We approached a major fast moving consumer goods company (FMCG) operating in Zimbabwe with the intention to help them understand why some of their retail outlets performed above expectation while others performed below expectations. The aim was to identify controllable factors and non-controllable factors that influence the performance of each outlet. The identification of controllable factors would then lead to strategies around manipulating those factors such that each retail outlet’s performance is optimised. More specifically, we sought to answer the following questions:
- Does the diversity of product lines in each outlet influence sales performance? Some outlets had one product line while others had over a thousand product lines.
- Does location have an impact on the sales of each retail outlet?
- Does the presence of certain employees (sales representatives, merchandisers, merchandising supervisors) have an impact on the sales of each retail outlet?
- Does the psychometric profile of employees influence the performance of their respective retail outlets?
- Does the brand name of each retail outlet attract more (or less) customers leading to differences in sales?
Answers to the questions above would lead to informed decisions when deploying merchandise, taking into consideration the effects of location, product diversity, presence of and profile of human capital as well as whether or not some brands of outlets should have more attention than others.
How we intervened
We worked together with the retailer’s Human Resources Department and gathered data to incorporate in our analysis. The initial stage of our analysis involved building a descriptive analytics model using regression techniques to identify the major drivers of the differences in sales performance of each outlet. The second stage involved interrogating further the factors identified by the descriptive analytics model as contributing more to the differences in sales performance. The third stage involved matching insights with appropriate recommendations aimed at improving the retailers’ overall performance.
Using our descriptive analytics model, our initial findings were as follows:
|Factor that drives sales||Does the identified factor significantly impact sales of each outlet?||Percentage of variation in sales explained by the identified factor||Impact of the identified factor on sales of each outlet|
|Product diversity||Yes||47%||If the number of product lines per outlet changes by X units, sales change by 1.30X%|
|Brand name of retail outlet||Yes||26%||There are differences in the average sales of selected outlets with different brand names.|
|Number of Sales Representatives||Yes||7%||Increasing the number of sales reps by one improves sales by 7%|
|Number of Merchandisers||Yes||5%||Increasing merchandiser presence by one increases sales by 39%|
|Location of Retail Outlet||No||–||The impact of location of outlet is inconclusive|
|Supervision of each outlet||No||–||The impact of the differences in supervision of outlets is inconclusive|
As seen in the table above, our descriptive analytics model managed to explain approximately 86% of the variation in sales. In other words, the four factors that we identified as significantly impacting sales performance of each retail outlet can be used accurately to explain why some outlets meet expectations while some fall below expectation.
Manipulating the presence of merchandisers to improve profit
Although the differences in merchandiser presence in each outlet explain only 5% of the differences in sales performance, as suggested by our model, changing the presence of merchandisers from one to two in an outlet has the greatest impact on sales when compared with other controllable factors. This implies that the FMCG Company stood to benefit from increasing the presence of merchandisers in all its outlets. Merchandiser presence could be improved by either rotating the available merchandisers such that each outlet has enough merchandiser presence or by recruiting externally if funds permitted.
Manipulating product diversity and the number of sales representatives per outlet
Product diversity turned out to be the factor that explains most of the variation in outlet performance. This retailer does not own shops. It distributes merchandise through contractual agreements on shelf space with wholesalers, supermarkets and other shop owners. Some outlets had only one product line whereas others had more than a thousand. Outlets that had more product lines consistently outperformed outlets with relatively less product diversity. It says in the table that if the number of product lines changes by X units, sales change by 1.3X%. This implies that, for example, improving product diversity by one more product line improves an outlet’s sales by 1.3%. This is almost insignificant. However, when ten new product lines are introduced in an outlet, this would translate into an increase in sales performance of 13% which is quite significant. For this particular retailer, product diversity is linked to the number of sales representatives who frequent each shop. Therefore, the retailer needed to do a cost-benefit analysis since introducing more product lines meant increasing sales representative presence which means more wages or overtime.
Taking advantage of the different brand names the outlets belong to
Among all the outlets, some outlets had better average sales because of the brand names they belong to. The implication is that when introducing new products, having taken all other things into consideration (e.g. the socioeconomic class of the outlet’s clientele), the retailer would consider trying brand names that consistently outsell the other brand names to ensure adequate market penetration.
Not focusing on the wrong things
Our analysis revealed that outlets under different Merchandising Supervisors had no significant differences in sales performance. Such a scenario is very unlikely considering that equality in performance is difficult to achieve when we have a number of employees. The expectation is that some employees excel while others fall behind. This result suggests that the way the role of the Merchandising Supervisor was structured was not contributing to the business. Also, any form of remuneration for these supervisors that is linked to the differences in their respective outlets’ performance was erroneous.
Broadly speaking, average performance of outlets in urban areas could not be differentiated from the average performance of outlets in peri-urban and rural areas. This means that, when deploying merchandisers, sales reps and merchandise, all outlets would have to be treated equally, without preferential treatment driven by location.
The Second Phase
Having identified that manipulating the presence of merchandisers in an outlet greatly improves sales, we sought to identify the attributes of the ideal merchandiser. Should the FMCG Company choose to recruit, it would not only have the right numbers but also merchandisers’ with the ideal profile. We sampled merchandisers from selected outlets and invited them to come for psychometric profiling. This allowed us to measure cognitive abilities and personality traits of each merchandiser. We also gathered demographic information (e.g. Age, Marital Status etc.) on all tested merchandisers to establish if there is a link between selected demographic factors and the performance of the merchandisers’ respective outlets.
We identified three personality traits which, in the right combination, explain 20% of the variation in performance of each outlet. In other words, the FMCG Company now knows the personality traits of its merchandisers that are associated with higher performance.
The three personality traits are defined as follows:
- Striving for social acceptance describes a person’s striving for social status and recognition. The higher the percentile rank, the more important it is for the person to feel recognized socially and to behave in a socially acceptable way.
- Aspiration level – This dimension indicates whether the respondent tends to set realistic or unrealistic goals (e.g. with regard to the quantity of work that can be performed or the accuracy and thoroughness that is being aimed at). Achievement-motivated people set themselves realistic but ambitious goals that somewhat exceed their previously achieved performance. Goals of this type are indicated by medium to high percentile ranks.
- Extroversion – The merchandiser enjoys being around people, would rather take charge than be reserved, always thinks the best
The insight above is invaluable when the company wants to recruit new merchandisers or take merchandisers that are already in the system through training.
Instead of linking psychometric profiles to the performance of outlets, it would have been ideal to link each psychometric profile to each merchandiser’s performance appraisal score. However, at the time, the retailer’s performance appraisal system for merchandisers was not reflecting true merchandiser performance. To monitor if the identified psychometric attributes are indeed proving to be vital in the company’s performance, we worked hand in hand with the company’s Human Resources Manager and Merchandising Manager to draft a new performance appraisal system that would not only accurately measure merchandiser performance, but also improve the contribution of Merchandising Supervisors to the business. The objective is to link performance appraisal scores from this new system to each merchandiser’s psychometric profile.
By following the stages of this project, we identified critical factors that are crucial in the performance of this FMCG Company’s outlets through the use of analytics methods. We also identified factors that did not have significant contributions to differences in performance. Combining all these insights provides an invaluable guide in relation to which factors the company strategy should and should not focus on.
We went on to interrogate further, a critical factor (merchandisers) that determines sales performance. The validity of the use of psychometric tests for recruitment and development purposes was confirmed. We conclude that basing your organisation’s strategy on scientific methods and moving on to validate the insights that subsequently come from this approach is of paramount importance.