"Great, thanks for the support" — then go push it alone. "Come find me" is textbook abdication: responsibility slides back to you while looking exactly like backing, so it takes three months before you notice nobody was ever holding it.
"Thank you. Let me break 'support' into three specific items — tell me which ones you'll take:
① Voice: at the Aug 5 all-hands, you (not me) announce this as one of three H2 must-wins, and name the cost of not doing it;
② Signature: the kickoff email to the 5 team leads goes out under your name — I'll draft it;
③ Arbitration: 15 minutes every two weeks, only for cross-team conflicts I can't move, first one Aug 12."
Compressing open-ended goodwill into dated, signed actions is the line between sponsorship that lands and sponsorship that doesn't.
"Then I'll re-plan on 'voice, no arbitration': cross-team conflict resolution goes from 3 days to 3 weeks, and the Q3 milestone moves out two weeks." Not a threat — it puts the cost of the sponsorship gap on the record now. Leave it off the record and it eventually gets recorded as your delivery problem.
Action: write your most important change effort into a one-page "sponsor schedule": three items, three dates, signatory fixed. Send it this week so all he has to do is answer yes/no.
Reflection: on what you're driving right now — if your sponsor did nothing from tomorrow on, how long before anyone noticed?
"We're running a bit behind — it might help if you reinforced the importance again." Vague, and it hands the judgment call back to him. "Sure, I'll mention it next time" disposes of you, and you still feel like you escalated.
"Two blockers, each needs one decision from you. Team A has sequenced migration behind features — I need you to state the ordering at next week's staff meeting; one sentence is enough. Team B's tech lead believes the design carries real technical risk — that isn't resistance, it's a genuine problem — I need you to approve a two-week spike. The first is yours to own, the second is mine; I'm reporting them separately."
Three 15-minute slots at weeks 6, 12 and 20, with a fixed title: "One decision needed from you this cycle." The agenda is always exactly one item; if there's no decision, cancel. Once status updates have eaten the sponsor's bandwidth, it's not there when you actually need arbitration.
Action: open your calendar and pre-book weeks 6, 12 and 20 sponsor syncs for your most important effort.
Reflection: if your sponsor had only three appearances left, which three weeks would you spend them on — and why not kickoff?
"This is a company H2 strategic priority." Two consequences of leaning on authority with a peer: he agrees on the surface and assigns his weakest person; and you take on a debt you'll have to repay later. More importantly — he isn't lying. It genuinely isn't in his Q3 goals.
"Let's leave priority aside — I want to check one thing: in your Q3 review, which line does this migration count toward?"
If he says "it doesn't" — "Then you're right to block me; I'd block me too. So I shouldn't be asking you for people. Two paths: I go to Sandy and get it written into your Q3 goals, or we carve the adaptation out as a deliverable counted on my side. Which do you want?"
"Sandy: the migration needs roughly 2 person-months from Platform, and it isn't in Ken's Q3 goals — so this isn't a cooperation problem, it's a goals problem. Two fixes: (a) add this line to Platform's Q3 goals; (b) we absorb the work and Ken only reviews. I recommend (a), because (b) makes adaptation quality depend on code we don't know well. Your call."
The cc is the dividing line: out in the open, you're solving a problem; behind his back, you're manufacturing an enemy.
Action: pick a peer you've filed as "immovable," look up his review metrics for this quarter, and find which one your ask lands in. If none, stop mobilizing and go change the goal.
Reflection: are you currently blocking something "reasonably" yourself — because it was never written into your goals?
"Understood, we'll do our best." The result is three efforts at 40% each: nothing lands by year-end, and what the team remembers is "we churned for a year and nothing stuck" — which doubles the startup cost of the next change.
"I agree with all three, but let me state the absorption ceiling first: my team can take exactly one change in Q3 that asks everyone to alter daily behavior. The other two either move to Q4 or get scoped down to a 2-person pilot. My proposed order: migration (external dependencies, least deferrable) → oncall (2-person pilot first) → performance process (naturally aligned to the Q4 cycle). You can change the order and I'll follow it — but I won't agree to all three in parallel, because that means none of them happen."
"In Q3 there's exactly one thing you actually have to change: the migration. I've got the other two moved to Q4 or scoped to a pilot. If anyone comes to you asking for those two, send them to me." — Blocking change traffic for your team and winning executive airtime for your project are the same job, pointed in two directions.
Action: list every change running on your team in Q3, including the ones other functions started. More than two — take the list to your boss and ask for an ordering.
Reflection: what was the last change on your team that finished and stayed? What was structurally different about it?