Cofounder : “Let’s shut down Augment. We are launching Clusto and we cannot maintain two products”.

Me*: “It has users - 3000 weekly active”*

Cofounder*: “Will they pay?”*

Me (downcast): “No, no one pays for stuff like this. I will send a goodbye mail to everyone”


This was 2016. Augment was a plugin inside gmail that helped send data from incoming mails to other tools like Evernote, Asana, Trello etc. Later gmail implemented it's own "side panel" with integrations to Google Products. Needless to say, in 2016, there was no alternative yet. We'd hit upon a need and hit it well.

The next day, I sent out the goodbye email.

But there’s a twist.

In the goodbye email, I included a link to a survey, knowing fully well that no one takes surveys. By now you can tell that I “knew” everything - what people pay for and what they do.

People took that survey - we got about 250 responses.

One of the survey questions was “How much would you be willing to pay for something like this.”

Of course, I knew people were going to say “nothing”.

The Twist

Most people selected the option $3 per month.

Augment was just 8 months old with 3000 weekly actives. We could have taken it to a decent revenues in another year.

But we “knew” people wouldn’t pay, so we had already pivoted - 3 months ago. Augment was no longer actively developed. We had already sent the goodbye mail.

We shut it down.

The pain of “what it could have been”, especially with some data to back it up, was immensely painful.

Lessons Learned

  1. You never know anything - ALWAYS ask users to validate what you think you know. Sometimes we do not ask customers because there is always other stuff to do - write code, send out mails, write blogs etc. For closed ended questions, snap polls work like magic. For detailed, open ended questions, nothing beats directly talking to customers.
  2. Ask the right questions - This is so much easier said than done. Customers always give wrong signals (unknowingly) if you ask questions where emotion might be involved. Eg. A) "would you like to pay" vs B) "here's the payment link" are same questions but have different "activation energies". The latter needs more psychological effort. Those who click on that link are actually more likely to be real customers than those who said, "yes, I'd like to pay". Somewhere between these two is the question C "how much would you pay". This also requires some effort, but is not as great an indicator as question B, which is more action than question.
  3. Build a detailed view of the customers: This can uncover many blind spots. We had seen people being unwilling to pay for similar products in the reviews of those products. But our customers were mostly freelancers who were saving time - the exact commodity they sold. Thus they were more willing to pay than customers of competition we checked out. Later, in other startups, I implemented a much more detailed customer profiling system to discover their interests, habits, job, company etc. In all these cases, the product features and marketing copy, both evolved the right way.

3 years later ...

Fortunately, the story doesn't end with the shutdown (it never does).

I started up again, 3 years later - this time with a wife and a baby in tow. Guess what I wanted to build ;-)

A tool that can stop such accidents from happening.

Maybe this time some other mistake would kill me … ;-)