A primer on data-driven selling

TLDR: When your business is not meeting sales targets, how do you know whether your sales team is on the right track? Data-driven selling can be a powerful tool for diagnosing and improving sales productivity issues. It is founded on the following four insights: (1) sales success is highly correlated to the right sales activities with the right frequency; (2) there is a real and measurable point of diminishing returns for any sales activity; (3) you can develop high quality expectations for new and retrained sales personnel pertaining to sales production over time; and (4) data infrastructure and dashboards are key to sales managing themselves and to management managing sales.

When your business is not meeting sales targets, how do you know whether you are on the right track? Do you trust that your high performers will somehow make it happen? What if your sales cycle are long, say on the order of months? How do you know whether you are doing today what you need to be doing to be successful in six months? 

Your sales activity mix is the new media mix

It turns out this problem facing sales leaders today is very similar to the problem faced by marketers decades ago. Lack of data, lack of accountability and traceability… The days of “half the money I spend on advertising is wasted; the trouble is I don’t know which half” are gone thanks to digital, data-driven marketing. With a modern marketing technology stack we can now track every single digital interaction and optimize our marketing mix to generate the most leads at the lowest cost.

The rise of CRM systems is now enabling the creation of large datasets capturing sales activities at a very granular level. CRM adoption rates are continuing to grow, from 57% in 2018 to an estimated 82% now. CRM data quality is improving through better data management (auto-population, cleansing, master records) and better training. This in turn enables us to use sales activity data to predict and optimize sales outcome both at the macro level and at the prospect or transaction level, just as you would with media mix modeling and attribution. 

Sales success is highly correlated to the right sales activities with the right frequency

With proper data management techniques (more on that below) we can now identify and quantify the impact of each sales activity on sales success. Specifically we can analyze and predict using machine learning techniques such as random forest or gradient boosting classifiers whether a prospect will turn into a customer or not and why.

Just as different types of media activities act over different time frames (think brand messaging over broadcast media vs. promotional advertising online), so do different types of sales activities. We can measure and quantify the fact that relationship-building type activities such as networking events and other client events do have a significant impact on sales but one that takes time, typically beyond a 1-year horizon. Tactical activities such as sales meetings on the other hand have as expected a shorter time frame for impact, typically inside of six months. 

There is a real and measurable point of diminishing returns for any sales activity

Following our media mix modeling path, we can not only look at the mix of activity (think reach in media mix modeling terms), but we can also look at the optimal frequency for each activity, e.g. how many calls, meetings, networking events is it going to take to maximize sales outcomes? When do we reach the point of diminishing returns? Knowing this means we can coach the sales team in a very targeted and effective manner.

In the example above for instance we can see that to maximize chances of closing a sale with a particular account a sales rep needs to get to at least two sales meetings and up to six office visits. This will of course vary by industry and market segment. The point is two-fold though: (1) there is a “minimum effective dose” which must be met to be effective, and (2) there is a real and measurable point of diminishing returns beyond which sales effort is probably wasted. 

You can develop high quality expectations for new and retrained sales personnel pertaining to sales production over time

Once you understand the mix and frequency of sales activities required for optimal sales outcomes you can develop a view of the right cadence of activities to develop or retrain reps. Because such a view is grounded in detailed CRM data it can be developed at an individual rep level and tailored to each rep for coaching purposes. It can also be used to set broad expectations at a team level. In our experience using a combination of both yields sales productivity improvements that are typically in the 5% to 20% range.

Data infrastructure and dashboards are key to sales managing themselves and to management managing sales

To enable and implement data-driven selling, businesses will first need to ensure the data infrastructure and necessary data ecosystem to support their CRM systems are in place. Most of our clients start out from a position along the lines of “we don’t have the right data,” or “our data quality isn’t good enough.” These are very common issues facing virtually every business. The truth is, as long as the sales data is being captured somewhere, somehow, applying the right set of ETL processes and techniques will usually yield a perfectly usable data set. Even in cases where the sales activity data is not being captured, that can be remedied through the use of basic auto-population and similar techniques that will quickly build up quality sales activity data over time.

Once the proper data ecosystem and data-driven selling analytics are in place rep-level dashboards can be pushed out to the field for coaching purposes; executive-level dashboards can be produced showing the aggregated mix of sales activities and sales performance at the team level.

Data-driven selling delivers outsized benefits to those who embrace it

Even with a limited investment, businesses who embrace data-driven selling quickly benefit in a number of ways: 

• The on-boarding phase of new recruits can be shortened by as much as 30% with a tighter range of performance across a new cohort, yielding a measurable financial impact on turnover costs
• Tenured team members that are not meeting quota can be retrained in an objective, data-driven manner with actionable pointers on what to do next to improve sales performance. Data-driven coaching can yield 5% to 20% productivity improvements across the team
• Executives can better anticipate sales results and set expectations accordingly. Data-driven selling yields better sales forecast, especially in long sales cycle environments when the forecast provided by the sales team may not be particularly reliable.