Looking back at Broadcast 1.0: What did we build and why did we move on?
We're building a new product for managers at growing companies! But here's a quick recap of the previous one.
PS. We’re launching a new product for managers @ growing companies next week and we’d love for you to try it out! If you want early access: DM me or reach out at varadh@withbroadcast.com. Really excited book some time: https://calendly.com/varadh-broadcast/30min, ha!
About 7ish months ago we moved on from the automated updates product to building a few other products (standups & changelog) which we’ve since moved on from as well. The bulk of our time was spent on Broadcast 1.0 and this post is a brief reflection of that journey.
This blog post covers 3 topics:
Product: Automated updates
Promising ideas: Split View and Smart Blocks
Why we moved on: Horizontal product challenges and frequency of usage
Product
Our north star has always been and will continue to be focused on bringing the power of AI and automation to your workplace to give you time back and achieve the company’s mission with less overhead.
The place we picked to get started on this journey was a place of personal pain and discovered pain and talked to hundreds of users at companies of various stages in the first year. Not everyone looks forward to writing their updates, collecting parts of it from their colleagues, figuring out what happened, etc. but they all admitted it was an important part of their job, especially if they were a manager.
They were not alone! Updates are how people find out what’s happening on other teams, drive accountability on their own, and know the status of projects.
How information travels is critical to the success of every single company and many have talked in depth about it.
Stripe goes in-depth on this topic: Email Transparency @ Stripe circa 2013
Shopify built an internal tool called Vault: Here’s a video walkthrough by Jean Michel Lemiuex their CTO at the time
Jensen Huang, CEO of Nvidia talks about how he receives short weekly emails from 50+ direct reports as a way to manage his company
I like the quote below as an encapsulation.
“Perhaps the CEO’s most important operational responsibility is designing and implementing the communication architecture for her company.” - Ben Horowitz via Gokul Rajaram’s blogpost on Communication Architectures LINK
Communication and coordination is a lubricant of capitalism. And it often it happens in patterns and rhythms and we wanted to own those via an automated updates product.
So we focused on automating the meta work (collecting updates, sending reminders) and automatically pulling in information from other tools that can be used.
Our product’s flow was:
Create a recurring update →
Set a template for people and apps to fill out →
Get reminders for the next editions →
Prefilled content →
Edit it →
Ship and share over email and Slack
People now know what’s happening!
While we made the entire process about 2-3x better in terms of time saved and annoyances had. This would’ve been sufficient to start generating revenue if we were competing against a fixed other tool but for a workflow that people have in docs, slides, and Slack that’s been hacked together, it turned out to be not enough. Get stuck had too much work to do, it wasn’t 100% automated yet!
Promising Ideas
We’d love to see more of these ideas in other tools!
1. Side by Side View
You’re always pulling information from other sources when writing updates. Whether it’s a previous update, somebody else’s update, or information from a tool of yours.
We built a beautiful editor where in the right sidebar, you could see your previous updates or any other information you might want to pull from to write the next update!
2. Smart Blocks
We believed deeply that pulling information out of your tools into structured narratives would ease the process of writing dramatically. This was a middle ground we built until eventually, the AI will be able to go pick the tools, filter the information, and bring it back to the app (though this is much further out than I originally believed it to be)
Moving On
Post-launch last-January, we onboarded about 50 ish companies into paid pilots selecting them from a list of 500 companies that told us they wanted what we were building in an in-depth way. A couple of thousand people in total signed up. So this overall tells there’s a significant signal to solve a big problem here. Thank you to all the people (you!) who talked to us about your problems and gave us a chance to solve them.
Horizontal Product Challenges
tl;dr Turned out to be extremely challenging for a small team of 5 on a pre-seed. Likely it would help if you had 3-5x the amount of capital we raised to fully test our hypothesis or within a larger organization.
Our approach to the solution required us to be a mile wide and an inch deep. The reason for this is we were taking on a workflow that many folks do in different ways in various other tools and we needed to meet parity with some of their existing workflows. While possible, doing that in a performant way while building out various innovations proved to be,
The other thing that’s very hard for horizontal products is who is their champion/persona inside an organization buying it. A proxy test of this is, to ask your friends to describe Notion or Slack and who on their team set it up, you’ll get an incredible diversity of answers because the marketing, positioning, and persona work is a very fun challenge for these products. Deciding between tops down vs. bottoms up. How do you position it in outreach efforts? This topic deserves its blog post.
Frequency of Usage
tl;dr Known issue before launch but it caught up to us through feedback cycles
We knew what we were building was only a starting point inside the organization. it’s very hard to build a company around a product that a person uses only 1-2x a week and we were aware of that. However, it seemed like a critical workflow in certain roles in the organization so a good place to get started. We didn’t realise until about 3-4 months post launch (1.5 years into the company) that if your user is only using it 1x a week, then unless you have an infinite warchest, you’re not iterating fast enough! This made it too slow a burn for us to demonstrate progress fast enough
But our work is not done yet! New stuff coming soon
Many folks are working on similar or adjacent problems, and I wish them all the best: Campsite, Syncup.ai, Stepsize, Velma, Spoke, Glean!
Also thank you to many of you who asked some great questions which I’ll cover at a different time! @gm_mertd @designpapa @theericanderson
Thanks for posting this - also feeling the frequency of usage problem in our current iteration of the product. That is so critical for speed of learning, I agree! Good luck on your next launch!