ideas that may help some startup teams price better and increase their chance of success.