Using Propensity to Buy analytics to focus resources on growing the most profitable accounts

Digital technology – notably the advent of advanced analytics and AI – offers tremendous potential to improve the productivity, engagement, speed, and financial contribution of revenue teams. One way these technologies can accelerate sales growth and create firm value by optimizing the allocation of revenue team resources and “cutting the customer tail” to focus selling time on only the highest potential opportunities and accounts.

Targeting accounts by potential is certainly not new.  Every business leader understands the 80-20 rule when it comes to targeting customers,  in theory at least.  But In our experience, most organizations we work with tend to target too many customers and develop too few of them.  There are good reasons for this – the optimism of sellers, the desire to realize more market potential, and the pressure to generate the most revenue growth from scarce selling resources. But they are also the result of some bad habits. For many businesses the “customer curve” remains too long and sellers continue to chase “tail accounts” that are unprofitable to pursue.  These common problems are usually the result of human nature – Not challenging entrenched belief systems. Using “gut feel” assumptions to size opportunities. Relying too much on historical sales data instead of predictive insights when planning. 

Almost every CXO we have interviewed in our Revenue Operations research agrees they  are chasing too many customers and that is hurting every measure of sales performance – quota attainment, coverage and cost of sell, customer lifetime value, and the customer experience.

For example, Elango R, the President of North America and New Clients Acquisition at Mphasis, believes in shrinking the client bullseye in order to grow faster. He advocates a level of account focus that may seem extreme – but in practice is highly effective as a way to grow customer lifetime value, margins, and revenues. “Most people try to cover too many accounts and that hurts their go-to-market effectiveness and margins,” says Elango. “Our best accounts are what we call Horizon accounts, and we only have about 100 of these because of the relationship depth and commitment they require. Our revenue portfolio is around half a billion and in all we are focused on 100 core customers. Using this highly focused approach, we’ve been able to win large deals of greater than $20 million in total contract value (TCV) year over year.”

Elango is not alone in his focus on zeroing in on the most actionable and profitable account opportunities. 85% of the CXOs interviewed in the book Revenue Operations are actively redefining sales force roles, assignments, customer treatment models, and targeting priorities to improve selling performance, realize more market opportunity, and growth customer lifetime value. This is because optimizing territory design by matching territory size with revenue and profit growth opportunities. can increase sales by 2 to 7% and reduce selling costs by 10-15% cost reduction without any change in total resources or sales strategy, according to the Data Driven Sales Resource Allocation Report.

Every decision about growth investment allocation, sales resource planning, market segmentation, and individual seller assignments is ultimately built on key assumptions about account potential. Estimates of the opportunity potential in an account, their propensity to buy, and the level of effort it requires to realize that potential in terms of sales are the atomic material of go-to-market strategy and execution. Most everything stems from these estimates. In most organizations they are little better than guesses based on historical data and undocumented beliefs about the future.

In the short term, some of the most practical and impactful ways data-driven algorithms can create value is to help with the basics of account sizing, prioritization and qualification, and reallocating sales resources to the right accounts, territories, and markets. These tasks are easier for organizations to execute with limited analytics acumen and data scientists in short supply, according to Leonard Lodish, Professor of Marketing at the Wharton School of Business. “Organizations are dramatically improving sales performance by using algorithms to help with the basics of account and lead prioritization and qualification, recommending the content or sales action that will lead to success, and reallocating sales resources to the places they can have the most impact,” reports Professor Lodish.

The majority (62%) of high-performing salespeople see a big role for guided selling that ranks potential opportunity value and suggests next steps. This is because a wide range of AI tools are now available that make it easier than ever for sales teams prioritize opportunities based on buyer intent, marketing signals, and customer engagement data. AI tools can also create algorithmically derived customer response models to help take the guesswork and gut feel out of aligning sales resources across geographies, accounts, and business lines.

Over half of CXOs interviewed in Revenue Operations  are using commercial insights to dramatically sharpen the targeting and focus their selling resources. For example, Peter Ford, the VP of Global Sales of iconectiv, was able to leverage analytics to “cut the tail” by becoming more scientific about prioritizing customers based on readiness and potential. This information enabled different levels of customer treatment to manage cost to sell in smaller “tail” accounts. “We use analytics to create the curve of our customers,” relates Ford. “Much like every other technology business, I have a relatively small number of high spend customers who purchase almost every single one of our products. And then we have a long tail of much, much smaller customers. We can’t manage all of those customers in the same way from a sales perspective. So, we treat a customer that generates less than $10,000 of revenue for us a year differently than a customer that generates in excess of $20 million a year. Understanding when to take a light touch approach versus hands-on sales or when an account should be more marketing driven is key.”

To do this, their operations teams are developing highly accurate propensity-to-buy targeting models using existing CRM and transaction data that will more accurately predict which customers are going to buy from you, with the least selling effort are seeing near term gains of 20% or more in conversion, sales quota attainment, and account development.

Growth leaders are using advanced analytics to create better predictions to inform their investment bets and evaluate more scenarios to optimize resource allocation decisions. AI-enabled algorithms and propensity to buy analysis can help you develop more accurate and predictive estimates of opportunity potential. They can also teach you more about customers – who they respond to, whether they intend to buy, and how they fit with your sellers, products, and treatment types.  In the absence of analytics that quantify account potential and propensity to close, most organizations chase too many low-quality clients and opportunities – ones where they stand little chance of winning but want to feel like they are still in the game, according to Cam Tipping, who has led over 100 customer targeting workshops and even more simulations with sales teams over the last decade. “This behavior leads to bad outcomes – like bad service, high cost to sell, lower margins, and unsatisfied clients. It’s a basic 80/20 problem. But a difficult one to solve. No business that I have seen is aware of this issue until they go through the analytics to understand it,” continues Tipping.

“Traditional CRM or financial systems are not set up to provide this type of information. So historically account prioritization and segmentation was developed using custom analysis and a review of client performance over multiple years, and a lot of gut feel and local market knowledge. That’s an unprofitable formula for allocating sales time and attention in 2022. The data exists to create the facts that let us ‘cut the tail’ . Today managers have ready access to effective  tools like deciling and propensity to buy analysis that use this data to identify and achieve consensus on the 80% of accounts not to call on.”

So why aren’t more organizations using algorithms to define or at least augment and refine their customer segmentation. There are a number of human factors at play.

People target too many customers because of the perceived risk to their revenue plan, the natural optimism of salespeople, the desire to help every customer. On a management level the biggest factors can be institutional bias about what accounts, industries and targets are most attractive and legacy assumptions that are neither documented nor challenged.  Most sales resource allocation and territory designs are actually based on estimates of the total addressable market when account teams need to understand the total actionable market to really be effective with their time.

Another reason is most managers are still uncomfortable relying on technology for this type of planning. Most executives are more comfortable trusting AI to make a million calculations a second to drive their car, then relying on analytics to analyze different customer data sources to refine their targeting and segmentation. Targeting is a far simpler analytic task relative to autonomous driving, but on one level, their familiarity with the information and the perception they can manage all the variables on a spreadsheet works against them.

So how can managers make the move from gut feel towards more fact based and algorithmic customer targeting?

The secret is to convert the customer engagement data you already have into targeting and segmentation insights. Sellers don’t need more data. They need actionable insights that inform account priorities, resource allocation decisions, and the level of effort to apply to specific target customers. Machine learning has been democratized over the last several years with advances in toolkits and packages making it possible now for all organizations to have access to these advanced analytics. It’s not just the realm of Google and Facebook. You can really find a way to bring practical analytics and big data to the kinds of problems that you’re trying to solve in the commercial organization.”

For example, it is possible to develop highly accurate propensity-to-buy targeting models using your existing CRM and transaction data. This can be done pretty quickly. A propensity-to-buy model will more accurately predict which customers are going to buy from you, with the least selling effort, and which ones are not likely to buy or will require too much work to convert. “When compared with the estimates of sales teams and local market leaders, most see near term gains in conversion, sales quota attainment, and account development.

For example, it’s pretty straightforward for most B2B organizations to build a propensity-to-buy model that segments their customers into three buckets used for account targeting – high medium and low. “This can be done by combining the traditional data points used for targeting, including firmographics around the targets – regions industries – with the CRM history of the customer to create a machine learning model. And it’s not about man versus machine. It’s about man plus machine. We get lift from the quantitative data, but even greater improvements when we work with teams to combine the quantitative learnings from ML with the human learnings of the sales team to get the best outcomes.”

Using analytics has advantages in accuracy, speed and scalability.  Obviously, the first benefit is to able to target higher probability leads. But these models also can do this at scale. It takes minutes to score the thousands of potential customers using our machine learning models. If they’re doing it by themselves with intuition, with gut feel – even if they are right, it still takes a lot more time and a lot more effort. This is about the long tail. This is identifying the customers that are not likely to buy. So when we predicted that a customer was low probability, they were low probability. The salespeople would be wasting their time going after these accounts.

You May Also Like