Why this role is different

Most analysts we speak to have the same story. They built the model, found the leak, wrote the deck. Someone said "great work, let's pick this up next quarter." It went into a folder. A year later they still know exactly what is wrong with the business and have no lever to fix it.

This role sits on the other side of that gap. You report into the Founder's Office. Your first two interviews are with the founder and the CTO, not with a panel three levels below them. When you find something, you say it to the people who can act, and it moves that week. Your analysis is not an input into the decision. Usually it is the decision.

The data you would be working with

We are a digital first agri company. We sell direct to retailers across rural India with no dealer or distributor layer in between, and we run apps and web for both farmers and retailers. Telecalling sits on top of the digital channels, field sales supports on the ground.

That means we hold first party behavioural data on both sides of the market: the farmer discovering a crop problem, and the retailer who eventually sells them the solution. Very few companies in Indian agri have that, and almost nobody has it digitally, at scale, without a distributor in the middle owning the customer and telling you nothing.

The open questions are the actual job:

- Does farmer side demand on our app predict retailer side ordering in the same pincode, and how many days ahead? If it does, that is a forecasting and inventory engine nobody in this category has built.
- Where is margin leaking between order and delivery, and which SKUs, geographies and channels quietly subsidise the rest?
- What is the true incremental contribution of telecalling? How much of what it closes would have closed anyway through the app?
- Three channels claim credit for the same retailer. App, telecalling, field. What is each worth per rupee spent?
- Our apps generate real product analytics. DAU and MAU, funnel drop off, checkout adoption, AOV, repeat behaviour, cohort retention. Nobody is systematically converting that into product decisions.
- Why does RTO vary so sharply by geography, agent and product, and what is it costing us?
- What is a retailer worth over three years when we own the relationship end to end?
- Nine sales teams, different structures,



different incentives, and the same metric defined differently in each. Which are actually creating value?
- What do we spend on data infrastructure per order, and which queries, tables and tools drive most of it? Nobody here can answer that today.

None of this data is clean. All of it matters. Almost none of it has had a serious analytical mind applied to it yet. What you own The function. You build it rather than inherit it. Charter, team, hiring, structure, standards, and how analytics engages the rest of the company.

The data layer. With engineering: modeling, warehouse architecture, pipeline reliability and metric governance, so "active retailer" means one thing whether it comes from the app, the telecalling floor or the field. Today it does not.

Instrumentation. Event tracking across app and web, done properly and owned by you. If the events are wrong, every number downstream is wrong, and right now nobody owns that layer.

The two sided view. Nobody here has properly connected farmer behaviour to retailer behaviour to revenue. That link is probably the highest leverage unbuilt thing in this company.

Profitability. Unit economics per order, per retailer, per channel, per SKU. Finding and closing margin leakage is standing work, not a project.

The cost line. Our tech and data spend has grown faster than anyone has audited it. Database cost, compute, storage, third party tools, redundant pipelines, unoptimized queries nobody was watching. Treat infrastructure spend as an analytics problem, because it is one. Instrument it, attribute every rupee to a system and a team, forecast it, and work with engineering to cut the waste. Done well, this part of the role pays for itself in year one.

The AI layer. Natural language querying over our own data, agentic workflows, anomaly detection that flags a problem before the month closes, reporting that automates itself.



Built as infrastructure by someone who has already shipped things like this, not piloted by someone who has read about them.

The truth. You become the person leadership trusts when the room disagrees about a number. That is real authority here, and we intend to protect it.

Who we are looking for

- 8+ years in analytics, at least 2 leading a team, ready to run a department rather than a pod.
- Real consumer scale behind you. Gaming, quick commerce, food delivery, travel, fintech, consumer internet. Millions of users and events. If you have worked on DAU, retention cohorts, monetisation funnels and LTV where the numbers were genuinely large, you will feel at home here.
- Robust quantitatively, and this is the hard filter. Statistics, probability, causal inference, experiment design, forecasting. Dashboard fluency is table stakes, not the differentiator.
- Deep database and systems knowledge. Advanced SQL, query optimization, schema and warehouse or data lake design, relational and NoSQL, and a real feel for how data moves through production systems.
- A cost instinct. You have forecast infrastructure spend, traced a bill back to specific queries and schema decisions, and cut it materially without breaking anything.
- Python at working depth, plus the modern stack: dbt, Airflow, BigQuery or Snowflake or Redshift, Looker or Power BI or Metabase or Redash.
- Genuinely AI native. We will ask what you built with AI in the last 90 days, and we would rather see a screen share than hear a summary.
- You speak to founders, not to a data team. Decision first, no forty slide preamble.

Useful but not required: agritech, D2C or rural exposure, product analytics on a consumer app, having built an analytics function or a data lake from zero, and public work that shows the quantitative edge. Who this is not for

If you want a mature stack, clean documentation and a tightly defined scope, this will frustrate you. Some of what you find will be broken, and some questions have never been asked here before. If that sounds exhausting, skip this one. If that is the interesting part, we should talk.

Compensation

Benchmarked for a department head, not a senior analyst. For a candidate who is genuinely exceptional, compensation will not be the reason this conversation ends.

📌 Business Analytics Lead (Bhopal)
🏢 Katyayani Organics
📍 Bhopal