One of our goals this year has been to systematize things in the Revelry process as much as possible. A thing our pal Rick Webb talks about in his new book Agency. Some of our work has involved creating documents like our handbook, onboarding process, and training materials. We’ve also been building a product internally to automate various parts of our business workflow, including performance reviews.
We’ve neglected, well, more like avoided, doing official performance and peer reviews. We all work closely together, follow our own version of Scrum/XP/Agile, but our reviews have always been random. And most of the time happen over a walk to grab coffee. Conversations, while awesome, would go untracked and because we love lean, that was generally fine.
Now we have a bigger team and if done correctly, reviews can be super powerful. So, we started re-thinking the coffee walk review processes and other review processes team members have experienced in the past.
Performance Reviews
Our thoughts:
- Performance Reviews suck. Let’s make them suck less.
- Point systems suck. Let’s not put numbers on people.
- Peer reviews are better than top-down reviews.
- Reviews should be as anonymous as the author chooses.
- 1:1’s should be focused and take place in the office. Save the coffee walks for strategy and brainstorming.
As we were going through our past experiences, and all the bloat that went along with them, we decided to trim down to 2 questions.
- What is she doing well?
- What can she do better?
We asked everyone in the company to answer these 2 questions about other team members.
How we collect the data
We looked at a few SaaS products on the market. There’s some good stuff out there, but none as simple as we wanted, so I spent a day or so building a simple web app.
It’s a single form that lists all team members and asks the two questions. Anything can be left blank, form submissions are totally anonymous, and only management can access the answers.
Pretty simple stuff, but what could be even easier? Slack integration.
We’re all in Slack anyway, so why not? Time for some
Messaging Experience Design (MXD).
Read more about MXD in Ben Brown’s Medium Post, Slack is the Operating System
Our Slack Commands:
/peer
Sends back a link to the form mentioned above.
/peer help
Sends instructions on how to submit reviews through slack
/peer better:gerard spend more time writing blog posts
This will save a "do better" entry in the app
/peer good:gerard hacking prototypes
This will save a "doing good" entry in the app
How To
So, let’s take a journey into building this or something like it for yourself.
Get started by creating a new slash command. This will be the command users enter into slack with a preceding slash to trigger different events. In our case, we used /peer.
My little web app replies back to the individual in slack, but it can also send messages with help text or links back to the public channel. If you want to send public messages back to a particular channel, you will need an incoming webhook.
“These tools are great for building your own, personal integration, but if you’d like your service to be listed as an Official Slack Integration, please get in touch before beginning development. We’ll help guide you in the right direction and provide you with resources you may need along the way.” — Slack
Super simple and no barrier to entry for anyone on our team to submit performance reviews at anytime. It’s all anonymous right now, but total anonymity is up for debate. Overall, we’ve had amazing responses and the review process has been more of a dream rather than a nightmare.
We may release this as open source at some point, but in the meantime, drop us a line if you would like to discuss implementing a solution like this at your company. This is really what we do best: look at workflow problems and try simple solutions that make things go faster.
We're building an AI-powered Product Operations Cloud, leveraging AI in almost every aspect of the software delivery lifecycle. Want to test drive it with us? Join the ProdOps party at ProdOps.ai.