Technical Program Manager

Tell me about a time you faced significant resistance or conflict from a team that had no reporting relationship to you, and how you resolved it.

Also asked as: Describe a situation where you had to align two teams with fundamentally different, competing incentives. · Tell me about a time you had to influence a team without having any direct authority over them. · Walk me through a conflict where trust had broken down before you even started — how did you rebuild it? · Give an example of resolving a conflict where the other side wasn't just skeptical, but actively opposed to your initiative.

Motivating TeamsConflict ResolutionCross-Functional CollaborationInfluencing Without AuthorityDriving ExecutionRelationship Building

Opening Statement (~60 sec)

As the Technical Program Manager driving new ad-product incubation across Amazon's streaming portfolio, I owned the launch of pre-roll video ads on Amazon Freevee — Amazon's free, ad-supported streaming service with 30M+ monthly viewers. The revenue case was clear: pre-roll CPMs ran 30–50% higher than the mid-roll-only model Freevee was running. The conflict was just as clear: the Freevee Content and Product teams, who reported into a completely separate VP org I had no authority over, saw a pre-roll unit as a direct threat to years of UX investment — backed by their own data showing higher ad load increased session drop-off.

I earned alignment instead of mandating it: a two-week stakeholder discovery sprint before proposing anything, a reframed joint problem statement tying ad revenue to Freevee's ability to keep licensing premium content, and a five-part guardrail framework — including an automated kill-switch the Content team could trigger without needing my sign-off.

The result: we launched to 100% of U.S. Freevee traffic within six weeks of the pilot, ad revenue per viewer hour rose ~38%, and session-start rate moved only -0.2% against an 8% guardrail. The Freevee VP — our most vocal critic going in — became the program's public advocate, and the guardrail-and-tiger-team model I built became the template for two subsequent streaming ad launches.

Situation

1. In late 2022, Amazon Ads identified a strategic revenue gap: Amazon Freevee, Amazon's free ad-supported streaming service (FAST), had grown to 30M+ monthly active viewers in the U.S. but ran ads exclusively as mid-roll breaks, which carried lower CPMs and shallower inventory depth than industry-standard pre-roll formats.

2. Amazon Ads was under pressure to diversify streaming ad revenue amid a broader slowdown in display and sponsored-product CPMs, and industry data showed pre-roll ads commanding 30–50% higher CPMs than mid-roll on competing platforms like Hulu, Peacock, and Tubi.

3. The Freevee Content and Product teams optimized for a fundamentally different metric — viewer engagement, session depth, and retention — and had internal engagement models showing ad load above a threshold historically correlated with session abandonment.

4. Adding a pre-roll unit that every viewer would hit before every piece of content read, to that team, as a direct threat to years of UX work. This wasn't a data disagreement — it was two orgs with genuinely competing North Star metrics.

Task

1. Senior leadership in Amazon Ads asked me to own the end-to-end program for designing, aligning, and launching the pre-roll ad unit on Freevee, with a target launch inside two quarters — a change-management and organizational-alignment challenge as much as a technical one.

2. My mandate was threefold: build genuine cross-org alignment (buy-in, not compliance) with the Freevee Content, Product, and UX teams; protect UX quality by designing an implementation that could address their engagement concerns directly; and deliver revenue impact on a tight timeline with success metrics both orgs would accept.

3. Critically, I had no direct authority over the Freevee Content or Product teams — they reported into a completely separate VP organization. My leverage was entirely relational and data-driven; I had to earn alignment, not mandate it.

Action

1.Ran a two-week stakeholder discovery sprint before proposing anything — one-on-one sessions with Freevee Directors, PMs, and data scientists, arriving with questions instead of a pitch deck. Their resistance turned out to be rooted in real internal data on session abandonment, and openly acknowledging that data — instead of dismissing it — immediately changed the dynamic from adversarial to collaborative.
2.Reframed the narrative with a jointly authored problem statement, shifting the framing from 'we want to add pre-rolls' (heard as 'we want to degrade your product') to a shared reality: Freevee's ability to keep licensing premium content depended on stronger ad revenue, and cited Tubi, Pluto TV, and Peacock as proof pre-rolls didn't have to cost viewer satisfaction.
3.Converted their objections into binding, measurable guardrails in the PRD and program charter instead of asking for trust — a 15-second duration cap, a one-pre-roll-per-session frequency limit, an A/B-tested skip option, a weekly joint KPI review with Freevee holding escalation veto, and an automated kill-switch that paused all pre-rolls if the 7-day rolling session-start rate dropped more than 8%, with no human approval required to trigger it.
4.Built a cross-functional tiger team — Ads Engineering, Freevee Product, UX, and Data Science on one shared channel and JIRA board — and proposed a time-boxed 4-week pilot at 5% of U.S. traffic with pre-committed exit criteria, so the Content team was agreeing to a controlled experiment, not a launch decision.
5.Surfaced problems before they were escalated to me — when pilot data showed a 3.4% spike in exit-before-play events specifically in reality TV, I called a cross-org review within 24 hours, shared the raw numbers, and proposed excluding that genre pending investigation, rather than waiting to be asked.
6.Separated legitimate concerns from scope creep when a Freevee PM, emboldened by that concession, requested expanding the guardrails to exclude three more content categories and shrink the frequency cap — changes that would have cut inventory by an estimated 40% and undercut the revenue thesis. I fast-tracked the one defensible piece (children's content) and held the broader ask to a scheduled post-launch review instead of either capitulating or flatly refusing.

Result

1. Pre-roll ads launched to 100% of U.S. Freevee traffic within six weeks of the pilot start, on schedule with the original two-quarter target.

2. Ad revenue per viewer hour rose approximately 38% in the first full quarter post-launch, and the unit became one of the top-3 highest-CPM placements across Amazon's streaming ad portfolio.

3. Session-start rate moved only -0.2%, well inside the 8% kill-switch threshold — proof the guardrail framework held under real traffic — and contextual, genre-matched creatives cut skip rates 22% below the mid-roll network average.

4. The Freevee VP, the program's most vocal critic at the outset, publicly advocated for it at an all-hands as the model for how Ads and Content should collaborate, and the guardrail framework and tiger-team structure were adopted as the template for two subsequent streaming ad launches.

Closing Statement (~60 sec)

What made this a conflict-resolution problem rather than a straightforward launch was that I had zero formal authority to force the outcome — every guardrail, every pilot gate, and every scope boundary had to be something the Content team chose to accept, not something I could mandate. Engineering could tell me whether the kill-switch was technically sound; it couldn't tell me how to turn a team that saw us as an existential threat to their product into one that co-owned the rollout.

The pattern I'd bring here is the same one that worked at Amazon: listen and acknowledge the other side's data before proposing anything, turn objections into binding, mutually verifiable commitments instead of asking for trust, and when a resolved conflict resurfaces as a bigger ask, separate what's legitimate from what's overreach and give a path forward instead of a flat no.

Follow-Up Questions

Two challenges determined whether this program could get off the ground at all: earning genuine buy-in from a team with zero reporting relationship to me and real data behind their resistance, and designing guardrails specific enough to earn trust without gutting the revenue thesis the program existed to deliver.

What were the biggest challenges you faced launching this program?

#	Challenge
1	No authority and a team with genuine, data-backed reasons to oppose the initiative — the Freevee Content and Product teams reported into a separate VP org, and their own engagement models showed real risk in raising ad load.
2	Designing guardrails concrete enough to earn trust (duration caps, frequency limits, an automated kill-switch) without so constraining the ad unit that it failed to close the CPM gap the program was built to capture.

How did you resolve the lack of trust and buy-in?

Step	Action
1	Ran a two-week, one-on-one discovery sprint before proposing any solution, arriving with questions instead of a pitch.
2	Openly acknowledged their internal session-abandonment data rather than disputing it, which shifted the dynamic from adversarial to collaborative.
3	Co-authored a joint problem statement reframing the initiative from "ads vs. users" to "sustainable platform investment," backed by competitive benchmarks from Tubi, Pluto TV, and Peacock.
4	Encoded their objections as binding, measurable guardrails in the PRD and program charter instead of asking them to simply trust the Ads org.

How did you resolve the risk of over-constraining the guardrails?

Step	Action
1	Ran a contained 4-week pilot at 5% of U.S. traffic with pre-agreed exit criteria before committing to any scale-up.
2	Used pilot data itself (the reality-TV exit spike) to make a targeted, genre-specific exclusion rather than a blanket guardrail tightening.
3	When a follow-on request threatened to cut ~40% of inventory, separated the one legitimate concern from the broader overreach and resolved only that piece immediately.
4	Deferred the larger guardrail-scope question to a scheduled post-launch review backed by 60 days of full-traffic data, instead of relitigating it mid-pilot.