Will Algorithms Replace Human Judgement in the B2B Sales Cycle?

With all the talk about advanced algorithms, artificial intelligence, and chatbots one begins to wonder when virtually every B2C or B2B transaction will be automated.  In this utopian (dystopian?) future, the machines will know exactly what you want, buy it for you, and deliver it to you the same day.  But what role will humans play?  Are we to be disintermediated by the machines?  Replaced by algorithms? Future thinker and researcher Andrew McAfee makes the case that algorithms can and do outperform “experts” that rely on accumulated experience and good old human judgement—but only under certain conditions.

Understanding where you can use advanced algorithms will help you think through where to apply investments in analytics and what complementary skills you need on your marketing and sales teams.

Humans vs. Machines

According to McAfee, there is an abundance of evidence indicating that algorithms outperform human experts in their prediction making prowess.  One research study he cited, which involved the meta-analysis of 136 different studies comparing the prediction accuracy of machine vs. man, showed that in only 8 of the 136 studies the “expert judgments were clearly better than their purely data-driven equivalents.”  He further noted that “Most of these studies took place in messy, complex, real-world environments, not stripped-down laboratory settings.” So, why is this the case? In what situations or conditions do algorithms have the advantage?  And what about human intuition? To answer this, he calls on a bit of theory regarding the ideal conditions for decisions made based on judgment and intuition.  The ideal conditions for human judgment include:

  • an environment that is sufficiently regular to be predictable
  • an opportunity to learn these regularities through prolonged practice

In the medical field, you can find examples of expert judgment that fit the above criteria.  McAfee notes that since human biology changes very slowly, medicine meets the first criteria, but gyrating stock markets certainly don’t.  The second condition benefits from fast and consistent feedback loops that promote learning that can be applied to future decisions.  Anesthesiologists working with dozens of patients experience rapid feedback loops --seeing the effects of their actions -- that help improvement decisions. Where human intuition and judgment tend to break down is in “noisy,” highly variable environments where there are large data sets that aren’t easily interpreted.

Implications for B2B Marketing and Sales

For marketing and sales professionals, there are two aspects of the above analysis that impact that effectiveness of algorithms for lead qualification and selling and perhaps tilt the scales to human decisions:  the availability of complete and accurate data to feed the algorithm, and unknown preferences and biases of real buyers. The quality of the output from an algorithm can be highly dependent upon the veracity of the inputs. If there is missing or incomplete data, or a small sample size is used that skews results, the algorithm could suffer. Additionally, the complexity of the typical B2B sale makes automation with analytics trickier.  In the B2B sale, there is often multiple decision-makers and influencers, they can be less than forthcoming in sharing their intentions, and they face reputational risk when making buying decisions.  All of which requires more handholding, coaching, educating, understanding, and communicating—none of which easily automated.  

Conclusion

If we apply the ideal conditions for intuitive decision-making listed above to the B2B marketing and sales functions, we can see where the line could be drawn between automated algorithms and human decisions.  The first condition describes an “environment that is sufficiently regular to be predictable,” which would apply to a sales process whereby qualified prospects have a consistent set of pain points and requirements.  The second condition is where sales team “learns these regularities” and becomes more adept at educating, positioning, and pricing based on this understanding, resulting in faster sales cycles and higher close rates.

The antithesis of this is random leads that have no consistency and high levels of variability, which limits the intuitive decision-making ability of your sales team in determining what is “right” for the customer.  In this context, the job of analytics and algorithms is to eliminate the randomness by combing through data to find patterns that help identify consistent characteristics of qualified leads and deliver them to sales.  It’s not utopian to imagine machines and humans cooperating to make better decisions, you just need to have them each focus on the task they are best equipped to handle.

Photo credit: MattHurst via Visualhunt.com /  CC BY-SA