The situation repeats itself: I walk into a company as its fractional CTO, the engineering org is a few dozen developers, and nobody — not the CEO, not the VP R&D, not the team leads — can answer "what is everyone working on?" without opening four different trackers.
The first week decides whether you're credible or just another consultant. Here's what I actually do with those seven days.
What I stopped doing
I stopped asking for status reports. I stopped asking for slide decks. Both are already out of date by the time they reach you, and both are written by the people you're supposed to be checking on.
Nobody in an 80-developer department is lying to you. Everyone is just reading a different tracker, and none of them match.
What I look at instead, in order
1. Where the work actually lives. Every company has an official tracker and several real ones. I've walked into places running Jira, Azure DevOps, and a GitHub Issues graveyard in parallel — active projects split across all three. Before you can say anything about delivery, you need to know which systems hold the truth, and the honest answer is usually "some of each."
2. Delivery rhythm per team. Lead time and deployment frequency tell you who's shipping and who's stuck. A team with a long lead time and low deploy frequency isn't slow — it's stuck behind something: a review bottleneck, an approval process, a database migration nobody wants to touch. The numbers don't give you the answer. They tell you where to spend your first week listening.
3. Workload distribution. Not rankings — I've written before about why leaderboards poison this data. The question is simpler: does the load rest on 3 people or 30? In most inherited departments it's closer to 3 than anyone admits. That's not a performance finding, it's a risk finding: your bus factor has a number, and it's small.
4. The quiet one. The engineer going quiet is more urgent than the engineer going loud. Tickets still close. Reviews still happen, eventually. But the activity curve — commits slowing, review comments drying up while the board stays green — tells you six weeks before the resignation email does. Managers miss it precisely because the tickets get closed anyway. The data doesn't miss it.
Why this became a product
I built the same dashboard at three companies in a row — pull the trackers into one warehouse, one board, the same four views. The fourth time, I built it properly and we open-sourced it.
Because the obvious next question is "fine, but what about my data" — it's self-hosted. Nothing leaves your server, no vendor operates it, you can read the source, and it starts at $600/month. That's the whole answer, and it's the answer because the alternative is sending your entire development history to someone else's cloud so a stranger can chart it.
Repo: https://github.com/Codpal-Limited/deckgauge — live demo with a fictional org at https://demo.deckgauge.com.
I'm Yaniv — fractional CTO, and I've stopped pretending the first week is about anything other than reading the data before anyone tells me the story.
Top comments (0)