← See All Case Studies
glowlit logo
Data Science, Food and Beverages





An international food & beverages company has an in-house data services team who was tasked with identifying which colmados (small shop owners in South America) were the most likely to manage successful, long-lasting businesses. Given this data, the company could invest in the colmados’ potential for growth in a variety of ways. The data team contacted Q to redesign a 15 question survey for their system that would be sent to thousands of colmado owners. By interpreting their answers, we would be able to accurately measure their respective level of motivation.


Without any previous validation study, we were careful not to construct an overly complicated analytical model for weighting the different questions. Instead, we considered each of the four factors independently and assigned an equal weight to each one in the final model (this model can change in the future when validation data is collected). When we looked at the distribution of answers, we found that by averaging all the four factors, a normally distributed bell curve of colmado owners’ potential emerged, as we expected. As such, the company could move forward and invest in the highest potential colmado owners with confidence.


We conducted an extensive literature review on sales and factors of business success. Through this research, we sought to understand the role of motivation and what additional aspects can predict a business’ future success. 

Surprisingly, we found that motivation is not a good predictor of business success. A lot of businesses are highly motivated, yet they still fail.

Based on the aforementioned literature review, we discovered the four biggest predictors of future success, namely: 

  1. Grit/ Resilience: how you fare when things get tough 
  2. Self Efficacy: the belief that you will be successful 
  3. Past Success: data on past performance
  4. Growth-Focused: how much you are occupied with your businesses’ growth

After presenting our findings, we designed the questionnaire, but were constrained by space. the company’s system called for a maximum of 15 questions, with the final version of the questionnaire narrowed down to merely 10 questions (despite the usual measures for these psychological variables involving as much as 6-8 questions per variable). 

To overcome the constraint for grit, we used survivorship bias to our advantage. Knowing that all colmado owners faced hardship because of COVID-19, we strategically wrote a question to learn how difficult the pandemic was on their business. We could already deduce that businesses who suffered more serious difficulties because of the crisis are likely to have a high level of grit, given that they are still in business. 

Furthermore, since the survey is based on self-report, it’s easy to fall into the trap of social desirability bias: colmado owners might answer questions according to what they think is “right” or what the company would “want to hear.” 

So, we baked the four determinants of success into how we framed the questions and answers, giving colmado owners the opportunity to answer in less optimal ways (sometimes even encouraging them to do so using specific phrasing), so we could better evaluate their business approach. The way colmado owners answered the questionnaire could, in and of itself, provide valuable insight into the colmado owner. 

As a global international CPG company we are proud to work with Q in order to better help our customers thrive. Roy and Haggai helped us rethink our strategy on how to cluster our customers and better assist them run their business and detect ambition and potential! Thank you very much Q team for the dedicated work, you are highly professional and GREAT team-players!

John Doe, CEO of the company's in-house data-services team