Managing Up: The Three Relationships That Decide Whether Data Leaders Get Funded, Trusted, and Promoted

On a Tuesday in October, a data engineering director named Matt learned the hard way what managing up actually costs. He did something out of character: he audited his own calendar.

The trigger was a comment from his team lead, made without malice in a sprint retro: “We mostly see you on Mondays now.” Matt laughed it off in the room. That evening, he pulled four weeks of calendar data into a spreadsheet, because he was still the kind of person who responds to discomfort by building a dataset.

The numbers were worse than the joke. Of his twenty-seven weekly hours not spent with his own team, eleven were consumed by exactly three relationships. Four hours went to his CDO, who had recently started requesting a written status update every Friday, then a revised format two weeks later, then a third format with a risk column added, and who opened every Monday one-on-one walking through the document line by line. Three hours went to recovering ground with the CTO, who had reassigned a major streaming project to his own platform team in a review Matt wasn’t invited to, on the logic that it was “infrastructure, not data.” And four hours went to an escalating sequence of meetings with a peer VP of Engineering, who had absorbed that reassigned project enthusiastically and was now proposing, in writing, that data pipeline reliability should move to his team too, “for consistency.”

The Forty Percent Problem

Eleven hours. Forty percent of his cross-functional time, spent managing one relationship upward that was tightening, one diagonal relationship that was dismissive, and one lateral relationship that was quietly annexing his charter.

His first instinct was the engineer’s instinct: this is overhead, and overhead should be optimized away. It took him a while to accept what a mentor eventually said to him directly: “That is not overhead on the job. At your level, that is the job. The only question is whether you do it deliberately or reactively.”

Managing Up Is Not Politics. It Is Information Management.

Most advice on managing up was written for a generic office, not for a data leader sitting between two languages that rarely overlap: executives who think in outcomes and quarters, and technical teams who think in systems and dependencies. Managing up in a data organization is not flattery or reading the room. It is the deliberate work of getting accurate information to the people above, beside, and diagonal to you before someone else’s incomplete picture fills the gap instead.

Here is how that plays out in the three relationships that eat most of a data leader’s calendar.

Managing Up: The Micromanaging CDO

Start with the relationship above, because it sets the weather for everything else.

A CDO who suddenly wants weekly written status, in a format that keeps evolving, reviewed line by line, is exhibiting a symptom. The disease is almost never what it feels like from below. From below it feels like distrust of you. In reality, it is usually pressure on her. Executives tighten their grip on the functions they expect to be questioned about. Matt’s CDO had spent a quarter defending her team’s reporting line to the CFO. She wasn’t auditing Matt. She was arming herself.

The two responses to micromanagement have opposite effects, and leaders under pressure usually pick the wrong one. Compliance with friction, producing the requested updates minimally and late, confirms the executive’s anxiety. The function that resists visibility must need watching.

The Thursday Note That Ended the Line-by-Line Reviews

The response that actually works is counterintuitive: give more information than requested, sooner than requested, in a format built for the executive’s actual use rather than your convenience. Matt replaced the Friday status document with something shorter, sent Thursday night: three lines of what moved, one line of what’s at risk and what he was doing about it, and one line of what she might get asked about by her own peers that week.

That last line was the unlock. It showed he understood the document’s real purpose, which was her survival in rooms he’d never see. Within six weeks, the Monday review shrank to ten minutes. Within a quarter, she canceled the written requirement entirely, because the Thursday note had made it redundant.

Micromanagement is almost always a trust deficit being serviced manually. You do not negotiate it away. You make it unnecessary, by proving the information arrives without being extracted.

The One Boundary Worth Testing

There is a version of executive tightening that is not pressure-driven but temperamental, the leader who will consume every byte of reporting you produce and ask for more, forever. The test: does proactive communication reduce the demands within a quarter? If yes, you were servicing a deficit, and it’s now serviced. If demands keep growing regardless, you are not facing a trust problem. You are facing a management style, and the response shifts from information to an explicit conversation about what level of reporting is sustainable.

Managing Up: The Dismissive CTO

The diagonal relationship is harder, because the CTO owes you nothing. He is not in your chain. He simply holds a theory, common among CTOs, that data engineering is a species of infrastructure and infrastructure belongs to him. A reassignment like the one Matt faced is rarely an attack. It is the theory, acting on the org chart.

The instinct is to contest the decision directly, escalate, argue charter, invoke your own boss. Sometimes necessary. Almost never sufficient, because winning one territorial fight does nothing about the theory that generated it, and the theory will generate another decision next quarter.

What actually changes a CTO’s theory is not argument. It is peer-level technical credibility, deployed where he can see it. Matt’s real problem was that the CTO had never once watched him operate technically. Every interaction between them had been status, budget, or escalation, formats in which Matt appeared as an administrator of a cost center.

So Matt changed the evidence. He asked to present at the CTO’s architecture review board, not about his own roadmap, but about a genuinely hard technical problem: the consistency tradeoffs in the very rebuild that had just been reassigned. There was no relitigating of the decision. In front of the CTO’s own architects, in the CTO’s own room, he demonstrated that the problem had depth the reassignment hadn’t priced in, and that the person who understood that depth ran the data team.

Within weeks, the platform team was quietly consulting his senior engineers. At the next planning session, the CTO proposed the rebuild be “co-owned,” which everyone in the room understood was the dignified word for returned.

A dismissive senior technologist is dismissing the evidence he has. Supply different evidence, in his forums, on his hardest problems, and let the theory update itself. Slower than escalation. The only version that lasts.

Managing Up: The Annexing Peer

The lateral relationship gets the least written about it and carries the most career consequence, because peers are the reference class promotions get calibrated against, and peer conflicts have no referee.

The VP of Engineering in Matt’s case wasn’t a villain. Absorbing data infrastructure looked like synergy from his side of the table: his team already ran reliability for everything else, and every responsibility his org added made his next headcount case stronger. Annexation among ambitious peers isn’t hostility. It’s gravity. Undefended charter drifts toward whichever organization has the most mass.

The defense has three layers, and only the last one is a conversation.

Legibility first. Charter nobody can see is charter nobody respects. A quarterly operating review that states, in front of a room of stakeholders, exactly what your team owns and what it delivered against that ownership, removes the ambiguity annexation thrives on.

Cost second. Absorption proposals are cheap to make when saying yes is free. Matt made the true cost visible in writing: the domain knowledge across dozens of upstream systems, the stakeholder relationships behind every quality commitment, the on-call expertise that could tell a data correctness incident from a plain availability incident. The proposal didn’t survive contact with its own price tag.

The conversation last. Matt eventually booked the VP for lunch, no agenda, and said something close to: “We are going to be peers for years. I’d rather trade with you than skirmish with you. Tell me what you’re actually trying to solve, and I’ll tell you what I’m protecting and why.” The answer was illuminating: the VP was under pressure to cut cross-team incident handoffs, a real problem annexation would have solved badly. They built a joint runbook instead. The skirmish line became a trade route.

Managing Up Is Not Politics

Managing up, done well, looks nothing like the politics people fear it to be. Somewhere in all this, a natural objection surfaces: isn’t this just politics? The distinction is real, and it matters. Corrosive politics is the manipulation of perception away from reality, claiming credit that isn’t yours, undermining rivals privately, telling each audience what it wants to hear. It works briefly and compounds negatively, because organizations are small and memories are long.

What actually works is the opposite: the deliberate transmission of reality to people who would otherwise act on incomplete information. The CDO tightened because she lacked information, so supply it. The CTO dismissed because his evidence was thin, so thicken it. The peer annexed because the charter was ambiguous and his real problem was unsolved, so clarify the one and solve the other. In every case, the move is to make the truth more available, not less.

The Skip-Level Most Data Leaders Never Use

There is one upward relationship rarely discussed, and for most data leaders it is the least managed asset they have: the skip-level, the executive above your own manager.

Most leaders treat it as somewhere between forbidden and dangerous, so they confine it to the mandatory annual meeting, perform thirty minutes of well-rehearsed positivity, and waste the single relationship best positioned to change their trajectory.

The fear is based on a real failure mode and a wrong conclusion. Using the skip-level to relitigate decisions or lobby against your own manager genuinely is fatal. But the legitimate channel is wide and almost entirely unused. Skip-levels want three things from the leaders two levels down, and providing them is service, not betrayal:

• Unfiltered signal about ground truth, what’s actually happening in the work, not complaints
• A read on who is real two levels down, because succession planning is a standing item on every senior executive’s private agenda
• Their own blind spots serviced, one honest sentence that says “here is something I think looks different from your altitude than it does from mine”

The protocol is short: a standing quarterly cadence, never let it decay into status reporting, tell your own manager it’s happening and roughly what was discussed so the channel stays clean, bring signal and one honest question, and never bring a verdict on your manager or an appeal the formal chain should handle.

When You Are the Problem

One honesty check belongs at the end of all this. Every framework above assumes the friction is a management problem the executive is having, not a performance problem you are having. Usually true. Not always.

Three signs distinguish a relationship problem from a performance problem wearing a relationship costume. The friction is coming from multiple directions at once, one difficult executive is a relationship, three simultaneously is a pattern, and you are the constant. The complaints are specific and dated rather than vague and stylistic. And your best people are not defending you, when an organization unfairly squeezes a good leader, that leader’s team usually pushes back through their own networks, and silence from your own team can mean they are seeing something you are not.

The test costs one honest conversation with the person in your professional life least likely to flatter you. Ask them directly how much of the friction you are generating yourself. Then absorb the answer without arguing with it.

The Takeaway

At the senior level, the relationships above, beside, and diagonal to you consume close to half your calendar and decide most of your outcomes, and they respond to deliberate management exactly the way your own team does, just with different tools. Upward, trade proactive transparency for autonomy. Diagonally, change the evidence rather than contesting the conclusion. Laterally, defend charter with legibility and honest pricing, then solve the real problem underneath the encroachment.

None of it is politics. Managing up, at this altitude, is simply the job. And the forty percent of your calendar it consumes is not the tax on the work. At this altitude, it is the work.

Three Steps to Start This Week

Run your own calendar audit. Pull four weeks of meetings and measure how much cross-functional time goes to managing upward, diagonal, and lateral relationships, and to whom specifically. The number is usually close to forty percent. The question it answers is not how to shrink it, but whether you are spending it deliberately.

Write your own Thursday note. If anyone above you is asking for more reporting than feels reasonable, invert the dynamic. Send a five-line proactive note before they ask for it: what moved, what’s at risk and your response, what they might get asked about by their own peers. Run it for six weeks and watch what happens to the requests.

Price one encroachment. Find the one place a peer is absorbing or proposing to absorb part of your charter. Write the true cost of that ownership, factually and without heat. Then book the no-agenda lunch and find the real problem underneath the proposal.

This is one relationship inside a much larger picture. For the complete framework on how data and analytics leaders build influence, navigate reporting structures, and lead through the age of AI, see The Data-Driven Executive: How Data and Analytics Leaders Build Influence and Lead in the Age of AI.