Should I do a lot of upfront planning around the work I will take on across the half?

Good question! I think you're already very much on the right track with that last point about having a pipeline of projects: That's definitely an example of E5 thinking!

I try to lean more structured, especially at Meta where things are so chaotic and blow up all the time. In more concrete terms, I would try to estimate projects in terms of weeks it would take with an uncertainty of +/- 2 weeks. However, there is a tension between taking a ton of time doing planning and moving really fast in the Meta style.

To clarify my thoughts here here, I can share how planning worked on my team at Instagram, Instagram Ads. I felt like the process didn't take too long (1 to 1.5 months) while being very structured and high-value. Here's how it worked:

• We have a huge list of projects that come from backlog, brainstorms, people's shower thoughts, etc.
• We would go through each project and estimate the impact. This was generally some math to calculate projected new revenue from each one.
• Alongside impact analysis, we would do a time estimation, trying to follow 80/20 rule (i.e. don't do a full blown tech review). Tech leads would give their input after thinking about it overnight, and if it was an XFN project, we would quickly ask other teams in Work Chat what they thought about this proposed project. If they looked at us like we were crazy people, we knew to jack up the time estimate or maybe even can the project. On a side note, make sure to include the following when it comes to estimates:
  • Planning time for a project. A lot of engineers only scope for execution time, which makes no sense as every project has so many phases on top of building.
  • PTOs (vacation, parental leave, etc)
  • Some guessed amount of sick leave. This is super relevant with coronavirus running around.
  • If your team is XFN heavy, throw in a couple weeks for XFN spaghetti, haha. It comes up all the time.
• After we have this list of projects with impact/time analysis, we would rank them by priority with P0/P1/P2, P0 being the highest. The goal would be to execute on all the P0s with a few P1s. If projects are blocked/killed for whatever reason, there's an ample backlog of P1/P2 projects to switch into!
• The end result of all this would be a giant spreadsheet in Quip, which we would then turn into Workplace posts to get max alignment and visibility.

All that being said, half planning is super important for almost all teams at Meta, and I highly recommend playing as big a role in it as possible if you're pushing for E5 to get a lot of credit on the Direction axis. And once you're pushing for E6, it is effectively required to be one of the most important people on the team in roadmap planning, if not just outright driving the entire thing.

In terms of how engineers can add value to this entire process:

• Help organize the brainstorm - Talk to your PM and designer to see how you can contribute!
• Evaluate impact of opportunities - At Meta, this means running the data and hybridizing as a data scientist.
• Propose ideas of your own - Make sure to create as compelling a case as possible behind them.
• Work with XFN to vet estimates - At a Big Tech company like Meta, it's vital to proactively figure out dependencies.

For tactics around how to come up with ideas to put onto the team's roadmap, I recommend these discussions:

• "How do I come up with innovative, impactful ideas and bring them to my team?"
• "How can I come up with big initiatives, especially those at a Staff level?"

You can not plan for everything, but you need to have enough clarity. Imagine a road trip over many days. You don't print out turn-by-turn navigation for the whole trip on day one. You don't know if there's traffic or detour. However, you have rough ideas of which cities you visit and which route you travel along in your mind. When your gas is low you look up the next gas station along the route but you don't plan them ahead of time. That's what it's like.

Learn through trial and error. Practice more and you'll gradually learn what you need to plan ahead and what you can improvise.