Escalations aren't dysfunction. They're evidence that people care enough to disagree. Clean escalations are open, objective, and fast. They accelerate decisions, deepen collaboration, and keep talented people engaged. We leveraged this framework at SendGrid, and I've reshaped it at every company since. The principles translate because the human patterns underneath are universal.
Two engineers can't agree on an approach and neither will budge. A PM and a tech lead have competing views on what to cut from a release. Or the quieter version: someone has a problem with a colleague, never raises it directly, and instead brings it to their manager's 1:1 every week until the manager is drowning in proxy conflict.
These are all escalations. Most happen sideways. And the sideways ones cause the most damage.
"Escalation" sounds adversarial, like something broke. That reflex to avoid it is actually the root of the problem.
Across every scaling organization I've worked in (SendGrid, Twilio, Contentful, now Siteimprove), the highest-performing teams aren't the ones that dodge disagreement or compromise their way to mediocre decisions. They're the ones with a shared method for surfacing disagreements and driving them to resolution.
We first formalized "clean escalations" at SendGrid during rapid growth. Cross-functional teams needed a way to resolve conflicts before they bottlenecked at the leadership level. Pairing escalation principles with the RAPID decision-making framework gave everyone a common vocabulary for who owns which decisions and what to do when people disagree. I've since adapted the approach at each company, adjusting decision roles for different team structures and shaping the norms to fit each culture. The framework endures because the human dynamics it addresses don't change.
Why escalations are actually great
When handled well, escalations reinforce the kind of environment top performers want to work in. They do three things:
They accelerate decisions. Your market isn't waiting for consensus. When disagreements sit unresolved for weeks, that's not rigor. That's drag. A clean escalation can collapse a multi-week stalemate into a single conversation.
They deepen collaboration. Real collaboration isn't finding the middle ground on everything. It's hearing different perspectives and making a sharp call when they conflict.
They protect engagement. When conflicts get addressed promptly, it lifts the emotional weight and keeps people focused. When disagreements drag on, people quietly check out.
Clean vs. dirty escalations
The question isn't whether your organization escalates. It does. The question is whether those escalations are productive or toxic. Here's the difference:
| Don't do this (Dirty Escalation) | Do this instead (Clean Escalation) |
|---|---|
| Don't duck the conversation. Ignoring opposing viewpoints and pressing ahead guarantees the conflict resurfaces later. Silence is not consent. | Engage directly first. Work it out with the other person assuming good intent. Give yourselves a real shot at resolution before pulling anyone else in. |
| Don't lob problems over the wall. Mentioning an issue in a status update and walking away so you can say "I flagged it" isn't decision-making. It's abdication. | Own it through resolution. Stay with the issue until someone makes a deliberate call. Even if that call is to accept the status quo and its risks. |
| Don't let it drag. Letting a disagreement linger for weeks, whether spinning in circles or waiting for alignment that won't come, is avoidance dressed up as patience. | Escalate early. When the conversation stalls or one side stops engaging, raise the escalation. The right time is when progress stops. |
| Don't weaponize it. Framing the escalation as a threat, or venting to your manager behind closed doors, contaminates the process before it starts. | Keep it open and unemotional. "I think we've hit a wall on this. I'd like to escalate so we can get a decision and keep moving." |
| Don't treat disagreement as failure. Smart people regularly arrive at different conclusions from the same information. What matters is that the process is rigorous, not that your view wins. | Advocate with conviction. You don't have to concede the moment someone pushes back. Bring your reasoning and your data so the organization can choose wisely. |
| Don't undermine the outcome. Griping about a decision or quietly working around it erodes trust across the entire organization. | Support the decision fully. If both sides had a fair opportunity to make their case, accept the outcome with humility and execute. You can always revisit later with new data. |
How to escalate: a practical framework
1. Get clear on the roles
Before escalating, align on the problem and who has a stake. I use an adapted version of the RAPID framework:
- Recommend: The person or small team with cross-functional visibility to synthesize inputs and propose a direction. This is where the bulk of the work gets done.
- Agree: One or two people whose sign-off is required. Cap it at two. More than that is a committee, not a decision process.
- Perform: The people who execute whatever gets decided. They often provide input during planning as well.
- Input: People with relevant knowledge that could shift the decision. Not every stakeholder needs a voice on every decision.
- Decide: One person who makes the final call. Only one D allowed.
When the Agree and Decide roles can't get aligned, that's your signal to escalate.
2. Find the right decision-maker
The default is for both parties to loop in their managers. That often works. The managers convene with both sides, weigh the evidence, and decide.
But managers aren't always the right audience. You want someone with enough context to understand the business impact but minimal stake in either position. For an architecture dispute, that might be a senior staff engineer. For a commercial tradeoff, a revenue leader.
Escalate to the lowest level where the decision will actually hold.
3. Frame the decision clearly
This is where escalations stall. Here's what keeps them moving:
- Write it down. Half a page covering the situation, the stakes, and what you need. If you can't make the case concisely, you probably need to sharpen your understanding first.
- Make the tradeoff explicit. Frame it concretely: "We need [team] to build [X], which takes [Y weeks] and delays [Z] by [W weeks]." This usually requires input from both sides.
- Have a real meeting. Walk out with either a decision or a next step with an owner and a deadline.
- Go higher if you must. But be honest: is this preference or a genuine mistake? Is it reversible?
4. Close it out
When the escalation meeting ends, the decision is made. There is no follow-up meeting to strategize a workaround. You're allowed to think the decision was wrong. You're not allowed to act like it was.
Speed matters
Most escalations should resolve in 24 to 72 hours. If yours routinely take longer, something in the process needs attention.
Why this matters for product leaders
Product leaders sit at the center of every cross-functional tension. Engineering bandwidth vs. customer commitments. Speed vs. quality. Near-term revenue vs. long-term investment. These tradeoffs never go away.
Without clean escalation culture, decisions default to whoever is most persistent, most vocal, or most politically connected. That's not strategy. That's organizational drift with a process label on it.
The most effective leaders I work with have one thing in common: they've normalized healthy disagreement and built systems that direct conflict toward decisions rather than away from them. Clean escalations are one of those systems.