There's a special place in user experience hell for the exit modal. You know the one—you're trying to leave a website, you move your cursor toward the browser's close button or back arrow, and suddenly: BAM. A modal appears, blocking your escape, demanding one more moment of your attention.
"Wait! Before you go..."
"Are you sure you want to leave?"
"One last offer just for you!"
At Synacor, working on AT&T's start.att.net portal, we had a particularly lucrative version of this dark pattern: an exit modal that played a video ad.
It made tremendous numbers in testing. The product team was convinced users would hate it. And navigating the tension between these two facts taught me more about the actual role of product management than any "PM as CEO of the Product" blog post ever could.
What Is an Exit Modal (And Why Does It Exist)?
For the blissfully uninitiated, an exit modal is exactly what it sounds like: a modal dialog that appears when a user attempts to leave your site.
The technical implementation is straightforward. JavaScript detects when the user's cursor moves toward the browser chrome (the back button, the tab close X, the address bar) and triggers a modal overlay before they can complete the action.
The business rationale is equally straightforward: you're about to lose this user anyway. Their session is ending. You might as well try to extract one more impression, one more conversion opportunity, one more ad view before they go.
In our case, that "one more thing" was an ad.
The CPMs were very high, so it allowed monetization of users that would otherwise just use our homepage as a place to get their email, resisting our other charms.
From a pure revenue optimization perspective, it was a no-brainer.
From a user experience perspective, it was questionable.
The Business Case: Money Talks
Let me be honest about the incentives, because understanding them is crucial to understanding the tension.
The revenue from exit modals was significant. It was a material impact to the bottom line on a portal that saw most of its traffic come from DSL subscriptions, which were under pressure from cable and fiber.
When you have millions of daily users, and even a fraction of them trigger an exit modal, and if each impression generates $3-5 CPM, the math gets real impressive real fast. We're talking hundreds of thousands of dollars annually—potentially more—from a single feature.
For a product that was already serving ads throughout the experience, the exit modal represented pure incremental revenue. We weren't displacing other ad inventory. We were capturing value from a moment that would otherwise generate zero revenue.
The business stakeholders loved it. Sales could show advertisers an additional high-value placement. Finance could count on predictable incremental revenue. Leadership could point to innovation in monetization strategy.
And technically, it worked. The implementation was solid.
From every business metric that mattered, the exit modal would have been a success.
The User Experience Reality: Rage and Resentment
Our product team despised it. And so did users, or so we thought they must. But how do you prove that?
We were able to look at how often people came back to our site after seeing the exit modal. It was abysmal, a final straw for people that at best tolerated us.
The PM's Actual Job: User Advocate, Not Revenue Optimizer
Here's where we get to the heart of what product management actually is.
There's a pervasive narrative in tech that the PM is the "CEO of the Product." It's a seductive metaphor—empowered, strategic, decisive, in control. Newly minted MBAs must find this appealing given how many of them try to snag PM jobs after graduation.
It's also mostly bullshit.
A CEO controls capital allocation, hiring, strategic direction, and ultimate decision-making authority. A product manager controls... well, mostly PowerPoint presentations and Jira tickets. We influence decisions, but we don't make them unilaterally. We don't have a PnL, we don't manage sales and marketing, we don't even manage the engineers on our team (well we shouldn't - if the PM runs ENG, that's a red flag).
More importantly, the CEO metaphor fundamentally misunderstands the PM's most critical function: being the voice of the user inside the organization.
In any company with revenue pressure (which is all of them), there are plenty of people advocating for business metrics:
- Sales wants higher ad inventory and better placement
- Finance wants predictable revenue growth
- Leadership wants to hit quarterly targets
- Advertisers want guaranteed impressions and completion rates
Who's advocating for the user experience? Some CEOs are product guys at heart, or take the long view, or value the brand highly enough to protect it ferociously. But all CEOs face a ton of pressure from all sides all the time, and have to chart a course through all of that for short term profitability and long term growth.
So customer advocate is the PM's actual job. Not to be the CEO. Not to maximize revenue. But to be the person in the room who says:
"Yes, this makes money. But it also makes users hate us. And that's a problem."
The Exit Modal Debate: Revenue vs. Advocacy
The debates around the exit modal were intense.
The revenue side argued:
- "Users are leaving anyway—we're not disrupting an active session"
- "The revenue is substantial and incremental"
- "Other sites do this all the time"
- "Video completion rates prove users are watching the ads"
- "If we don't maximize revenue, we can't invest in better content"
The user advocacy side (which included me) argued:
- "Users explicitly chose to leave—blocking that is hostile"
- "The last impression shapes their overall perception of our brand"
- "High completion rates don't mean users are happy—they just mean they can't escape"
- "This could damage AT&T's brand relationship with customers"
- "Short-term revenue gains create long-term retention problems"
Notice something about those positions? Both sides had legitimate points.
The revenue argument wasn't wrong. We did need sustainable monetization. The incremental revenue was significant. Users were leaving anyway.
But the user advocacy argument was equally valid. We were creating a hostile experience. It did damage brand perception. Users did resent it. At least, they weren't coming back as often once they experienced the exit modal. It was burning long term revenue streams for a short term burst.
This is the actual job of product management: sitting in that tension, understanding both perspectives, and trying to find a path that serves the business without betraying users.
The Compromises and Constraints
In a perfect world, user advocacy would always win. We'd build delightful experiences, never compromise on user trust, and everyone would be happy.
In the real world, companies need revenue to survive. Synacor needed to deliver returns to AT&T. AT&T needed the portal to be financially sustainable. Advertisers needed inventory to fill.
So we couldn't just kill the exit modal entirely, even though from a pure UX perspective, that was clearly the right answer.
Instead, we did our homework.
A/B testing impact on retention: We tried to quantify the long-term cost of the exit modal by measuring whether users exposed to it had different retention rates. (They did.)
Were these compromises satisfying? No. Did they eliminate the fundamental tension? Absolutely not. But they represented the PM trying to move the needle toward user experience within business constraints.
That's the actual job.
What "User Advocate" Really Means
Being a user advocate doesn't mean you always win. It doesn't mean every business decision goes your way. It doesn't mean you can veto revenue-generating features because they're not perfectly delightful.
It means:
You bring the user perspective into every decision. Even when it's inconvenient for business goals.
You quantify the user experience cost. Not just "this feels bad" but "here's how this impacts retention, brand perception, and lifetime value."
You push for the least-bad compromise. If you can't kill the dark pattern entirely, can you make it less dark? Less frequent? Less intrusive?
You document the trade-offs. Make sure everyone understands what you're sacrificing for the revenue. Don't let the organization pretend there's no cost.
You build the case for long-term thinking. Help leadership understand that maximizing this quarter's revenue might damage next year's retention.
You fight the battles you can win and accept the ones you can't. Save your political capital for the fights that matter most.
In this case, we won in the short term. Once we were informed that our contract was up with AT&T though, the exit modal came back. There was no long term, so, with AT&T's blessing, we maximized revenue now.
That's user advocacy in the real world.
The "CEO of the Product" Problem
The "PM as CEO" framing is not just wrong—it's actively harmful because it misrepresents the power dynamics and obscures the PM's actual value.
CEOs make final decisions. PMs influence decisions that other people make.
CEOs control resources. PMs negotiate for resources they don't control.
CEOs set strategy. PMs advocate for strategic directions within constraints set by actual leadership.
CEOs optimize for shareholders. PMs should optimize for users (within business viability constraints).
Everyone thinks they can do product management. A lot of the job is navigating choices that are obvious to their proponents and letting down people if not easily than rationally and diplomatically.
When you tell PMs they're the "CEO of the Product," you set them up to think they should be maximizing revenue, optimizing for business metrics, and making hard trade-offs against user experience in service of growth.
But we already have plenty of people doing that! Finance does that. Sales does that. Leadership does that.
The organization needs someone to say: "But what about the users?"
That's the PM. A user advocate with business acumen.
The Long-Term Cost of Short-Term Revenue
Here's the thing about the exit modal revenue: it was real, but it wasn't free.
Every user who encountered that exit modal had their perception of AT&T and start.att.net slightly damaged. Most didn't leave immediately. Most didn't write angry emails. But the resentment accumulated.
It showed up in subtle ways:
- Slightly lower return visit rates for users exposed to exit modals
- More negative sentiment in qualitative feedback
- Higher support burden from confused and frustrated users
- Brand perception damage that's hard to quantify but very real
This is the fundamental tension in product management: the revenue is immediate and measurable. The trust damage is delayed and diffuse.
Business stakeholders naturally weight the certain, measurable revenue more heavily than the uncertain, long-term cost. It's the PM's job to keep bringing up that long-term cost, quantifying it where possible, and pushing for decisions that preserve user trust even when it costs revenue.
Sometimes you lose those fights. You advocate, you present data, you make the case—and the business decides the revenue is worth the trust cost.
That doesn't mean you failed as a PM. It means you did your job by ensuring the decision was made with full information, not because no one thought about the user experience cost.
Then you disagree and commit. Or quit.
Conclusion: Advocacy, Not Authority
The exit modal playing ads when users tried to leave start.att.net was a dark pattern. It generated significant revenue and damaged user trust. Both things can be true simultaneously.
As a product manager, my job wasn't to be the "CEO" who optimized for revenue. It was to be the advocate who made sure the user experience cost was understood, measured, and weighed against the business benefit.
Sometimes advocacy wins and the right thing happens. Sometimes advocacy loses and the business priorities prevail. But the presence of that advocacy voice in the room—the person who keeps asking "but what about the users?"—is essential.
That's the actual role of the PM. Not CEO. Not revenue maximizer. Not growth hacker.
User advocate with business acumen.
The exit modal taught me that lesson more clearly than anything else in my career. I hated that feature. I fought against it repeatedly. I lost in the end.
But if the PM doesn't do that, who will?
Have you dealt with similar tensions between revenue and user experience? How do you think about the PM's role as user advocate? I'd love to hear your perspectives.