When I left IndiaMART in May 2024, most people reacted with some version of disbelief. Why would you leave something so stable? And more than once, from people who knew the company well, how did DA even let you go?
What almost nobody knew is that it was my fourth attempt. The first one was back in January 2020.
Four years, four conversations. So if you're trying to work out how to know when to leave a stable job, the useful part of my story isn't the exit at all. It's why the first three didn't happen.
Key takeaways
- Wanting to leave and being ready to leave are two different things. Mine were four years apart, and I don't regret the gap any more.
- Something stops you each time, and it's usually one of two things. Circumstance, or readiness. Knowing which one you're facing changes what you should do next.
- A bounded experiment beats an irreversible leap. "Six months is enough to know" turned an impossible decision into a manageable one.
January 2020: a thought with no shape
The first time, it was really only an instinct. I told Dinesh Agarwal (DA), founder and CEO of IndiaMART, and across fourteen years my boss, mentor and toughest critic, that I was thinking of building something of my own.
He didn't react much. He gave me one piece of advice.
Whenever you start anything, start with a service, not a product. A service gives you cash flow from day one. Then you build the product gradually, the way IndiaMART did.
So I went away and thought hard about what that business could actually be. And I couldn't figure it out. So I dropped the idea myself.
Nobody talked me out of it. I simply ran out of answer.
April 2022: a decision without a destination
The second time I was far more certain. I told DA I wanted to leave, and I meant it. He asked why. I said I wanted to do a startup.
That conversation ran for a month. And what he kept coming back to was one question I couldn't answer.
If you don't know what you want to start right now, why do you want to leave now? First think of an idea. Let me know when you are ready.
He told me plainly to rethink it. He wasn't protecting his org chart. He was pointing straight at a hole in my plan that I'd been very carefully not looking at. So I pulled the decision back.
May 2023: the three-month test that said no
A year later, in May 2023, I came back with something specific. I wanted to build an insurance startup, an app to help people manage their insurance policies.
This time the answer was different. Try it alongside your job for three months, he said, and see how far you get. Meanwhile we'll start looking for someone to take over your role.
Notice that second sentence. He was willing to actually begin replacing me. This was not a retention tactic dressed up as encouragement.
Two months in I had my answer, and it wasn't the one I wanted. Being a startup founder demands an enormous amount of time. With two small children at home, I just couldn't give it what it needed.
So I went back to him myself, my decision and not his, and said I'd delay until my kids were older. And I stayed.
But here's the thing. The experiment had worked exactly as it was supposed to. It gave me a clear no before it cost me my career.
How to know when to leave a stable job: the two blockers
Looking at those three attempts together, what was missing was never courage, or money, or permission. It was an answer to a very simple question. To do what, exactly?
They failed in two different ways, and the difference really matters.
- No destination (2020 and 2022). I wanted out more than I wanted anything specific. If you can't describe the work you'd go and do in one sentence, and if that sentence needs the word "something" in it, this is where you are.
- A destination that didn't fit my life (May 2023). The idea was real. The market was real. It just needed a version of my week that didn't exist, not with two young kids in it.
The first one needs thinking. The second one needs testing, because no amount of reflection would have told me what two months of actually trying told me straight away.
Most of what we call fear of leaving is one of these two, wearing a disguise. And neither of them gets solved by waiting for a braver morning.
March 2024: the fourth conversation
By then I knew what I wanted. Not a startup. Coaching, full time.
I told DA. And this time he said:
You've expressed your desire to build something of your own so many times. Go attempt it. Don't wait too long, six months is enough to know if you're meant for it. And if you ever feel otherwise, just tell me.
He also pointed out, quite accurately, that this was the third time I'd come to him saying I wanted to do something different and leave. Go ahead and do it.
That wasn't permission. Permission would have been "fine, go". This was a man who'd spent four years testing whether I was ready, finally telling me that I was, and that the door behind me would stay open.
That's faith. And faith was the only thing that had ever actually been missing.
Was it hard for him too? He didn't express much. That's simply how he is, and after fourteen years I didn't need him to.
What the six months actually looked like
Here's the part I couldn't have predicted.
Almost immediately after leaving, I picked up two technology advisory assignments with startups. That's how the journey actually began. Not with coaching at all, but with advisory work that paid from the very first month.
The coaching assignments followed soon after. And then there was no looking back. I never came close to using DA's "just tell me" clause.
It took me a while to notice what had happened. Start with a service, not a product. Cash flow from day one, then build gradually. That was the advice he gave me back in January 2020, when I had nothing but a vague itch and no idea what to do with it.
Four years later I followed it without even realising I was following it. The advisory work was the service. It funded the runway while the coaching practice grew into the thing I actually wanted to build.
What to do on Monday
- Write the sentence. One line describing what you'd go and do. If it contains the word "something", you've found your real blocker, and it isn't the job.
- Design a three-month test you can run while still employed. Not research. Real attempts at the actual work, with real people who can say no to you.
- Decide in advance what a negative result looks like, and give yourself full permission to accept one. Mine saved me from a bad exit.
- Find your service. Something you can sell from month one that funds the thing you actually want to build.
- Tell one senior person who'll be honest with you. Not to get approval. To find out whether somebody who knows your work rates your odds.
Frequently asked questions
How do I know if it's the job or me?
Ask yourself whether a better version of the same job, different scope, different manager, more autonomy, would fix it. If yes, you have a role problem, and changing companies is far cheaper than changing careers. If no version of this work interests you any more, that's completely different information.
Should I tell my boss I'm thinking of leaving?
I told mine four times over four years and it worked out. But that rested on years of trust built long before the first conversation. If you can't predict how the news would land, you already have useful information, and it's worth sitting with what that tells you about where you work.