Case Studies
1. How Repositioning ZET's Core Message
Dropped CAC by 31%
Before
.jpg)
After
The Problem
When I joined ZET, the performance marketing strategy was built on a reasonable assumption: Indians love rewards, so lead with rewards. Every ad, every creative, every influencer brief pushed the ZET FD Credit Card as a cashback and
benefits product.
The problem was that this was the same thing every other card in the market was saying. We were competing on a dimension where we had no real advantage, in a space already crowded with bigger brands, bigger budgets, and bigger reward programmes. CAC sat at roughly ₹2,500 per conversion, and there was no clear path to bringing it down.
The Insight That Changed Everything
The shift came from an influencer campaign that outperformed everything else we had run. The creator had spoken directly to people who had been rejected for a credit card because of a low credit score, and framed the ZET card not as a rewards product but as an FD-backed credit card that helps build a 750+ credit score.
That framing unlocked something. The audience was not aspirational credit card users. They were people who had already tried and been turned away, and were looking for a genuine solution, not another cashback offer. The FD was not a limitation to explain away. It was the product's actual value: a savings tool that gives you access to credit you could not get otherwise.
What We Did
We rebuilt the messaging architecture from the ground up around three themes: loan rejection due to low credit scores, 100% approval guaranteed, and the FD as a smart savings instrument. Every performance creative across Meta and Google was reoriented around this.
We also moved to a video-first creative strategy, with significant investment in the first five to six seconds of each ad. Hooks were written to stop the scroll immediately, specifically for people who had experienced credit rejection. We went further by producing regional language video ads tailored to audience-specific themes, making the messaging feel local and relevant rather than generic.
The Result
Within three months, CAC dropped from ₹2,500 to ₹1,730, a 31% reduction, without a significant increase in spend. The influencer channel, now briefed with the same messaging framework, began generating approximately 2,000 secured credit card conversions per month at ₹1,400 CAC.
The bigger lesson was that message-market fit is not about finding the most attractive claim. It is about finding the problem your audience is actually living with, and making it clear you have the only solution that works for them.
2. How Data-Led Thought Leadership Took Onsurity's Media Share of Voice from 3% to 50%+
The Problem
When I joined Onsurity, the brand had minimal media presence. Share of voice in a category dominated by legacy insurers and better-funded healthtech/insurtech players sat at around 2 to 3%. The sales team was doing the work, but the brand was invisible, which was a real liability in B2B sales. SME founders and HR decision-makers make high-trust buying decisions, and trust is built before a sales call ever happens. We had no earned credibility to do that work.
The PR brief I inherited was reactive: respond to journalist queries, send out press releases when there was funding or a product launch, and hope for coverage. The result was sporadic, shallow, and transactional.
The Strategy
I made a decision early on that we would stop competing for attention through announcements and start competing by being genuinely useful to journalists. The insurance and employee wellness beat in India was data-starved. Journalists covering HR, workplace health, and SME business had plenty of opinions but very little original research to cite.
That was the gap we could own.
The plan was to build Onsurity's brand on proprietary data, not press releases. If we could consistently produce research that answered questions journalists were already trying to answer, we would stop being a source they called occasionally and start being a source they cited without being asked.
What We Did
We designed and published two major whitepapers backed by primary research. The first, "Are Indian Employees Unlocking the Full Potential of Their Healthcare Benefits?", surveyed 1,150 individuals pan-India and surfaced numbers that were genuinely alarming: 83% of employees were completely unaware of the healthcare benefits their employer offered, 60% had no idea who their benefits provider even was, and less than 1 in 10 had used their employer plan to file a claim. These were not soft insights. They were the kind of statistics that write themselves into headlines.
The second, conducted with the Knowledge Chamber of Commerce and Industry, found that 43% of Indian tech professionals experience work-related health concerns, and received coverage in the Economic Times Health and Outlook Business, among other media outlets.
But the turning point was a third study: Burying the Burnout: Decoding the Health Challenges of India’s Tech Geniuses. Its timing was not accidental. Narayan Murthy's comments on 70-hour work weeks had ignited a national debate about overwork in India's tech sector, and every journalist and columnist covering the story needed data. Our report gave them exactly that. It began getting cited organically, without any outreach, in stories about workplace burnout, tech sector health, and the employer's role in employee wellbeing.
Alongside the research programme, we sharpened the company's overall narrative, built a consistent founder and leadership voice for IRDAI policy announcements, and placed Onsurity's leadership at relevant industry events as speakers.
The Result
Within six months of continuous execution, Onsurity's media share of voice crossed 50%, from a base of 2 to 3%. Coverage shifted from occasional press release pickups to proactive citations by journalists who had come to see Onsurity as the authority on employee health and workplace wellness in India.
The work was recognised with a Gold Title for Content Leadership at the Inkspell Media Awards in 2023 and Best Community Engagement Campaign PR at Adgully in 2024. The underlying insight was simple: in trust-sensitive B2B categories, the brand that teaches the market wins the market.
3. How We Built an AI-Powered Content Engine at ZET That Doubled Creative Output
The Problem
At ZET, we were running a performance marketing and product communication operation that was bottlenecked entirely by human output. Three designers could produce around four static performance creatives with 4 adaptations each in a single day. For automated user journey content, which included push notifications, WhatsApp messages, and in-app banners, we could close content for a maximum of two funnel stages per day between two writers. The team was not underperforming.
The process itself was the ceiling.
This was a real problem in a category where speed of testing determines speed of learning. We needed to know which message worked for which audience at which stage of the funnel, and we needed to know it fast. The pace at which we were creating and deploying content made that kind of iteration impossible.
What We Built
The first thing I did was standardise what we knew. We documented our brand voice, product CVPs, funnel architecture, conversion triggers at each stage, and the historical data on which creatives had performed best and why. This became the knowledge base.
We then set up a Custom GPT, fed it the full creative library alongside brand guidelines, tone references, writing style, product messaging, and funnel logic. Critically, we also fed it performance data, telling it which ad angles, hooks, and message structures had driven the lowest CAC and highest conversion. The system was not starting from scratch each time. It was building on what had already proven to work.
The brief format we used was standardised: channel, funnel stage, audience segment, objective, key message, and any creative constraints. When a brief came in with this structure, the GPT could produce multiple copy variants, each mapped to a different hook or angle. The output covered performance ad copies, product journey content across push notifications, WhatsApp, and in-app banners, as well as decks, flyers, newsletters, and video scripts.
Later, we extended the same setup using Claude, which we found particularly effective for longer-form content and product communication that required tonal consistency across multiple touchpoints. We also found Claude to be more accurate with regional content.
We also built a separate workflow using ElevenLabs and Envato to produce product explainer videos across multiple languages. These videos significantly reduced inbound queries to the customer support team by answering common product questions before users needed to ask them. The team adopted this not as a replacement for their judgment, but as a way to move faster. Designers who had previously been producing static assets only began using AI video tools, expanding their own scope of work. The up-skilling happened organically because the tools were embedded into the actual workflow, not introduced as a training exercise.
The Result
Weekly creative output went from 35 to 70+, without adding headcount or compromising on the quality. The more important shift was in the testing cycle: we could now try more angles, fail on the ones that did not work, and double down on the ones that did, all within the same week. This directly contributed to the CAC improvement documented separately, where we brought acquisition cost down from approximately Rs. 2,500 to Rs. 1,730.
The broader outcome was a team that became genuinely more capable. Graphic designers were working on video. Stakeholders were writing better briefs because they had to think in structured inputs. And the organisation had a repeatable, scalable content infrastructure that did not depend on hiring more people or onboard external agencies every time volume needed to go up.
The principle behind all of it was straightforward: AI does not replace creative judgment. It removes the operational drag that prevents good judgment from being tested at scale.