Break Orbit, Not Systems – LaunchDarkly’s First Conference

We’re so excited that we have announced our first conference and are actively collecting talk submissions and selling tickets for it!

We thought you might be interested in hearing why we’ve decided it was the right time for a conference about feature management, feature flagging, testing in production, and trunk-based development.

We want to learn from you!

As the developer advocate, I spend a bunch of my time in the field, going to conference talks and listening to what’s happening in the world. And I know that there are people who are stone-cold experts in flagging systems out there. We want to give you a reason to get together and talk to us, and each other, about what you’ve found out about scale, technical debt, and best practices. We saw that Gremlin’s ChaosConf and Honeycomb’s o11ycon gave practitioners a chance to exchange information and cross-polinate ideas, and we’d like to carry that on. I firmly believe that getting people with a common interest together leads to leaps for the entire category they talk about. Look at DevOpsDays!

People are hungry for best practices

Sure, I can tell you about what kinds of flag naming schemes work best, but wouldn’t you rather hear it from someone who is doing it at massive scale, and has been for years? Getting people together, from beginners, to people who built their own systems, to our customers who are using us at scale, gives us all a chance to learn and codify best practices. At LaunchDarkly, we feel like this year is an inflection point for feature management adoption, and as a part of that industry, we want to help the best practices apply as broadly as possible. We hope that Trajectory will let us talk about how testing, design, development, release, and sales can all be accelerated and made more powerful. How are you reaching escape velocity? We’d like to hear.

We think you’re awesome and have great ideas!

We have plans for what direction we’re going to take LaunchDarkly, and what’s next on our wishlist, but we want to make sure it’s going to serve you, our customers and potential customers. It’s no fun building a great technological achievement that doesn’t actually help people.

When people ask me how Edith and John got the idea for Feature Flags as a Service, I always tell them what Edith told me, that she was on so many late-night, weekend deploy bridges that she grew to hate them, and eventually, to build a product that means people can launch and test their software without late-night drama or stress.

We always want to be doing that, making your life easier – so tell us how! We made a wishlist of talks that we hoped would get proposed, but I bet you have ideas on things that you learned that you can talk about or write about. Here are some of the things we thought of:

  • Transformation or disaster stories – learning that we need new software to safely operate in this new world
  • How we deconstructed a monolith
  • How to knock down the cultural barriers and get DevOps/feature flag adoption at massive scale
  • GDPR – 1 year after
  • The new tech stack for brownfield development
  • How to test in production
  • The path to continuous delivery

Other awesomeness

  • The venue is at the Oakland Museum of California
  • Oakland-based catering is gonna be amazing!
  • You! And people like you! And I’m helping pick the swag!

I hope to see you there! Tickets are limited, so don’t overthink it!

On the Origin of the Speciated Conference

I go to so many conferences! It’s an awesome and amazing part of my job. I speak at them, but I also attend them. I sit in the front row and live-tweet. I attend talks. I participate in unconference sessions. I talk to people in lines, and at lunch, and at the afterparty. I give out stickers and I say hi to the vendors. Conferences are something I’m an expert at. And when I’m not doing technology stuff, I am support crew for science-fiction conference runners.

Given that, I’m surprised that I haven’t seen a taxonomy of technical conferences, something that would help you understand which flavor of conference you’re about to go to. As near as I can tell, about 10 years ago, there was a great flowering of conference types.

Originally, we had The Technology Conference. People paid a lot of money, they sat in large conference rooms (over 100 people, say), and they listened to industry experts. Some/many of the industry experts were also vendors giving pitches.

Then, a decade ago, many people decided that this method was not meeting their needs, and they wanted more interaction, more peer contact, more connections. We ended up with some new species of conference:

  • The regional variant of a language conference. No longer just PyCon, but PyConAU, and EU, and the same for JSConf and Ruby. It was cheaper to get people to someplace close to them, the conferences were smaller, the odds of meetings speakers and experts was higher.
  • Single-track conferences with registration caps. Write the Docs, The Lead Developer, and (I think) Monitorama use this method. Everyone attends the same talks, but the registration cap means that it’s still possible to identify and talk to a speaker. A well-run single-track conference allows a lot of time between talks so people can mingle and talk.
  • DevOpsDays. The DoD format is flexible, but tends toward the single-track morning, unconference afternoon. They also work really hard to fit into budgets that allow people to attend on their own, with lowish registration fees and locations all over.
  • No Fluff, Just Stuff. The first unsponsored conference I’ve spoken at. No vendors, and a high rate of repetition for speakers and a lot of tracks, so odds are good that you will be in a small group.
  • Birds of a Feather. Not unique to any one conference organizing system, but a way for people interested in a similar problem to find each other and do collaborative learning. Mostly these happen during non-programming time.

All of these conference styles prize collaborative learning over authoritarian instruction. If you’re a speaker coming from a more authoritarian background, I have to imagine the change is a bit of a shock. I know I have felt weird when presented with a large audience that I can’t see. I don’t want you to think that one way or the other is better – depends on what you need. 50k people wouldn’t go to AWS Reinvent if there wasn’t a value to be found in it. And Reinvent has small, unrecorded sessions as well as the massive keynote sessions.

Hotel ballroom filled with a few hundred people, all facing toward a podium and the camera.

DevOpsDays Toronto

So who are the stakeholders for running a conference?

  • Attendees
  • Sponsors
  • Speakers
  • Organizers

When you maximize the happiness or utility for one group, the utility for other groups goes down, or may go down. There are some overlaps. Attendees want content that answers their questions. Speakers want to provide content that is new and promotes their personal brand. Organizers want to select speakers who bring good value and are reliable. Sponsors want their speakers selected because talking about a product drives sales. Attendees, on the whole, don’t want sales-pitch talks. You see the problem!

As a speaker, I prefer single-track conferences. That way, I never miss other people’s talks! The talks are also usually very highly curated, since a day-long conference might only have 7 speakers, so it’s pretty darn flattering to get picked. As an attendee, I like conferences that are sized so each speaker ends up talking to about 50 people. It’s small enough that I feel engaged, and big enough that the speaker doesn’t feel like they have to stop to take questions. As a sponsor, I want multiple tracks with large spaces where people have to walk past my booth to get caffeine. As an organizer, well, I’m still working on that.

I’m thinking about this because LaunchDarkly is assembling our first conference this year (2019), in the spirit of Gremlin’s Chaos Conference and Honeycomb’s o11ycon. What do we want to give people, how many people do we think we’ll have, and how do we make the experience useful?

In the spirit of testing in production, we’re going to try a combination of things – keynotes will be one-track, so everyone has a common thing to talk about, and then we’ll split into other configurations in the afternoon.

We’re looking for people who want to join us on April 9 at Trajectory to talk about feature flagging, trunk-based development, devops tools, testing in production, blue-green deployments, and other ways to speed up your development and delivery…safely.

https://www.papercall.io/trajectory

If you want help with your pitch, or want to noodle around an idea, let me know. I’ll be back at work on the 7th and ready to think it through with you! (Yes, we’ll do bigger announcements later!)