Why We Built a Behavioral Simulation Engine for India

Every year, hundreds of FMCG products launch in India. Most fail within eighteen months — not because the product was bad, but because the company misread how Indian consumers actually make decisions.

The problem with traditional research

Survey panels tell you what people say they’ll do. They don’t tell you why — and in India, the why is shaped by forces that Western consumer models weren’t designed to capture.

Consider a simple question: will a middle-class family in Pune buy a premium baby skincare product priced at ₹499?

A survey might tell you 62% said “likely to purchase.” But that number collapses when you account for:

• Aspiration-Permission Tension (APT): The mother aspires to premium care but needs permission from the household’s financial decision-maker — often a mother-in-law or spouse.
• Channel dynamics: Is the product available at the local kirana store, or only on Amazon? Channel access shapes consideration sets differently across SEC segments.
• Household structure: In a joint family, the purchase decision involves more people. A nuclear family in the same SEC bracket behaves differently.

These aren’t edge cases. They’re the default mode of consumer behavior for 1.4 billion people.

What we built

Negenco’s simulation engine models individual Indian consumers across 18 behavioral dimensions — not as averages, but as distinct synthetic agents with internally consistent profiles.

Each agent runs through a four-step evaluation pipeline:

1. Brand Trust Gate — Does the agent trust this brand given their SEC, category involvement, and prior exposure?
2. Benefit Relevance — Does the claimed benefit map to what this agent actually values?
3. Price Evaluation — Is the price acceptable given the agent’s sensitivity, household income, and aspiration level?
4. Innovation Risk Gate — Is the agent willing to try something new in this category?

The key insight: these steps are sequential. An agent who doesn’t trust the brand never gets to evaluate the price. This models real decision-making, where early filters eliminate options before rational cost-benefit analysis kicks in.

Why dimensions, not demographics

Demographics tell you who someone is. Dimensions tell you how they decide.

Two 28-year-old women in SEC A1 Mumbai might have identical demographic profiles but completely different decision patterns. One is a maximizer with high need-for-cognition who researches extensively before buying. The other is a satisficer with high brand trust who buys whatever her preferred brand launches.

Our 18 dimensions capture this variance:

• Person-level (14 dimensions): City, SEC, life stage, household structure, price sensitivity, brand trust, benefit sought, risk tolerance, APT, tradition-self-direction, openness, conscientiousness, maximizer-satisficer, need for cognition
• Contextual (1): Occasion — derived from the product concept itself
• Derived (3): Category involvement, channel orientation, and innovation adoption — computed from the person-level dimensions and the product category

What this means for product teams

Instead of a single purchase intent score, you get a narrative understanding of which consumers accept or reject your concept, why they do, and what you could change to shift the outcome.

The simulation runs in under 30 minutes with 500 synthetic agents. Each agent’s reasoning is fully transparent — you can read their evaluation at every pipeline step.

This isn’t a replacement for real consumer research. It’s a way to test and iterate on concepts before you invest in panels, prototypes, and production runs. Think of it as a behavioral wind tunnel for Indian markets.

---

We’re publishing our validation methodology and results as open research. Follow this blog for updates on our retrospective studies, where we test the simulation engine against real market outcomes for products that have already launched.