Dairy FMCG
Base-case performance gap identified and quantified
Risk-adjusted expected value after conservative and optimistic scenarios
About
Two branches in the company's network served similar markets but performed very differently, and correlational analysis kept pointing leadership at the wrong lever. We used causal — not correlational — analytics, validated across four independent methods, to close a Rs. 76.4 crore performance gap and prioritise branch-level action with real confidence instead of debate.
Industry
Dairy FMCG
Company size
5,000 – 10,000 employees
Founded
1963
The Company
Two branches, one unexplained 74% gap
The company serves daily-replenished categories through a national branch network. Two branches serving broadly comparable markets — call them Branch A and Branch B — showed a 73.8% performance gap, equal to Rs. 76.4 crore in annual revenue, that leadership could see clearly in the numbers but could not explain with confidence.
In daily-replenished categories, small execution gaps in forecast accuracy, stock tracking, route speed, and partner capability compound quickly into large P&L swings, which is exactly what appeared to be happening — the question was which gaps mattered most, and in what order to fix them.
The challenge
Correlational analysis kept misreading the confounders
Standard regression-style analysis on the two branches kept attributing the gap to factors that, on closer inspection, were confounded by seasonality, product mix, and other noise rather than genuine causes. Interventions based on those readings were consistently expensive and consistently failed to move the number.
Leadership needed measured bets rather than another round of debate — a method that could isolate what was actually causing the gap from what merely correlated with it, and put a defensible rupee value against each real driver.
The Solution
A causal map validated across four independent methods
We built a causal map — a "GPS for the business" — of how inventory discipline, channel-partner capability, and customer value mix drive branch revenue, estimating effect sizes, confidence intervals, and significance over an 18-month panel with explicit corrections for seasonality and mix. Econometric modelling, a machine-learning ensemble, direct benchmarking, and a DoWhy/EconML causal network were run independently and converged on the same base-case figure.
The analysis found inventory delivered the single largest controllable gain (Rs. 46.4 Cr), followed by channel-partner capability (Rs. 30.5 Cr) and customer-value mix (Rs. 20 Cr) — and, critically, that fixing them together compounds rather than simply adds: gross effects summed to Rs. 96.9 Cr, and a 0.79 synergy factor produced the Rs. 76.4 Cr risk-realistic base case.
The Results
A sequenced, risk-adjusted playbook instead of another debate
The four-method convergence gave leadership a base-case opportunity of Rs. 76.4 crore and a risk-adjusted expected value of Rs. 71.8 crore — enough headroom, even under conservative assumptions, to justify funding the transformation. A root-cause layer traced Rs. 39.4 crore of the gap to a five-to-seven-year technology lag in planning, tracking, and CRM systems.
The resulting playbook sequenced action deliberately: fix the tech spine first, then tighten inventory, then elevate channel execution, then shift toward higher-value customers — because the synergy factor confirmed that fixing these out of order left real money on the table.
KEEP READING
Explore more studies on AI, analytics, and enterprise intelligence.
