Stone: recruiting high-performance salespeople for increased revenue


A Stone Payments was born within a market known for archaic and inefficient practices; and came to this market with the promise of "being the option that the current payments market does not offer, with the goal o empowering the Brazilian entrepreneur". In this context, the dream of its founders was created to provide a different business model to the Brazilian entrepreneur.


Founded in 2012 by André Street and Eduardo Pontes, Stone has ventured into the payment market, dominated by giant and consolidated companies. With a focus on the customer, the company promised to subvert the market by offering lower rates and superior service to its customers.


Combining technology with a team engaged driven towards growing and improving the business of its customers, the company experienced rapid growth and in just 6 years it reached an IPO in 2018; becoming the largest independent player in the Brazilian payment service provider sector.


A The startup became a quick success, and faced great challenges to gain market share, including the acquisition of another payment service provider twice its size. Always concerned with people management and striving to build a strong and resilient culture, the company started an expansion plan in 2016, growing its team 15 times and almost 1,500 times its customer portfolio in one year, becoming the fourth largest payment service provider..




In order to fulfill the expansion plan, one of the most critical areas for the company was the commercial area. Stone needed to ensure that salespeople performed wel, selecting candidates with the potential to deliver consistent results.



Using Mindsight tests and Machine learning, a tailored predictive algorithm was built based on the high- performance professionals in Stone's commercial department.


The algorithm was used to identify the characteristic patterns of the salespeople that generated the most results. Based on this data and knowledge gathering system, the software would be able to indicate among the candidates in the selection processes, which ones had the potential to sell more.



With such strong business growth occurring, Stone Payments needed to optimize its sales staff to keep up with this robust development. At that time, there were 600 salespeople in the company, of which 30% were considered to be high performers, according to the evaluation criteria of the area.


To build the algorithm, all salespeople performed Mindsight tests. Machine learning requires that a sample of people be provided to the algorithm, so that it learns the performance standards. Then, the rest of the vendors were used to test whether the algorithm was working.


With the artificial intelligence algorithm in place, the test was then performed on the remaining sample. Among the people that the system indicated most likely to have high performers, 45% were, in fact, high performers. That is, using the tests in the screening, Stone would have 15% more high-performance finalists.


Based on the results of the professionals and the sales model of the organization, it was projected that these people would reach ai increase of 10% in sales quantity and 6% in topline revenue..


Even with the promising forecasts, Stone was experiencing a very intense hiring moment, and it was still necessary to test whether Mindsight's algorithm would work during the day-to-day operations. After all, Stone was our first customer. Therefore, it was not mandatory for recruiters to use the indication of the algorithm in hiring.


9 months after the artificial intelligence (AI) implementation, the team evaluated the variation in performance between the people who entered with the indication of the algorithm and those who were hired without this AI filter.


It was then found that the people who were indicated by the algorithm reached an average of 16% more in sales quantity than those not indicated and still generated a 5.2% higher revenue. An even better result than initially projected.


Due to the models superior forecast capacity versus traditional methods, the custom algorithm started to be generally adopted by (Stone) recruiters to indicate more high-performance professionals and to support the company's rapid growth path.