The reputation of dynamic pricing has evolved almost as rapidly as the technology and talent that power it. What started as a way for airlines and hotels to balance supply and demand for perishable inventory has transformed into a way for any company to deploy artificial intelligence (AI) to process the myriad financial and commercial inputs in order to achieve pricing’s Holy Grail: the right price decision at the right time for a specific customer.

But many B2B companies remain skeptical. Reinforcing their doubts are several myths and misconceptions that label dynamic pricing as too daunting or as high-tech overkill. This resistance persists, even though several trends are creating compelling opportunities and strong arguments for B2B companies to adopt it. These trends include increasingly volatile demand—especially during the current COVID-19 crisis—along with accelerated product life cycles and the proliferation of readily accessible data.

Our research shows that B2B companies can generate higher revenue and margins by implementing a more dynamic approach to pricing. The first critical step is for companies to see through the myths around dynamic pricing, which will help provide the ambition and the direction to start their journey.

Myth: Dynamic Pricing Is Just “Normal” Pricing With Real-Time Updates

It is no surprise that many managers believe that dynamic pricing is synonymous with real-time price changes, given the way that the business press and others use travel and leisure companies as their go-to examples. But dynamic pricing takes on many forms besides constant, real-time price updates. In the broadest sense, a company’s journey toward dynamic pricing is one that moves away from static, one-size-fits-all price recommendations informed by occasional inputs. At their more lucrative destination, companies use AI to analyze numerous inputs in near real time to generate pricing outputs that can be tailored all the way down to the individual customer level. (See Exhibit 1.)

Moving away from a static pricing approach means listening for changes in market conditions, competition, and demand, and then using those inputs to make the right pricing decision. A company with long product life cycles and only a small, homogenous group of direct customers may require only infrequent price changes. A company with a large and diverse customer base, multiple distribution channels, and shorter product life cycles will have a greater need for more frequent and finely calibrated price changes. In either case, though, the company benefits from the control and confidence that AI-powered dynamic pricing provides.

The transformation at one major automotive manufacturer shows how a company can move from static pricing to a more profitable dynamic pricing approach. The company improved the cost-efficiency of its incentive program 10% by shifting it from one based simply on country and vehicle, to a more sophisticated program based on criteria such as region, vehicle trim, and the type of incentive. The granular inputs were the purchase data on each individual vehicle: model, options, time and place of purchase, etc. Using an optimization engine, the automaker estimated elasticity by incentive type (such as cash, finance, or lease) at the microsegment level in order to generate the granular output it needed: the expected time each individual car would spend on a dealer’s lot until it was sold, based on all possible incentive scenarios. Using those insights, the manufacturer selected the optimal incentive scenario, making tradeoffs across the current month’s goals for profit, share, and volume as well as the number and mix of vehicles left on dealers’ lots.

Myth: Dynamic Pricing Is Useful Only for Rapidly Changing Industries

Strictly speaking, this is not a myth. It relies, however, on the mistaken belief that some industries do not change rapidly. All industries—even those with longer product life cycles and relatively little disruption—are prone to political, environmental, societal, or technological shocks. The recent COVID-19 crisis is an extreme example. But even without this kind of disruption, the availability of data alone offers a compelling rationale for apparently stable industries to reinvent themselves with dynamic pricing. The way that freight companies have used traffic information to optimize their costs is one of many examples of data utilization.

A petrochemical products wholesaler demonstrated that dynamic pricing can generate significant bottom-line impact in industries that are not generally perceived as rapidly changing. It improved margins by between 100 and 250 basis points by shifting from a mostly manual, less disciplined pricing approach to an automated pricing solution based on machine learning. But that is only one example of a dynamic pricing journey in a B2B industry. Dynamic pricing encompasses several use cases that address different needs and unlock significant value. (See Exhibit 2.)

B2B companies with list-and-discount pricing structures can use dynamic pricing to set optimal list prices that take into account market, competitive, product, and cost factors. Tech, software, and telecommunications companies, among others, rely heavily on subscription business models and contract- or deal-pricing mechanisms and now use dynamic pricing to optimize their on-invoice and discretionary discounts. Distributors and B2B retailers can use dynamic pricing techniques to optimize their promotion planning by calculating promotional impact in advance. For B2B companies with the customer base and the market conditions to warrant a yield management approach, dynamic pricing means using AI to match demand and willingness to pay in real time at the most granular level, even down to the segment of one. B2B e-commerce platforms use efficient auction management to optimize prices, accounting for real-time demand and the willingness to pay of individual buyers.

Myth: Dynamic Pricing Is Just a Black Box

The technical heart of dynamic pricing is obviously the pricing engine or algorithm, but treating that engine as simply a black box is a recipe for failure. Dynamic pricing is not an “either-or” choice between machines and people. Rather, it is a “both-and” solution that combines the strengths of both forms of intelligence.

People create the pricing benchmarks, constraints, and business rules that inform the dynamic pricing engine’s willingness-to-pay and competitor modules. Likewise, human judgment and intervention are vital to transforming the engine’s outputs into actual price decisions. Finally, the algorithms and engines have to be understood—at least to some extent—by those who will use them daily. The key is to invest in automation, strong teams, and robust processes, and to involve users continually in the design process in order to ensure that employees see dynamic pricing as more than merely “black boxes” and “cool tech.”

Salespeople don’t need to become data scientists themselves, but they won’t embed dynamic price recommendations into their routines without trust and confidence in those recommendations. The pricing team will need to explain how the engines work and ideally bolster salespeople’s confidence with in-market pilot tests.

A B2B distributor in the building material sector increased its EBIT margin 100 basis points by using this combined human-machine approach to implement its dynamic pricing program. The company manages a catalog of hundreds of thousands of SKUs sold to tens of thousands of clients from many diverse industries. In order to set and continuously update billions of price points, the company implemented an algorithm that pulled in a complete set of historical transaction data as well as detailed characteristics of the clients from both internal and external sources. Machine learning computed a personalized willingness to pay for each product and each client. But in order not to jeopardize long-standing client relationships, the sales teams systematically reviewed the recommended price positions. A feedback loop fostered an ongoing exchange between the users and the advanced analytics and data teams to improve the model.

Myth: Dynamic Pricing Requires Perfect Data…and Years to Implement

Who wouldn’t want perfect data for any task? A lack of perfect data, however, is not an impediment to dynamic pricing, nor is it a reason to delay it or avoid it. In fact, it is an incentive to implement it. An organization’s ability to collect and analyze granular inputs improves with time and repetition, because it prioritizes the data and makes its collection more efficient.

No company starts with perfect data or a perfect infrastructure. But starting with a small number of use cases and implementing agile development cycles will allow a company to quickly identify and prioritize the data it needs.

Think of dynamic pricing not solely as a technical, tactical, data-driven tool, but rather as a journey towards more automated, frequent, and personalized price recommendations. The companies that have quickly and effectively integrated dynamic pricing into their operations have treated this journey as a cross-functional design challenge, not as a technical challenge for the data scientists and the IT organization. Companies generally build capabilities sequentially, focusing on enhancing either inputs or outputs first, rather than both at the same time. Some will build personalization capabilities first, while others focus first on capturing real-time inputs. Each successive step in the journey adds value.

The journey can also move quickly. A large industrial distributor of maintenance, repair, and operations supplies needed only 12 months to shift its list price setting to a dynamic pricing engine, augmented by competitive price scraping. The move generated $40 million in near term profit improvement and drove significant volume growth: an increase in the first year of 10% for the entire business and 20% in its most profitable segment.

How B2B Companies Can Accelerate Their Dynamic Pricing Journeys

The adoption of dynamic pricing is not plug and play, but rather a journey unique to each organization. Regardless of where a company is on its journey, though, it can accelerate its progress by making deeper, ongoing self-assessments with respect to the must-have capabilities for successful implementation of dynamic pricing. (See Exhibit 3.)

Based on these self-assessments, a company can estimate and recalibrate the financial upside, the commercial effects, and the time required.

• Technology and Data. How much high-quality data do you have, and how automated are the current processes for data collection and processing? In some industries, the pricing mechanisms might limit the granularity of inputs that companies can collect. A building materials company, for example, may not have much use for input capture capabilities in real time, but may have a keen interest in other use cases, such as optimizing its segment-specific pricing or discount programs. Such a company would have a rich set of data to draw on: information on customers, products, costs, and competitor prices, as well as contract details, wins and losses, sales overrides and comments, escalations, and negotiation histories.
• Algorithms and Pricing Engine. How does your current pricing logic work? What inputs and calculation logic could you add in order to derive optimal price recommendations? Some B2B companies have already made steps toward rigorous, data-driven pricing, and others have some history and experience with pricing tools for marketing and sales. But the needs and potential will differ greatly. A subscription-based tech company that uses contract pricing will not apply dynamic pricing the same way as a transport-and-logistics company doing capacity-based quotes, or the same way as a telecom carrier that already quotes prices based on multiple inputs (capacity, competition, call duration, data usage patterns, etc.)
• Process and Governance. How well do you enable coordinated and controlled pricing decisions? As powerful as the analytical tools become over time, successful dynamic pricing ultimately depends on how well people exercise control over the process. Sales and marketing contribute their market knowledge and the human side of pricing, supplemented by the technical expertise of the data science team. The pricing team controls the model and improves it over time, and IT runs and enhances AI systems.
• Team Enablement. How should the change to dynamic pricing be hard wired into your everyday pricing routines? Developing dynamic pricing capabilities is not an easy journey; it will require a strong mandate and the support and involvement of leadership to be successful. Management’s role is absolutely essential to ensure cohesion across the multidisciplinary teams. An enablement process helps the teams use dynamic pricing to their advantage, and make it stick. In addition to training, that process helps instill an agile mindset and a test-and-learn culture. Governance, pricing and sales processes, and organizational responsibilities will evolve as the company creates feedback loops, focuses on continuous improvement, and trains the salesforce on how the pricing engine works. Communicating success stories along the way helps the whole organization embark on the transformation effort.

B2B companies that successfully implement dynamic pricing started their ambitious journeys by learning from one use case, then continuing to build and improve capabilities one use case at a time. An initial assessment of their pricing strategy and pricing capabilities enabled them to define their journeys and how to manage them.

To accelerate their journey, these successful companies took an agile approach with the right mix of talent from throughout the organization. The longer a B2B company takes to recognize how it can benefit from dynamic pricing, the harder it will be to seize its opportunities to pursue that Holy Grail: the right price decisions at the right time, derived with confidence from the right mix of human judgment and technology.