Managing Managers of Managers: Leading a Product Org You Can't See

When you manage managers of managers, you rarely see raw product work — it reaches you filtered through a director's judgment. Your leverage shifts from making good decisions yourself to installing good standards and questions that shape every review beneath you. Get this wrong and you either bottleneck the org two layers too deep, or become irrelevant to it.

Quick answer: Managing a two-tier product org means directors sit between you and PMs, so your primary tool becomes the standard questions you ask in director-level reviews — not personal review of PM work. Spot a struggling director two levels down by watching patterns in how they run their own reviews, not by pulling their PMs aside.

The Job Changes: From Reviewing Work to Reviewing How Work Gets Reviewed

Once directors stand between you and individual PMs, reviewing product work directly becomes rare and often destructive. Your real job shifts to reviewing your directors' reviews — the questions they ask, the bar they hold, the rituals they run with their own teams. Skip this shift and you either bottleneck decisions two layers too deep, or fade into irrelevance.

In a single-tier org, a sharp VP can read every PRD, sit in every roadmap review, and catch a bad prioritization call before it ships. That instinct — stay close to the work — is exactly what made you a good PM and then a strong first-line leader. It actively works against you the moment a second management layer exists.

Andy Grove's framing in High Output Management is worth borrowing here: a manager's output is the output of the organization under their supervision, not their own individual output.

For a VP over directors, that means your output is the quality of decisions made when you're not in the room. You can't personally inspect all the raw material anymore — there's too much of it, and your directors need room to actually manage.

What breaks if you skip the shift:

• You become a bottleneck two layers too deep. Directors route decisions upward instead of owning them, because you keep answering questions that are theirs to answer.
• Your directors stop developing. Fix a PM's roadmap yourself and the director never builds the judgment to catch the next one.
• You lose the leverage you actually have. One VP reviewing twenty PM artifacts a week is a rounding error. One VP shaping how four directors run forty reviews a week changes the whole org.

This particular transition gets little airtime because most advice on the VP of product role addresses the first scaling jump — IC PM to group leader — not the second one, from leading PMs to leading the people who lead PMs. The underlying skills don't transfer without deliberate re-tooling.

Managing the Operating System: Your Questions Become the Culture

Every question you ask a director in a review is a proxy for what you actually value, and directors repeat that exact question with their own PMs within days. Manage the operating system, not the outputs: the recurring questions and standards you apply consistently are what cascades three and four layers down, long after you've forgotten asking them the first time.

Think of it as building a leadership equivalent of a product operating model that runs without you in the room — except what you're instrumenting isn't a roadmap process, it's your own behavior inside reviews. If you consistently ask "what evidence says this is the top problem," directors start asking their PMs that before they ever bring you a roadmap. Ask about trade-offs instead, and teams learn to lead with trade-off reasoning rather than feature lists.

Three ways the operating system compounds, for better or worse:

1. Directors imitate your literal words, not your intent. Ask "what's the ARR impact" and you'll get ARR-justified roadmaps even in cases where ARR is the wrong lens for the decision.
2. Silence trains behavior too. A question you stop asking is a standard you've quietly retired — teams notice what you stopped checking almost as fast as what you started checking.
3. Inconsistency reads as the real standard. Ask a rigorous question once and skip it next time, and the org learns the bar is negotiable, not that the question itself matters.

Camille Fournier makes a related point in The Manager's Path: as you move up, your job becomes ensuring good judgment exists in the organization even when you're not present to supply it. For a VP two layers up, that's close to the entire job description — the org's default behavior when you're not watching is the only behavior that actually matters at your altitude.

Leverage Questions vs. Micromanaging Questions

A leverage question probes the reasoning, evidence, or trade-off behind a decision. A micromanaging question probes the decision's specific content and quietly substitutes your judgment for the director's. The test: if you heard only the director's answer, with no artifact in front of you, would the question still make sense? If not, you're operating one layer too deep.

The same review moment can produce either kind of question — the difference is altitude, not topic. Here's how five common scenarios split:

Notice the pattern: leverage questions almost never contain a verb you'd perform yourself — build, pull, add, fix. They contain nouns like evidence, job, bar, and ranking, which only make sense if someone already did the analysis and you're auditing it. That's a usable, live tell in the room, before you ask something you'll regret.

Signs you've slipped into micromanaging:

• You're naming the solution, not the criteria. "Do X" instead of "what did you weigh before choosing X?"
• The director's answer to you differs from what they'd tell their own PM. That gap means you're being managed, not informed.
• You feel the urge to fix it inside the meeting. That urge is real information — about a process gap, not about this one decision.

How to Spot a Struggling Director Two Levels Down

You can't read a PM's backlog to know whether a director is struggling — you have to watch for breaks in the director's own operating rhythm. Cancelled 1:1s, roadmaps that don't move even as evidence changes, PMs who quietly stop disagreeing, and escalations that skip the director entirely are all visible from your seat, without ever pulling a PM aside to ask what's wrong.

None of this requires an investigation — it requires attention to a rhythm you already observe. Gallup's long-running manager research has repeatedly found the manager accounts for a large share of the variance in team engagement, directionally around seventy percent in some analyses, which means a struggling director shows up in their team's energy well before it shows up in a metric. Google's Project Oxygen research found similarly that the specific behaviors of a manager, more than their technical depth, predicted team performance most reliably.

The moment a skip-level starts feeling like an audit to your directors, it stops working as a listening tool and starts working against you.

The discipline is resisting the urge to fact-find with the director's own PMs first. A skip-level conversation should be listening, not an audit — ask what's energizing and what's frustrating, and if a name comes up repeatedly, take it back to the director as a coaching conversation, not a verdict delivered behind their back.

Rituals That Give You Visibility Without Hovering

Visibility at your altitude comes from a small set of repeating rituals, not from spot-checking work: a real skip-level cadence, directors reviewing each other's specs and roadmaps as peers, and a standing translation of what's happening in the org into language your own boss and the board can use. Build these once and they run without you inserting yourself into every review.

• Skip-levels on a fixed cadence, not a crisis cadence. Monthly or quarterly, scheduled in advance, so they read as normal practice instead of a warning sign when they happen.
• Peer review circles among directors. Have directors review one another's specs or roadmaps periodically — it surfaces inconsistency in standards faster than you reviewing each one separately ever could.
• A standing narrative you can hand upward. Translating what four directors and their teams are doing into one coherent story is a distinct skill from running the org itself — the same discipline behind explaining product's business impact when the board asks.

None of these rituals replace judgment, and none of them are free. Peer review circles take director time away from their own teams, and skip-levels can misfire if directors feel bypassed. Frame them explicitly as listening and coaching, not evidence-gathering against a director — say that out loud, repeatedly, or the ritual itself becomes the thing that erodes trust.

Give Your Directors a Shared Rubric, Not Just Your Opinion

Coaching a director on how they coach their own PMs is hard without a shared vocabulary for the behaviors you're both trying to cultivate — discovery rigor, trade-off reasoning, stakeholder communication. A shared rubric turns a vague "raise your team's bar" into specific, discussable competencies you can both point at, instead of the director guessing what's in your head.

This is the gap Prodinja's Leadership Suite is designed to close with its Growth competencies — a shared rubric of the PM behaviors you want cascaded down the org, visible to you and your directors alike. Rather than coaching a director from your own private mental checklist, you're both looking at the same defined competencies, which makes it easier to agree on where a director's team is genuinely strong versus where the director's own coaching has a gap. It's a shared reference point for the conversation, not a replacement for having it.

A rubric doesn't do the coaching for you — it just means you and a director are debating the same five things instead of five different things each of you privately assumed mattered most. That alone removes a surprising amount of friction from one of the hardest conversations in a two-tier org: telling a director their coaching, not their output, needs to improve.

Key Takeaways

• Your leverage moves from decisions to standards. At this altitude, your output is measured by the quality of decisions made in rooms you're not in, not decisions you personally make.
• Every question you repeat becomes policy. Directors copy your literal questions onto their own PMs within days — choose them as deliberately as you'd choose a roadmap priority.
• A leverage question probes reasoning; a micromanaging question probes content. If your question only makes sense with the artifact in front of you, it's one layer too deep.
• Struggling directors show up as pattern breaks, not confessions. Cancelled 1:1s, static roadmaps, silent PMs, and escalations that skip a layer are all visible without ever going around the director.
• Skip-levels are for listening, not investigating. Frame them that way explicitly, or they become the thing that damages the trust they're meant to protect.
• Rituals scale your attention; spot-checks don't. A fixed skip-level cadence and director peer-review circles give you real visibility without inserting yourself into every review.
• A shared coaching rubric turns opinion into a discussable standard. Whether it's a framework you built yourself or one like Prodinja's Growth competencies, you and your director need to be pointing at the same list.