2026-11-05
November 5, 2026
Most TPMs invest in managing up. The trust that decides whether your program ships is sideways, and you can audit it with a four-term equation.
Ask a senior TPM where they spend their relationship budget and the answer is usually "up." The sponsor, the skip-level, the VP who owns the roadmap. That instinct is not wrong. But look at where programs actually stall: a platform director who deprioritizes your dependency, a product peer who re-litigates scope in every review, a security lead who finds a blocker the week before launch. None of them report to you or to your sponsor. All of them can slow you down at will.
The standard response is rapport. Coffee chats, a friendly Slack tone, a reputation for being easy to work with. Rapport is real, but it is one input, and it is the one people over-index on because it feels like the work. Trust is a composite. If you cannot say which part of it is weak with a given peer, you are guessing at the fix.
There is an old, plain tool for this: the trust equation from David Maister, Charles Green and Robert Galford's The Trusted Advisor (2000). Credibility, reliability and intimacy in the numerator, self-orientation in the denominator. This issue argues that a senior TPM or tech leader should run it on their five most important peers, not their boss, and that the denominator is where most of us lose.
Why sideways is the hard direction. Managing up has a built-in gradient. Your manager has a reason to hear you, a calendar slot for you, and a stake in your success. Peers have none of these by default. They have their own roadmap, their own sponsor, and a legitimate interest in saying no to you. Dependencies between their team and yours are, from their side, a tax. That is the structural fact. Your charm does not change it; the shape of the exchange can.
LeadDev has run material on this for years. Aish Raj Dahal's Staff+ London talk, published in 2023, makes the case that peer relationships between staff-level engineers get far less attention than mentoring or partnering with product and design, even though those peer relationships often cross org boundaries and shape the technical roadmap. The TPM version is the same observation with a different job title: the people who decide whether your program lands are largely people you cannot instruct.
The equation, term by term. Maister, Green and Galford define trust as a ratio. In the version most people know, trust rises with credibility (do they believe you know what you are talking about), reliability (do you do what you say, repeatedly), and intimacy (is it safe to tell you the real problem), and falls with self-orientation (do they believe you are mostly serving yourself). Fiona Siseman and Alia Rose Connor of Frank & Eddy Leadership use exactly this equation in their LeadDev piece on managing up, as a way to assess where a relationship with a manager is strong or weak. The same lens transfers cleanly to peers, which is the leap this issue is making.
Here is why the structure matters. The numerator terms are additive. A peer can find you credible and reliable and still not tell you what is actually blocking them. The denominator is different: it divides everything. If a peer suspects your agenda is your milestone rather than the company's outcome, high credibility makes it worse, not better, because a competent person pursuing a private agenda is more dangerous than an incompetent one.
Where TPMs specifically leak trust.
Self-orientation, by process. The TPM's job description invites suspicion. You track other people's commitments, you report status upward, you hold the dates. To a peer director, a dependency tracker can look like a surveillance instrument pointed at their team. Every status ask that serves your dashboard rather than their decision raises the denominator. The test is whether your last three requests to this peer were things they needed or things you needed.
Reliability, by small defaults. Reliability is built in unglamorous places: the pre-read that arrives when you said it would, the "I will confirm by Friday" that you confirm on Thursday, the meeting that ends when scheduled. Peers who have been burned by a TPM who over-promised on their behalf will discount everything else you do.
Intimacy, by the bad news channel. Intimacy in this context is not friendship. It is whether a peer will tell you early that a date is going to slip. You build it by how you react the first time they do. If the response is an escalation, you have just taught every peer in the room to tell you later.
The sideways audit. Pick five peers who can stop or slow your program. For each, score the four terms from one to five, honestly, on a single page. Then write one sentence per peer: "The weakest term is X, because of Y." Do not average them. The point is to find the one term that is dragging each relationship, because the interventions differ completely. A credibility gap is closed by doing something visibly hard and correct, for instance a sharp technical or domain read in their area. A reliability gap is closed by a boring run of kept commitments. An intimacy gap is closed by asking a real question and not using the answer against them. A self-orientation gap is closed by doing something for their outcome that costs your milestone, at least once, visibly.
That last one is the expensive move and the one most leaders avoid. It is also the only move that lowers the denominator. Will Larson, in the description of his 2024 LeadDev talk on policy, consensus and conviction, says he "routinely stepped on my own foot" by using a single leadership style in every situation. The same failure shows up sideways: running the one relationship pattern that has worked with your sponsor on peers who need something different.
What this is not. It is not a manipulation checklist, and peers will smell one. If you score yourself at 2 on self-orientation for a given peer, the honest question is whether your program is actually hurting their team, not how to look less self-interested. Sometimes the fix is renegotiating the dependency, not repairing the relationship. Siseman and Connor make the related point that staying silent to avoid an awkward conversation usually costs more over time. The same holds sideways.
Try this week. Pick the one peer who most often slows your program. Spend ten minutes scoring the four terms, then identify the single weakest one. Before your next interaction, do one thing aimed squarely at that term: if it is self-orientation, ask what they need from you this quarter and deliver one piece of it ahead of your own ask.
What it is. A diagnostic that treats trust as credibility, reliability and intimacy divided by self-orientation, so you can find the specific weak term in a relationship instead of "improving rapport."
When to use it. Before a program kickoff that depends on peers you do not know well, when a dependency owner keeps slipping, or when a peer relationship has gone cold after a disagreement.
How to run it:
When NOT to use it. Do not run it as a one-time rating you share with peers, or as a substitute for fixing a structural conflict of incentives between teams.
Note on provenance: the four components are as defined in the book and in the LeadDev write-up cited above. The 1 to 5 scoring and six-week re-score are a practical adaptation, not part of the original.
Managing up is not doing your boss's job for them (Siseman and Connor, LeadDev) — November 2024. Uses the trust equation to assess a manager relationship and argues silence about a weak one costs more than the awkward conversation. Read it, then point the same lens sideways.
The Manager's Path in the age of AI (Camille Fournier, LeadDev) — Fournier's LDX3 New York 2026 talk, published September 29, 2026, on which parts of the management job AI actually changes and which skills stay essential. The page is a description and the full talk needs a free account, so treat it as a pointer, not a summary.
Beyond Senior: consider the staff path (Joel Hawksley) — A GitHub staff engineer's write-up using Larson's four archetypes. Useful for the line about reducing ambiguity and then handing the work off, which is the TPM job in one sentence.
"...routinely stepped on my own foot using that style for every situation."
— Will Larson, as quoted in the description of his LeadDev talk on policy, consensus and conviction
Don't miss what's next. Subscribe to Critical Path: