Horse's Mouth

Horse's Mouth was a two-sided marketplace, and the alumni side was the constraint. Students didn't arrive browsing. They came with target universities already in mind, and they wanted to talk to alumni from those exact universities. If we had nobody from a student's shortlist, that student was usually lost. We'd suggest comparable universities, but it isn't the same thing as talking to someone who went where you're trying to go. Alumni were the supply side and students the demand, and I worked both: onboarding the alumni and running the marketing that brought the students in.

We started with the UK because it's the top destination for students from India, Pakistan, Nepal, Bangladesh and Sri Lanka, which was our core market. Once that base existed, the network widened outward from it.

The network reached 3,000+ alumni in total. I was directly responsible for 1,000 of those, across 20+ countries — roughly 600 through two influencer campaigns and the rest through one-to-one outreach before and after them. The biggest markets were the UK, US, New Zealand, Australia, Germany, France, Japan, Canada and Poland, plus a growing Southeast Asia cluster.

The process was research first, outreach second. I'd work through a university's alumni ecosystem and use targeted LinkedIn research to find people whose course and career actually matched what students on our list were asking about, then approach them on LinkedIn. Most didn't reply the first time, and the follow-up is where the conversions came from. We streamlined onboarding so that becoming a mentor took under two minutes. Coverage of the universities people were actually applying to was the real target.

Demand was the other half of the job. I ran content end to end across Instagram Reels, YouTube Shorts and LinkedIn — writing, editing and designing it myself.

The content bucket was made up of mentor focus and user acquisition: Mentor of the Month, session videos of students working through a problem, product utility, and mentors resharing their own reviews on social media.

The two influencer campaigns were the biggest thing I ran. I built the long list, then vetted candidates using influencer tools and their own Instagram analytics: audience make-up, past brand work, whether their videos actually performed. That got us to a shortlist of five or six.

I approached them, negotiated per-reel rates, got contracts signed, then worked directly with each on the creative — what to say, how to say it, what not to claim — through several drafts before anything shipped. The two campaigns did 700,000+ impressions and brought in ₹25L+ worth of leads in the form of users.

We also ran an online community of alumni, which was there to hold alumni retention up. When I joined it was almost dead silent. We built an activation strategy and brought a few members in to help carry it, and made content that was actually useful to mentors — events happening in their countries, student deals and offers on everyday apps. Responsiveness in the community went up, and that turned into more willingness to get sessions scheduled.

A booked session isn't a delivered one. Once a student was booked, both sides had to actually turn up, and they were usually in different time zones.

One of my main responsibilities was making sure the sessions actually happened. We had automatic reminders going to both students and mentors over email and WhatsApp, but the absence rate was still there — either our video calling software failed, or a session needed rescheduling and the communication didn't happen.

I added my own WhatsApp reminders and phone calls to both sides before every session, which cut the no-show rate sharply — and made every booking depend on someone chasing it through reminders, technical faults and rescheduling. At the peak, that was 50 bookings a day moving through that cycle, and it held for more than a week straight.

The harder part was failure. Our video tool was unreliable and often one side or both couldn't get in, so I'd reach both people and either spin up a Google Meet on the spot or find a new slot that worked across their two time zones. That was the actual job — not scheduling, but recovering sessions about to be lost.

5,000 minutes of sessions completed during my nine to ten months; the company's lifetime total was over 10,000. Bookings ran higher than completions, and closing that gap is what this work existed to do. Logged-in users went into the sales funnel and the leads went through to our sales person.

The company was four people — the founder, a tech lead, one salesperson, and me on operations and marketing. On a team that size, you end up in rooms an intern normally doesn't.

I represented Horse's Mouth at an investor meetup at the IIT Roorkee incubator, where we were based, having conversations about our model with several investors. I followed up with a few afterwards and got their read on it, which was more useful than the pitch had been.

For the raise, the founder and I built the deck, and I worked the numbers: CAC, revenue, monthly active users, alumni count and its country spread, runway, unit economics, plus the argument that we had built an unconventional go-to-market in a very conventional market. Our CAC came in well under the far bigger players in it. We also worked up what the whole machine was worth. It was my first actual valuation of a business in real life.

To take it up a level, I wrote the acquisition plan, two of them, because there were two ways in. Plan A was carpet-style coverage of the world's top 50 universities, Ivy League included: pull the current students and alumni for each, then run targeted LinkedIn outreach through automation software. We weren't guessing at the yield, because we already had our own numbers on how many we reached against how many actually onboarded. I also wanted brand ambassadors inside those universities' clubs and societies, responsible for bringing alumni across, so we covered the spread of degrees and specialisations rather than whoever we happened to catch.

Plan B was influencer-led, 80% influencer marketing, 20% one-to-one outreach, with one or two creators per country making videos aimed at students of the universities there. We could identify them almost everywhere outside China and a few Southeast Asian markets, and our own past conversion rates set the onboarding flow rather than an assumption. Both plans went in with unit economics, past data analysis, influencer research, content ideas, and a full cost.

The situation changed, and instead of raising, we were acqui-hired by Yocket, the fifth-largest study-abroad company in India. The co-founder and the tech lead chose to get hired, and by the time the transaction went through I'd already been placed through my college placement that January, so I didn't go across.

This was my first real internship and it's still the most exciting work I've done. The founders gave me a goal and left the method to me, and that's the reason it worked.

I was doing it alongside college. I worked late-night shifts, skipped classes when a campaign needed me, and there were stretches at 15+ hours a day. Over the summer break, when everyone else took two months off, I worked from home and kept the same pace.

What I took from it is specific: give me real responsibility, a process I can see, and some room to run it my way, and I'll beat what both of us expected. That has held since.

When Yocket acquired us they kept the founder and the tech lead. I'd already been placed through my college placement, so I went to Garden instead.