| Company | AI-driven alternative data platform, 250+ data sources |
|---|---|
| Buyer | Enterprise, with traction concentrated in banks and NBFCs |
| Role | Product Development and Marketing Executive. Led marketing. |
| Tenure | Nov 2021 to Jan 2023, 15 months, Mumbai |
The company
Data Sutram is an AI-driven platform built on alternative data drawn from more than 250 sources. When I joined, it sold sector-focused data packages that helped companies acquire customers, assess credit risk at postal-code level, choose brick-and-mortar store locations, plan resources, benchmark competitors and inform overall growth strategy.
The founding team was technical, the data asset was real, and there was no marketing function and no predecessor. Nobody had yet been accountable for deciding what the company should argue.
Data Sutram went on to narrow decisively into financial services, raising $3M explicitly to accelerate its BFSI product and later a Series A led by Lightspeed. It now counts HDFC, Axis, ICICI, Kotak, L&T Finance and Tata Capital among its customers. I was there in the phase before that, when the company was still horizontal and the job was to find out which argument actually worked.
The problem as it was handed to me
“We need marketing.” Which in practice meant: we have a powerful data asset, founders who can explain it precisely to other technical people, and no repeatable way to start a conversation with an enterprise that has not already heard of us.
The instinct in that position is to start producing. Content calendar, ad account, posting cadence. There was real money available to do it with, so the expensive mistake was fully within reach: spend hard against the assumption that reach was the constraint, and find out a year later that it never was.
So I did not ask for the full budget up front. I asked for Rs 10 lakh to run three months of testing, on the basis that I did not yet know which channel would work and neither did anyone else.
What I diagnosed instead
1. Six use cases is the same as no use case
That list is a genuine strength as an engineering achievement and a genuine liability as a market position. A company that can help you acquire customers, price credit risk, pick store locations, plan resources and benchmark competitors is describing a capability, not a market. Every one of those is a different buyer, with a different budget line, a different trigger and a different objection.
Nobody buys 250 alternative data sources. They buy one decision they are currently making badly.
2. Reach was not the constraint, and treating it as one would have burned the budget
The addressable enterprise buyer list was countable. You can name every large bank, NBFC and insurer in India on a single page. A marketing function built for volume is the wrong machine entirely when the buyer universe is finite. Every rupee spent on reach was a rupee not spent on credibility.
3. Where the traction was, a veto seat sat in the room
The enterprise pull was concentrated in financial services, and in regulated financial services the person who wants the product and the person who can stop it are different people. Risk and compliance are in the meeting. They are not moved by product benefit. They are moved by evidence that the vendor is safe to be associated with.
Almost everything a startup naturally produces speaks to the enthusiast. None of it speaks to the person who can kill the deal.
The diagnosis: marketing here was not there to generate demand. It was there to narrow one sentence, and then dismantle one objection held by one person in a room I would never be in.
What I built
Language first, because everything else depended on it
The foundational work was refining how a complex, data-heavy platform described itself. Technical founders describe mechanism. Enterprise buyers buy an outcome against a risk they are already accountable for. Rewriting the argument so it led with the decision the buyer was making, rather than the 250 sources underneath it, was the piece that did not exist and the piece nothing else worked without.
The effect showed up first as something unmeasurable and then as something obvious: conversations with large banks and NBFCs started going better with the same team, the same product and the same deck structure.
Events, because that is where the committee actually is
In enterprise financial services the buying committee is reachable in person in a way it is not reachable through a form. Events put the company in front of named institutions in a setting where credibility transfers by association rather than by claim. This became a primary source of qualified pipeline.
Thought leadership, as proof rather than as content
Not a publishing calendar. Material substantial enough that a risk officer could read it and conclude the vendor understood the problem better than they did. In a market this small, being right in public is a sales channel.
PR, social and content, in support
- PR and communications to establish that the company existed and was serious, to people who would never respond to an ad
- Content and SEO mapped to how enterprise buyers actually research vendors, which is fewer terms, higher intent and longer documents than consumer search behaviour
- Social and brand growth as the connective layer that made the thought leadership compound rather than disappear
Test first, then scale ten times over
The budget ran in two phases, deliberately. Rs 10 lakh across the first three months to find out which channels actually produced qualified enterprise conversations, then Rs 10 lakh a month once the answer was clear.
That structure is the reason the numbers hold. A ten-fold step-up in monthly spend is only defensible if you already know what you are buying, and three months of cheap testing is what bought the right to ask for it. It also meant the channels that did not work were killed at trial cost rather than at scale.
And consulting the clients themselves
Part of the role was advising Data Sutram’s own enterprise clients on their marketing strategy, implementation and optimisation. That is unusual and it was useful: it put me on the buyer’s side of the table repeatedly, which is the fastest way to learn what language actually lands in a category you are trying to define.
The pattern worth noticing. The two channels that produced the pipeline, events and thought leadership, are both credibility channels. Neither is a demand generation channel. That is not a coincidence, it is the diagnosis proving itself. In a market with a countable buyer list and a compliance veto, the job is to be trusted, not to be found.
The result
100+ qualified leads into pipeline within the first six months, from a standing start, in a category where the buyer had no existing budget line for the product. Sourced primarily through events and thought leadership.
Total spend across those six months was about Rs 40 lakh, roughly $48,000. That is Rs 10 lakh for the three-month test phase and Rs 10 lakh a month for the three that followed, against 100+ qualified enterprise leads.
Alongside the pipeline: a materially sharper company argument, which showed up as better traction with large enterprise clients, particularly banks and NBFCs, than the same team had been getting before.
What I would do differently
I would have forced the segment decision earlier. We had six credible use cases and I treated that as a range to be communicated rather than a choice to be made, so the first months of positioning work carried more optionality than they needed to. The enterprise pull was concentrated in financial services well before the messaging fully committed to it, and the signal was there to read sooner than I read it.
I would also have reallocated harder after the test phase. The three-month test did its job and told me events and thought leadership were carrying the qualified conversations. But when the budget went to Rs 10 lakh a month, I scaled the whole mix rather than concentrating it, and the supporting channels kept close to their trial weight for longer than the evidence justified. Testing to find the answer and then funding everything anyway is a familiar way to blunt a good decision, and I made it.
Why this travels
The specific market was Indian enterprise, with the traction in BFSI. The shape of the problem is not specific to either.
Any company selling something technical into a committee, where a compliance or security or procurement function holds a veto and the buyer universe is countable rather than infinite, has this problem. So does any platform whose real difficulty is that it can do too many things for too many people. Enterprise software. Regulated categories. Infrastructure. Most B2B AI products being sold right now.
The mistake is nearly always the same: building a demand generation machine for a market that does not need demand generated. It needs one sentence narrowed and one objection dismantled, in front of one specific person.