Overall repeat rate
49.3%
Repeat behaviour stays inside a realistic retail range.
Commercial framing
Which baskets create the most valuable retail behaviour?
The core question is not whether baskets are getting larger. It is whether specific basket compositions and store formats create stronger commercial behaviour through basket value, category mix, and repeat shopping patterns. That is the layer this case study is designed to surface.
Core business question
Which basket compositions and store formats in Gauteng create the highest-value shopping behaviour?
This analysis compares basket archetypes, category combinations, store formats, and promotions across Makro and Game. The outcome is not descriptive reporting. It is a set of ranked decisions about what Massmart should prioritise next.
Decision summary
Massmart priority
Massmart should prioritise mixed baskets because they deliver the highest average value while still carrying strong repeat behaviour.Promote together
The strongest category combinations to promote together are led by the full shop basket because it contributes the most value and penetration in the final model.Makro wins where
Makro outperforms Game on basket size, category diversity, and bulk-led value because shoppers use it for larger stock-up and liquor-linked missions.Game wins where
Game outperforms Makro in more accessible, lower-commitment shopping missions where household and quick-trip behaviour are easier to convert.Basket behaviour
Mixed and stock-up baskets create the strongest value patterns.
Quick trips are still important because they keep traffic moving, but they do not create the same commercial depth as mixed or stock-up baskets. The stronger growth opportunities sit where category breadth and repeat behaviour begin to reinforce one another.
Store formats
Makro and Game do not play the same role in the basket story.
Makro performs better where shoppers are willing to build larger, broader baskets, especially with stronger liquor and mixed-basket weight. Game performs better where basket formation is more accessible, more frequent, and more likely to stay tied to convenience or household needs.
Store-format read
Makro is not just bigger. It is structurally broader.
The final model shows that Makro's edge comes from its ability to hold more categories inside the same trip. Game's advantage is not scale. It is accessibility, especially where the mission is smaller, faster, or more household-led.
Promotion impact
The highest uplift combinations also demand margin discipline.
Promotions help, but not every basket-building offer is equally attractive once margin compression is considered. That is why the ranking matters more than a generic yes-or-no view on promotions.
Decision layer
What Massmart should do next.
This section answers the business problem directly. It moves from pattern recognition into decision-making across basket strategy, category pairing, store-format role, and high-value shopper identification.
Massmart should prioritise
Massmart should prioritise mixed baskets because they deliver the highest average value while still carrying strong repeat behaviour.Promote together
The strongest category combinations to promote together are led by the full shop basket because it contributes the most value and penetration in the final model.Makro outperforms because
Makro outperforms Game on basket size, category diversity, and bulk-led value because shoppers use it for larger stock-up and liquor-linked missions.Game outperforms because
Game outperforms Makro in more accessible, lower-commitment shopping missions where household and quick-trip behaviour are easier to convert.High-value customer signals
The clearest high-value customer signals are mixed baskets, broader category mix, stronger repeat patterns, and basket value above the portfolio median.Promo caution
The highest uplift combinations should be treated carefully because Game / Liquor-led basket also carries visible margin compression risk.Evidence and links
Insights, outputs, and public project links.
The insight cards show the main commercial patterns. The exported visuals support the story, and the public links connect the standalone page, the III.IV case file, and the underlying code and outputs.
01
Mixed baskets carry the strongest value pool
Mixed Basket has the highest average basket value at R493.64.
02
Repeat behaviour is strongest in planned shopping missions
Mixed Basket has the highest repeat-shopper rate at 64.1%.
03
Makro wins on scale and basket breadth
Makro leads on average basket size (12.8 items) and category diversity (2.80).
04
Game wins where baskets are more accessible and frequent
Game trails Makro on basket scale, but remains more exposed to smaller convenience-oriented missions and household crossover.
05
Category combinations matter more than single-category pushes
Full shop basket has the highest penetration and value contribution in the final model.
06
The best promotion opportunities also carry margin risk
Game and the liquor-led basket show the strongest uplift in the final model at 67.3%.
Project links
Explore the work from summary to source.
The standalone page presents the commercial narrative. The III.IV case file keeps the structured portfolio write-up, and the GitHub repository contains the dataset generator, notebook, and exported decision tables.
