Advanced analytics in software pricing: Enabling sales to price with confidence

| Article

Discount management is returning to the center stage of commercial management in enterprise software: traditional software players face increasing discount pressure as they compete with disruptive next-generation players, while vendors migrating to subscription models or to software as a service (SaaS) find that discounts in initial deals are the main determinant of future customer lifetime value. However, loose discounting practices are hard to rein in: sales representatives are often motivated purely on bookings, the low marginal cost of software drives an “every dollar is a good dollar” mentality, and the common and myopic quarterly management approach of closing deals at any price at the end of the quarter all drive average software discounts to levels not seen in the past.

Sales reps often argue that higher discounts are necessary to win deals, but our research across companies consistently shows the opposite to be true: successful deals almost always have lower average discounts than deals that were lost.

Management usually responds in one of two ways to counter excessive discounting. The first is by tightening the approval process. However, this slows the system down without significantly changing behavior; typically, more than 90 percent of deals that get escalated for more scrutiny get approved anyway, and the discounting continues. The second is by increasing list prices. But this tends to perpetuate the spiral of higher and higher discounts, leading to the 60 to 80 percent average discounts that we often see for perpetual licenses for on-premises software today.

As a result, discount variability in software is enormous. A certain degree of variability is expected—the right discount often depends on many different factors, such as customer segment, product mix, size, and geography. But these factors do not account for all of the pricing variability we see in the field (Exhibit 1). That unexpected and unnecessary variability represents lost value.

1
Significant, unexplained deal-pricing variability is a common challenge.

The root of the problem is a lack of insight into objective comparison points. Although sales reps have a good feel for the market, they don’t have much information about how their colleagues price similar deals. And to managers, every deal can look unique, forcing them to make approvals based more on gut—or accept the sales reps’ argument that the proposed discount is what it takes to win the deal—instead of an actual objective fact base. In software as in other B2B industries, the pricing challenge is not so much how to deal with big data—it’s how to get valuable insights out of limited data sets.

Sales leaders can address these challenges head on with new advanced-analytics techniques to gain quantitative deal-structure insights in a B2B setting. It changes the sales conversation fundamentally by moving the focus away from margins and discounts and instead putting it on objective deal scores. Embedding insights deep into the commercial process—including quote configuration, compensation, streamlined approval levels, and a new approach to sales performance management—can create significant improvements in return on sales: 4 to 10 percent and sometimes even better is typical in software, higher than in any other industry.

Would you like to learn more about our High Tech Practice?

Rewiring the commercial process to leverage advanced analytics for discount management does not only improve commercial productivity and effectiveness—it is also one of the few actions that sales reps actually embrace. Instead of unpopular, top-down corporate discount guidance, reps can use real information to compare their own deals with those their peers are making, which empowers them to make decisions, reduces the red tape they have to deal with, and ultimately allows them to capture a share of the upside through higher commissions.

Three ways to drive better discount management through dynamic deal scoring

Uncover the true drivers of discounting performance

New advanced-analytics approaches dig deep into deal characteristics and identify the factors that truly drive pricing and discounting. They uncover the similarities among deals that previously looked like a blizzard of snowflakes, each different from the next. Deal size, product and service mix, deal type, customer history, and region are typical parameters that drive discounting; however, dozens of different deal parameters often turn out to be relevant in identifying similarities. Similar deals can then be clustered into peer groups, allowing a true “apples to apples” comparison across deals—a deal score. This objective score creates transparency for sales reps into how their deals truly stack up compared with those of their peers, moving the deal-assessment conversation from assertions to facts.

Dynamically score deals as they are created to inform decisions at the moment of pricing decisions

For pricing and discounting guidance to be effective, it needs to be given at the moment that pricing decisions are made, not just at the end of the process when a deal is being submitted for approval. Typically, this means that deal scoring gets seamlessly integrated into the CRM (customer-relationship-management) or configure-price-quote system. Simplicity is key; as sales reps price deals, they receive instantaneous and dynamic feedback through a deal score that assesses how well a deal is priced compared with successful peer deals in the past (Exhibit 2). Sales reps systematically inch up prices to see if they can reach the next better score. These small pricing improvements across hundreds of deals can create substantial impact for software vendors.

2
Dynamic deal scoring provides sales representatives objective guidance on the quality of a proposed deal.

Win the support of the sales team by streamlining the commercial approval process and sharing the upside

Initiatives to improve deal pricing are usually received with suspicion by sales leaders and reps because they can reveal a disconnect between corporate ambitions and the reality in the field. Dynamic deal scoring, on the other hand, empowers the sales team by linking the deal score itself to approval levels; reps who price well compared with what the market can bear will get automatic approvals for their deals. Only objectively poor price levels will receive the additional scrutiny they deserve. This also changes the game for approvers, who are otherwise forced into “rubber stamping” deals. With the deal score in hand, they can instead spend quality time on critical pricing decisions informed by an objective fact base.

Finally, for pricing guidance to truly change sales reps’ behavior, it needs to link to sales compensation. Sales reps who outperform their peers on pricing deserve a share of the upside, with attractive compensation boosters. On the flip side, low deal scores need to be accompanied by an appropriate compensation penalty. Only then will sales reps hold the line on deals that can be won without excessive discounting.

Three common pitfalls in dynamic deal scoring

Blindly trusting advanced analytics

While analytics approaches have advanced dramatically, pressure-testing algorithms with business logic is a key step for a robust deal-scoring algorithm. Not doing so risks providing incentives for wrong behaviors and leaving the algorithm vulnerable to gaming by sales reps. Here’s a simple example: although data will invariably show higher discounts toward the end of the quarter, an algorithm that builds this into its guidance will perpetuate this behavior and give sales reps an incentive to delay the closing of deals to the end of the quarter to optimize their commission.

the Shortlist

Subscribe to the Shortlist

McKinsey’s new weekly newsletter, featuring must-read content on a range of topics, every Friday

Treating dynamic deal scoring as an IT project

Dynamic deal scoring is not simply another sales-tool implementation, nor is it primarily an IT project; we have seen organizations that treat it as such fail to capture the full value of it. If the algorithm is not trusted by sales reps, or if sales reps do not see the win–win nature of deal scoring, the scoring tool is likely to end up on the shelf after six to 12 months, with little to no impact realized.

Underestimating change management

Moving the commercial organization from assessing deals based on discounts or margins to objective deal scores requires an end-to-end change-management effort. The link to compensation, the integration with approvals, and the focus on making deal scoring an integral part of deal and performance reviews are necessary to capture the impact to the bottom line. Tackling these changes as an afterthought will set the dynamic deal-scoring project on a path to failure.


Advanced-analytics tools offer software leaders new ways to address the growing challenge of excessive discounting and associated inefficiencies in the commercial process. Dynamic deal scoring is one no-regrets move to reliably improve software companies’ commercial performance by 4 to 10 percent. But only those companies that truly commit to embedding it in their core commercial process will reap the benefits; otherwise, misaligned incentives and a lack of effective deal reviews will erode the value of the insights.

Explore a career with us