Nobody tells you what it feels like to bet on yourself with someone else's business.
In December 2025, I became the operating head of Innoventry Software — a GST billing and inventory platform built for Indian small businesses. The company was 15 years old. It had a real customer base, real revenue, and a real product that people depended on daily.
The founder — family — was transparent with me from the start. He told me multiple times: don't leave your job for this. It's harder than it looks. Make sure you know what you're getting into.
I heard him. I also believed in my father's conviction that something real could be built here, and I trusted my own gut that I could figure it out. So I said yes.
This is what I found.
The headline number and the real number
The business had 1,542 paying customers and ~₹10 lakh in annual recurring revenue. Handed over at zero acquisition cost. For a 24-year-old looking for a real problem to work on, that's an interesting starting point.
But revenue and what you actually take home are two different things. After the founder's share, after salaries, servers, support, and operations — the net profit splits to roughly ~₹15,000 a month between two people.
That's the real number. Not the headline. I'm not sharing it to complain — I knew the unit economics before I signed up. I'm sharing it because anyone thinking about acquiring or inheriting a business should do this math first, in full, before they get excited about the top line.
The customer data told a more interesting story
The software had over 101,000 total licenses on record. That number is technically accurate and practically meaningless — the bulk of it is pre-generated inventory and old bulk orders that were never activated.
The real picture: roughly 8,300 customers who ever actually used the product. 1,542 paying today. Another 684 who let their subscription lapse in the last year — they used the product, they just didn't renew. Over 2,000 more dormant beyond a year.
That gap between 8,300 and 1,542 is not a failure number. It's a roadmap.
The 684 warm lapsed customers are people who had the product, used it, and stopped renewing. They didn't leave because the product failed them — they left because nobody followed up. That's fixable. So that's what we're doing first: calling every single one of them.
The tech stack is exactly what you'd expect from a 15-year-old Windows product
The product runs on Java SWT and JFace — a desktop-first stack from an era before cloud, before mobile, before most things users now consider standard. It is Windows and android - only(working on making it available for MAC as well). There is no real API layer. Subscription renewal reminders were largely manual, handled by a support employee making calls.
That last part is the one that stayed with me. In 2025, a software company's primary retention mechanism was a person picking up the phone. There is no automated reminder. No WhatsApp follow-up. No email sequence.
This is also, strangely, the most tractable problem on the list. Innoventry already supports bulk WhatsApp messaging — I just need to set up Meta's Business API and write the templates. A weekend of work that should have existed years ago.
Why I didn't walk away
I ran a proper analysis on the business. The verdict came back unfavourable — not a good investment by traditional metrics.
But I think "investment" is the wrong frame. This is a learning environment with a live customer base attached. 1,542 real businesses are using this product today to manage their billing and inventory. Their problems are visible, specific, and solvable. That kind of access — to real users with real pain — is something most early-stage founders spend years trying to get.
The other thing that kept me in: the loyalty. A distribution partnership that ended years ago still has 933 of its original customers actively paying. People who have every reason to have moved on have not moved on. That tells me the core job this product does is genuinely valuable — even if the product itself needs rebuilding around it.
What the next 90 days look like
Three things, in order:
Revenue recovery first. Calling all 684 warm lapsed customers. Target is 35–40% conversion. If it works, that's roughly ~₹1.9 lakh in recovered ARR and a meaningful bump to monthly net.
Then automation. Setting up the WhatsApp reminder system so renewals stop depending on manual outreach. This should run itself once it's built.
Then listening. Before writing a single line of new product code, I'm doing 50 churn interviews. I want to understand, in customers' own words, exactly why they leave. Product decisions come after that data exists — not before.
I'm not rebuilding the software yet. I'm not pivoting yet. I'm fixing the basics and listening first.
The parallel track
While this is happening, I'm also building Ravix Studio — a SaaS product house I started in January 2026. The first product is Moneta, a payment follow-up tool for freelancers with recurring clients. The community is at 1,100 members on Discord.
Running both simultaneously is the actual experiment. Can you modernise a legacy business and build something new, with the same hands, at the same time?
I don't know yet. But the founder warned me it would be hard, and I chose to find out anyway.
That feels like the right place to start writing from.
This is the first post in an ongoing series about building — the real version, not the highlight reel. If you're working on something unglamorous and want to compare notes, my DMs are open.