Critical success factors of (AI assisted) Win-loss analysis

While AI usage in B2B sales is becoming commodity, the critical success factors for Win-Loss analysis are becoming even more critical. In this blog, I highlight two critical success factors, which I already discovered and published 6 years ago at Global Sales Science Institute (links to full article below).

Warning: If these two success factors are not considered, there's a high risk that we create a massive "Garbage in, garbage out" system for business critical decision making. 

Critical success factor #1: Data integrity

Data integrity needs to be high, in order to assure correct interpretation of data and corrective actions for improving competitiveness. However, analysis of Win Rate provides massive new insight to most companies, so one should not use poor data quality as an excuse for not starting today. Information quality has been recognized as a potential contributor in achieving strategic advantage over competitors (Baskarada & Koronios, 2014) and this applies also to B2B sales.

Root cause analysis capability and skills are especially important in evaluating the reasons for winning or losing tenders. While Lean Six Sigma (L6S) methods provide an extensive toolbox for high quality root cause analysis and continuous process improvement, starting with basic tools is already a huge improvement in an instinct based sales environment. 

Opinion based data, i.e. reasons for losing/winning/satisfaction, is often thought and entered in a rush (regardless of whether it comes from customer or sales personnel). Thus, the results also partially represent symptoms rather than real root causes. There are a couple of easy, yet powerful, root cause analysis methods which are recommended, such as standardizing the data structure, cause-effect diagram (ASQ 2020a) and the 5 why method (ASQ 2020b).

=> It is highly recommended that the input of Win Loss analysis data is done by human.

Critical success factor #2: data confidence level and quantity

Data confidence level and quantity is another critical success factor. The more data we have, the more we can deep dive to different types of cases and identify strengths and weaknesses. It is commonly assumed that AI requires massive amounts of data. This, however, is not necessarily the case in B2B sales when applying IoC. Pestorius captures the essential point in his book (2007): “In a World where 20% margins and 10% growth is considered successful, making critical decisions with 50-60% certainty, rather than 0% is an enormous and profitable improvement”.

This contradicts heavily with typically used 95% (P<0.05) confidence levels for correlation and regression testing as well as “Six Sigma” level of defect ratio where only 3.4 defects in a Million samples are allowed.

To make this concrete, who would even dream to win 9,999,966 tenders out of a million attempts?  My analysis shows that a 95% confidence level for decision making is reached with ~80 cases. And moderate (50%) confidence already at ~30 cases. Even though one can start with only ~30 won/lost tenders, the full power of IoC comes when the volumes are high. Not only large volume allows more focused and accurate corrective actions, but also totally new applications and use-cases for the competitiveness data.

=> It is highly recommended that prior to launching and Win Loss analysis results, there are at least 30 won/lost cases behind the analysis.

Source and contact for more datails

Leijala A., 2020, Disruption in B2B sales, Internet of Competition (IoC) model changes the way we sell, Global Sales Science Institute (GSSI) publications.

Antti Leijala, https://www.linkedin.com/in/anttileijala/, https://www.ultraleansales.com/en