I presented this talk at VSLive Las Vegas 2026. Here's the deck and some notes on what it covers.
The Post-Agile Reality
Everybody's doing something. Some teams are doing Scrum. Some are doing Kanban. Some invented something last Tuesday and aren't sure what to call it. Most are doing Scrum but skipping half the ceremonies. And the world has gotten pretty sick of all of it.
The complaints are real — story points lead to hours of debate and the estimates are still wrong, rigid sprint boundaries create artificial pressure, and required meetings become zones where people check out. But the answer isn't "no process." The answer is figuring out what survives any methodology and doing that well.
Three questions. That's it. What are we going to work on? What are we working on right now? Are we on track? If you can answer those, you have a process. Everything else is optional.
Why GitHub Projects
Your code is in GitHub. Your PRs and CI/CD are in GitHub. Your project management is in... Jira? A spreadsheet? Someone's head? Context-switching kills more projects than bad methodology. GitHub Projects puts your work management in the same place your code lives.
This talk is a live build. We go from zero to a working project workflow, covering the full loop: creating a project, building out a backlog, custom fields, prioritization, the whole draft-issues-vs-real-issues situation (which is genuinely confusing the first time you hit it), iteration planning, and views for different roles.
My practical tip that I think saves people the most headaches: create a dedicated "project issues" repository with the same name as your project. GitHub Projects can pull issues from multiple repos, and it will get tangled and messy if you let it. Keep it simple.
Flow Metrics: Stop Estimating, Start Measuring
The back section of the talk is the part that tends to light people up. Two numbers tell you almost everything: throughput (how many things you finish per week) and cycle time (how long things take from start to finish). No story points. No velocity. Just count and measure.
Once you have those numbers, you can run Monte Carlo simulations — basically, run 1,000 versions of your next few weeks based on your actual historical data and give probabilistic forecasts. "We're 85% confident this will be done by March 28." No estimation meetings required.
I wrote a free tool that does this for both GitHub and Azure DevOps: HonestCheetah.
Download the Slides
If you want to go deeper, I have full YouTube series on both Scrum with GitHub Projects and Flow Metrics Explained.
—Ben
