Akanksha Barot · How I can help · Startups and scale-ups
Startups and scale-ups
Nought to one, then one to ten, without breaking the field team on the way.
I have done both stages and they fail differently. AgriCentral went from an empty repository to more than ten million downloads. Jiva’s direct-to-farmer platform went from pilot to fifty thousand active farmer touchpoints across four services. The first stage kills you by building something nobody wanted. The second kills you by scaling something that only worked because twelve very good people were holding it together by hand.
Who this is for
Whether this page is about you.
- Seed to Series B, with a farmer-facing or field-facing product already in market.
- A founder who has an install number they are quietly unsure about.
- A team that has never had a product person who has sat in a farmer’s yard.
What usually goes wrong
Three problems I keep meeting.
- 01
The install number looks fine and nothing else does
AgriCentral passed ten million downloads and the weekly retention curve told a completely different story. That gap is the normal case, not the exception. The work is finding out which cohort came back, in which district, and what they did differently, before the fundraise makes the install number load-bearing.
- 02
The field team is the product and nobody treats it that way
A farmer platform in a low-trust market is delivered by people. If the agents are unplanned, unsequenced and unaccountable, the app is a rounding error. At Jiva, sequencing the day produced most of a roughly 30 percent productivity gain. The map produced far less, which was not what anyone expected.
- 03
Full-time or fractional, decided by cost rather than by stage
Both are defensible and they suit different moments. A team still deciding what the product is needs someone in the field with them for months. A team that knows and is executing needs review, structure and unblocking. Choosing on budget alone tends to buy the wrong one.
What backs it
Every claim names where it came from.
AgriCentral was built from scratch: product vision, UX, content systems and engagement loops, with a 26-person team including agronomists. Google named it among the "Best Apps from India for India".
Olam Information Services (Mindsprint), Head of Product, 2018 to 2021Weekly active usage rose 22 percent from experience-led engagement loops. Field staff productivity rose roughly 30 percent from the visit planning product. I owned both of those numbers.
Jiva AgServices, Indonesia, Product Lead, 2021 to 2025Eighteen people across Product, Design, Engineering, Data, Research and Field Ops, reporting to the Chief Product Officer.
Jiva AgServices (Olam Global Holdings)The year AgriCentral barely grew
AgriCentral launched on 31 December 2018 and adoption was slow for months. The assumption underneath it was that the ground teams would carry acquisition, and nobody had established which farmers in those districts actually had a smartphone. Growth only moved once acquisition went to Facebook, Google and YouTube, targeting farmers who were already on WhatsApp and Facebook. The ten million figure belongs to the period after that correction. I would rather a founder heard this than the version where the number was always going up.
Told in full on The Crop Chronicles, episode 5, and in the Jiva interview. Both linked at the foot of this page.Relevant here
The product streams this segment usually needs.
- Farmer advisory
Agronomy guidance a farmer acts on the same week, in the language and at the reading level they actually have. In AgriCentral that meant seven languages, activity schedules built from the crop and the sowing date, and a question channel answered by agronomists and by other farmers.
Source: AgriCentral, Head of Product 2018–2021, launched 31 December 2018. Jiva, Product Lead 2021–2025. - Agri e-commerce
Input sales to farmers who compare against the dealer they have borrowed from for fifteen years. Price intelligence is the other half: AgriCentral carried more than a hundred crops across roughly two thousand Indian markets.
Source: Jiva agri-ecommerce. AgriCentral Market View. Marketplace GTM across Amazon, Flipkart, eBay, Snapdeal and ShopClues at BoostMySale. - Field agent management
Visit planning, sequencing, coordination and what each agent is accountable for by the end of the day.
Source: Jiva field coordination product, Indonesia. Roughly 30 percent field staff productivity gain. - Leadership reporting
The internal view leadership actually steers on, rather than the deck assembled the night before the board call.
Source: Internal operations tooling at AgriCentral and Jiva. - Weather forecast communication
The forecast is a commodity. Turning it into an instruction someone trusts enough to change a spray date is the product.
Source: A fifteen day outlook inside AgriCentral, and the forecast panel on the Jiva advisory screen.
How it runs
What working together looks like.
- 01
Two weeks, in the field
Not a workshop. Farmer interviews, agent ride-alongs, and a look at whatever your analytics already know and nobody has read.
- 02
One metric, one roadmap
A single number the team is accountable for, with the roadmap arranged behind it and the things that do not serve it removed.
- 03
Ship, then hand back
Working alongside your product and engineering leads, with the goal that the structure survives after I stop being in the standup.
Questions
Asked before, answered straight.
Do you work full time or fractionally?
Both, and the stage should decide. If you are still finding out what the product is, I need to be in the field with your team for a stretch of months. If you know and are executing, a standing two days a week does more than a full-time hire who arrives with no context. I will say which one I think you need before we agree anything.
We already have a product manager. What would you actually add?
Usually two things. Field judgement, which is the difference between a roadmap that reads well and one that survives a district with two bars of signal. And the operation around the product: agent workflows, agronomy content, reporting lines and the dealer relationships. Most farmer-facing startups are under-resourced on the second and do not know it yet.
How do you measure whether it worked?
On a number agreed at the start, and it will not be downloads. At Jiva it was weekly active usage and field staff productivity. At AgriCentral it was retention by cohort, once we stopped celebrating installs. If your board is being shown a number you would not defend in private, that is the first thing to fix.
Our users are in a market you have not worked in. Does the experience transfer?
The constraints transfer and the agronomy does not. Direct sunlight, patchy signal, a shared household phone, several scripts and a user who will not read instructions are the same in Lampung and in Maharashtra. Crop calendars, price behaviour and trust structures are local, so I would expect to spend the first weeks being taught them.
What do you need from us to start?
Access to farmers and field agents rather than to a deck. Whatever analytics exist, however messy. And one person on your side with the authority to change the roadmap, because discovery that cannot change anything is theatre.
What this record is, and what it is not
Nine years of this was done in-house, owning the outcome, at AgriCentral and then at Jiva. It is not a consulting portfolio of twenty logos. For a startup that is usually the right way round: the useful person is the one who has carried a number through a bad quarter, rather than the one who has advised on twenty.
Published elsewhere
- PodcastLearnings from 10M downloads with AgriCentral PM Akanksha BarotThe Crop Chronicles, by Jiva. Episode 5, 31 minutes
How AgriCentral reached ten million downloads, why the first year of adoption was slow, and what transferred to Indonesia.
- InterviewLessons from 10 million downloadsJiva, on Medium
The written version. Acquisition, the seven languages, and the interface decisions that came out of watching farmers hold a phone.