How we messed up being one of the early Slack bots

The idea of getting work done within Slack was exciting and we were the first to build something like that on top of Slack. It was cool to create a trello card, a new calendar event, etc.. within Slack. The team got excited about that idea, and we built a prototype in a week. And then we did an animation explaining the concept and the syntax. It took us about two weeks to get this all completed.

Building

It was exciting to see how cool and productive it was to do all this within Slack. We had no clue about product launches at that time. We also didn’t have any idea about product onboarding. We didn’t test this with a small subset of the audience. We decided to launch this on Product Hunt, which still is an excellent place to discuss about ideas and launch products.

The launch

I still remember the day where I had a chat with Ryan Hoover about this, and he was ready to approve the launch for me when I was ready :) Once we finished development, we got on Product Hunt https://www.producthunt.com/posts/interbot-2, and to our surprise, we ranked the #1 product of the day. We had more than 1000 teams signing up to try the product.

That was a huge milestone for us, and we, as a team, never expected this to happen. As Product Hunt usually does, the next day an email was sent to the whole community since we ranked #1. It gave us good traffic and an uplift in the spirit for us to keep moving.

We had the project manager of HipChat mailing us to build the integration for them.

And Slack was recommending us to their customers on Twitter.

As you can see we were in cloud nine.

The fall

Our rush to launch cost us a lot. We had one of the worst onboarding. At that time Slack didn’t have an App store, and we were impatient to get it approved by the Slack team. They told us it might take a month to approve, and they were building the initial structure for an app store.

The setup was hard

Users found it hard to set up the integration with Slack. They were not sure if they should be pasting the organization code or an individual code. Many people lost interest along the way and left the system.

The Command wasn’t easy

We showed @mention as the command to initiate this process in the demo (facepalm). But we developed a Slack command to do that process. We should have changed the marketing image as soon as possible, but we didn’t. And the command wasn’t easy to remember, so people typed the wrong commands and expected a different result. These started to result in a lot of people leaving the system.

A lost opportunity

We didn’t realize the responsibility of being one of the early Slack bots to do the job right and have a smooth and seamless onboarding. We missed the boat. Instead of iterating slowly and deeply, we spread too thin too fast.

Now Slack has built the exact thing by using cards https://slackhq.com/turn-your-conversations-into-action

Though we can tap our shoulder and feel happy that we were able to see this use case 4 years ago, we couldn’t make use of it.

Lessons learned

Though we weren’t able to make it through with the idea, we did have a ton of learning which we are applying to ReportBot.

Iterate slowly

We added too many apps before we launched. We could have built a better UI/UX or command recommendation engine if we had started with just Trello.

Launch slowly

Instead of doing a launch to all the users, we should have launched to say first 10–20 users taken a lot of feedback and improved the process, so the rest of the users would have stayed.

Stay put and be patient

Some opportunities don’t last long, and it also takes time. Patience should be a virtue. If we could have waited for Slack to bring it as a bot in their backend, it would have given us a marketing milage and would have helped us move forward.

Build a minimum viable product not a minimum product

There is a huge misconception with MVP. MVP doesn’t mean broken product. It’s the minimum usable product. To make it usable, you need a proper onboarding for users. Skipping all those small things will make your product unsable.

Interbot did lead us to our next path in which we failed with a ton of learning as well. I will be sharing that as a separate blog post along to this series of Journey to ReportBot

ReportBot is a Slack bot that sends vital reports such us revenue, subscribers etc. from services like PayPal, Stripe, MailChimp to your Slack team every day, week or month.