In this article, I offered a simple and powerful AI framework that sales leaders can use to meet and exceed their goals through better planning, prioritizing, and ideating. Planning is about predicting what will happen so that you can be adequately prepared. Prioritizing, or targeting is about identifying the best opportunities for action.
Put another way, planning is about accepting what will happen, while prioritizing is about finding the right opportunities to change what will happen. For instance, if I know that a valuable customer is likely to churn, planning allows me to prepare for that eventuality. Prioritizing allows me to target that customer with an intervention that will compel them to stay.
Ideating, on the other hand, focuses on generating ideas for effective interventions. It’s one thing to identify a valuable customer as one who should be saved. It’s another to know how to save that customer.
Ideation through AI can be tremendously valuable. By identifying the factors that relate to an outcome, we can start to generate potential solutions that lead to better outcomes. Better outcomes can be anything that’s important to a business – less customer churn, more closed deals for a sales team, lower mortality for a hospital, fewer machine failures for a manufacturer, and so on.
Often, the factors that relate to the outcome are referred to as “drivers.” By examining drivers of an outcome, we often can come up with ideas about how to cause good outcomes to occur.
For instance, imagine that an AI model reveals three factors that are related to whether salespeople beat their quota in a given month:
Driver 1: whether that salesperson has engaged in a sales training program
Driver 2: how many times the salesperson has called or visited their customers
Driver 3: the salesperson’s number of years of selling experience.
It is helpful to note that Drivers 1 and 2 are behaviors. Driver 3 is a characteristic, or demographic. How can these drivers be used to generate new and valuable ideas? First and foremost, we must recognize the difference between causation and correlation.
For Driver 1, Imagine that those who attend the training sell more than those who don’t. Does this mean that the training is effective? Maybe. Perhaps the training is full of great information that causes the salesperson to more effectively convey the benefits of the product to the customer. If the training is, indeed, effective, then a company might be wise to mandate it for the entire sales team.
However, just because the training is correlated to higher performance doesn’t mean that the training necessarily causes higher sales. It is entirely possible that only the most motivated sellers choose to attend the training. If so, then their sales success might not be due to the training, but rather to their pre-existing level of ambition and motivation. If more motivated sellers take the training and less motivated sellers don’t, we would be foolish to conclude that the training was necessarily the cause of the difference in performance.
In this case, there is a third, separate factor (motivation) that is related to the outcome. It would be unfortunate for a sales team to mandate training that, in fact, was not actually effective.
Consider Driver 2: the greater the frequency of sales calls, the more likely a deal will close. Does this mean that placing more calls will yield more closed deals? Perhaps – it is possible that more calls give the customer more insight into the offering and the seller a greater chance to develop a fruitful relationship.
On the other hand, it is possible that only serious buyers will spend time regularly engaging with the sales team. And a seller will only place sales calls to buyers who demonstrate high levels of interest. Here, more sales calls don’t cause a customer to be more interested; rather, high levels of customer interest cause a salesperson to call more frequently. This is often referred to as reverse causation. In this case, asking your team to place more sales calls than normal would likely yield no benefit.
A business experiment consists of rigorously assessing the effectiveness of an initiative (training, more sales calls, etc.) by creating two randomly selected groups and assigning one to receive the initiative and one to not receive the initiative. For instance, to test whether Driver 1, a sales training program, is effective, we would want to randomly assign some sellers to take the training and some sellers to not take the training. If the training group were to outperform the non-training group, this would be good evidence that the training was an effective way to boost sales. In some cases, experiments are not feasible, requiring technical, second-best, analyses to assess whether an initiative actually works.
How, then, can we know whether an initiative is truly effective at causing good outcomes?
In short, the best answer to this is to run a business experiment, described in the sidebar. Business experiments are excellent ways to move from ideation (“here’s a possible solution”) to knowledge of whether a solution is truly effective.
Driver 3 tells us that salespeople with more experience tend to sell more. “Years of experience” is a characteristic, not a behavior, so we no longer care about causation versus correlation. (Since we cannot change someone’s age, it doesn’t matter if changing their age would cause sales to increase.). However, demographics still have a valuable role in ideation. Here, we can explore the reasons why experience matters. This could lead to more effective training or best practices for less experienced sellers. Often, drivers provide opportunities for more learning, instead of directly actionable insights.
AI-fueled ideation is a powerful tool. It allows us to find stories, trends, and phenomena that, without computational sophistication, would remain hidden from the human eye. By knowing what factors relate to important outcomes, we can empirically generate ideas that lead to more business success.