The message that changed how we built FlashSeat

The problem we didn't know we had

Sarah wasn't alone. Once we started paying attention, we saw the pattern everywhere. Users would load FlashSeat, see a flight to Barcelona for £34 or a sold-out Coldplay ticket suddenly appearing at face value, get excited, and then... nothing. They'd close the app. Twenty minutes later, the deal would vanish. And they'd miss it.

The traditional model of deal hunting assumes you're glued to your phone. You're not. You're at work. You're in a meeting. You're having dinner with friends. Real life happens. We'd built a platform that surfaced brilliant deals, but we'd built it for a version of humans that doesn't exist. The ones who refresh their browser every five minutes.

So we asked ourselves a harder question: what if we could catch deals on behalf of our users? What if you could tell FlashSeat exactly what you wanted, and we'd do the watching for you?

Saved searches felt obvious, but it wasn't

In theory, saved searches are straightforward. You set parameters. The system monitors. When something matches, you get a notification. We'd seen it done before in other spaces. But this wasn't just about flights to Barcelona on any Tuesday in June. Our deals move in hours, sometimes minutes. A saved search that sends you an alert three hours after a deal posts is useless.

The challenge was speed. We needed our systems to scan incoming deals, cross-reference them against what thousands of users had saved, and push real-time alerts without becoming a spam machine. We also needed to respect people's attention. One of our early prototypes would have alerted users to every tangential match. We scrapped it after testing with a few power users; they uninstalled within a week.

We settled on a tight logic: only alert when there's a genuine match. Only send during reasonable hours unless a user opts into 24-hour notifications. Make it easy to adjust what you're watching. We quietly launched it in the Premium tier first, then to Pro members, and watched how people actually used it.

What people actually do with deal alerts

Once saved searches went live, something unexpected happened. Users started saving very specific things. Not "flights under £50 anywhere in Europe". More like "Manchester to Malaga, departing Thursday evening, under £45". Or "any Stereophonics date within 100 miles of Liverpool". They were treating it like a hunting rifle, not a shotgun.

The data told us something important: people don't want to be overwhelmed with options. They want to be notified when something they actually want becomes available at a price they actually want to pay. It sounds obvious, but it changes everything about how you design the feature. Every notification is a moment of trust between us and the user. Send the wrong one, and you've burned that trust.

We also noticed that saved searches changed user behaviour in subtle ways. Members would set up a search, go about their day, and when an alert came through, they'd act on it immediately. The conversion rate from alert to booking was significantly higher than from browsing. The friction was lower because they'd already decided what they wanted; they just needed permission to pull the trigger. A good deal alert gives them that permission.

The integration nobody saw but everybody felt

Behind the scenes, saved searches connect to everything else in the app. When you get an alert, you tap it and land directly at the deal. Three steps from there and you're checking out with your QR ticket already generated. For Premium members, that checkout fee is 4 to 5 percent instead of 8 to 10 percent. For Pro members, there's no buyer fee at all. The deal alert doesn't just notify you; it puts you in the fastest possible path to actually claiming the deal.

We spent weeks on something most people will never think about: ensuring alerts land when they're most likely to be useful. Weekend deals get different timing than weekday ones. Theatre tickets behave differently from flights. We learned that people check their phones differently on Sunday morning than Thursday afternoon. The engineering felt invisible, which is exactly right.

Why this matters beyond the feature

Building saved searches taught us something fundamental about our audience. The 18-to-40 travellers and event fans we're serving aren't sitting around waiting for deals. They're living their lives. They have jobs, commitments, social calendars. A good deal app doesn't interrupt their life; it fits into it. It watches when they can't. It alerts when it matters. It gets out of the way the rest of the time.

That philosophy now runs through everything we build at MRVL. We don't assume users are addicted to their phones. We assume they're busy. We assume they have limited attention. And we assume that if we respect both of those things, they'll keep coming back because we've made their lives better, not because we've made ourselves impossible to ignore.