Which product lines and pack formats create the strongest repeat purchase and cross-category lift in Gauteng households, and where can Tiger Brands unlock sustainable volume growth without leaning too heavily on promotions?
Problem statement
Large household brands do not grow evenly across a portfolio. Some categories build repeat purchase, some broaden the household basket, and some respond to promotions in ways that make the margin trade-off harder to ignore. This project tests those differences across bakery, grains and cereals, snacks, and culinary staples in Gauteng.
The analysis is designed for portfolio, category, consumer-insights, and growth roles in FMCG environments where repeat demand, price sensitivity, and private-label pressure all matter at the same time.
Dataset summary
The project uses a synthetic Tiger Brands portfolio built around Albany, Jungle, Bakers, Koo, and All Gold. The data is structured to reflect everyday household purchasing rather than idealised or perfectly clean behaviour.
| File | Rows | Purpose |
|---|---|---|
| transactions.csv | 11800 | Transaction-level portfolio activity across categories, product lines, pack formats, value, repeat purchase, and cross-category behaviour. |
| customers.csv | 3000 | Household context across age band, household size, location, shopping frequency, and loyalty segment. |
| products.csv | 15 | Product-line and pack architecture, including pack prices, costs, and private-label pressure. |
| promotions.csv | 620 | Promotion periods and uplift assumptions by product line and category. |
| category_summary.csv | 4 | Category-level summary of value, repeat purchase, cross-category lift, margin, and promo sensitivity. |
What I analysed
The analysis is structured around category role, product-line performance, pack-size fit, repeat behaviour, cross-category lift, and promotion quality. The goal is not just to see where volume sits, but to understand which parts of the portfolio create stronger household behaviour.
Compared bakery, grains and cereals, snacks, and culinary staples on value, repeat purchase, and margin quality.
Tested how single, family, multi-pack, and value-pack formats align with different household sizes and loyalty levels.
Measured where the portfolio encourages wider category participation instead of one-line purchasing.
Balanced volume uplift against margin compression so promotional support is judged on quality as well as scale.
Key insights
Staples hold the portfolio together
Grains & Cereals is the strongest repeat-purchase category, showing why everyday household demand still anchors the portfolio.
Pack architecture matters
Family packs and value packs are doing more than raising basket value. They are also improving repeat behaviour in larger households.
Snacks broaden the basket
Bakers is the clearest cross-category connector, making it useful for wider household missions.
Promotions need discipline
Promotions are helping volume, but the margin trade-off is visible enough that not every uplift should be scaled.
Gauteng growth is specific
Jungle value pack stands out because it balances repeat strength, basket participation, and commercial quality.
Visual evidence
The visuals below show where category contribution, household fit, cross-category lift, and promotion trade-offs are strongest.
Commercial recommendations
Defend staples where private-label pressure is strongest, especially where repeat purchase is still high enough to justify stronger pack-value communication.
Use household-sized packs more deliberately. Larger households respond best when family and value packs match their shopping rhythm.
Use snacks to widen the household basket, but keep staples as the base that protects repeat demand and category consistency.
Promote with more discipline in categories where uplift comes with visible margin pressure and weaker repeat quality.
Tools used
Python, pandas, numpy, matplotlib, scipy, Jupyter, GitHub Pages.
Synthetic data design, portfolio analytics, exported evidence, and editorial case presentation.