Selective Deep-Dive Leadership
Instead of skimming 100 projects, go bare-metal deep on the 7-10 that matter most
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 4
- Confidence
- 90%
A method for leading a large portfolio without spreading yourself thin across status updates. Rather than staying high-level across every project, you pick the handful most impactful to customers right now and go extremely deep on them — sitting with engineers, reading code, understanding root causes. The scrutiny creates a discipline flywheel across every other team, even the ones you aren't reviewing this cycle.
Origin
Described by Dmitry Zlokazov, Global Head of Product at Revolut, as how he changed his operating style after joining the founder-led company. Contrasted in the episode with Brian Chesky's opposite approach at Airbnb (cutting down how many things happen so leaders can be involved in everything).
Core principles
- 01Catching up on statuses across 50-70 projects adds value to no one and eats your entire week
- 02Depth on a few beats shallow oversight of all
- 03Public, unpredictable scrutiny of some teams builds discipline in all teams
- 04Autonomy scales; micromanagement does not
- 05Going deep is a spot-check, not a takeover — you're the last line of control, not the driver
How to run it
- 1
Inventory everything in flight
Acknowledge the full portfolio — at Revolut this is closer to 100 projects, features, and product launches happening simultaneously.
Watch out Do not try to track all of them equally; that just means spending the entire week catching up on statuses.
- 2
Pick the 7-10 most impactful to customers right now
Select the small set where things going wrong carries the most downside for customers and the business. This is a personal call by the leader, not a committee decision.
Pro tip Let priorities shift over time — the set is dynamic, not a fixed list.
- 3
Go bare-metal deep on the chosen few
Sit with engineers at the reading-code level, understand exactly how each thing is built and what the underlying issue is. Zoom into details, then zoom back out to a helicopter view to sense whether the team is on track.
Pro tip Alternate between zooming in on details and raising to a helicopter view to judge whether a team is on good track.
- 4
Let the scrutiny signal ripple outward
Teams talk to each other formally and informally, so they know which areas are being scrutinized. Teams not being reviewed now infer that meticulous, high-quality execution is expected because they could be the next deep-dive target.
Watch out This only works in a culture of ambitious people who want to prove they can deliver; in a low-accountability culture it reads as random policing.
In the wild
With ~100 things in flight and a role spread across the whole product function plus retail, Zlokazov personally selects seven to ten of the most impactful projects and dives to the reading-code level with engineers, rather than sitting in status meetings across all of them.
→ Teams operate autonomously, discipline stays high across the org, and Revolut ships an unusually large number of products fast for its lean headcount.
Common mistakes
Staying high-level everywhere out of fear
Many senior managers avoid going deep because it feels scary not to track everything — 'what if something goes wrong and I can't steer it?' The result is a week of status catch-ups that adds no value.
Turning a deep-dive into micromanagement
Going deep is a last line of control to verify things make sense, not taking over the team's decisions. If a product owner can show the logic behind their roadmap, they keep the credit of trust to build their way.
Is it for you?
Best for
Senior product or engineering leaders overseeing a large, fast-moving portfolio of parallel projects in a high-accountability, autonomy-driven culture
Not ideal for
Small teams with few projects, or low-trust cultures where selective scrutiny would feel arbitrary rather than motivating
From the transcript
“we take maybe seven or 10 projects which are most impactful for customers right now and Um we go super deep into them like really…”
“it's a signal to other teams that if they are not proactively executing things on the expected level”
“just just spend your entire week just just catching up on statuses. Uh not really adding value to anyone”
“it's more like let's say a last uh line of control to make sure that what we're building is all all makes sense”
From the episode
How Revolut trains world-class product managers: The “local CEO” model, raw intellect over experience, and a cultural obsession with building wow products
Dmitry Zlokazov (Head of Product)