04 How Atmosphere makes or breaks user trust and retention
Atmosphere in action: Hopper and the Las Vegas effect
It was 11pm and I’d just landed at Heathrow Airport — exhausted and in need of a hotel before my connection the next morning. I was interviewing for a product lead role at Hopper at the time, and thought: what better way to get to know the product than to use it myself?
In that moment, I needed reassurance that I was getting a comfortable room, quickly, and without hidden surprises. Instead, what I experienced became a case study in how Atmosphere, the first dimension of ARCH, shapes whether users feel safe enough to buy a product — and return to buy again.
Case study: Hopper app (2023-2025)
Take a moment to scan these two screenshots from 2023 — the home screen and the hotel detail view. What do you notice first? Where does your gaze travel?
What Hopper got right
The app’s bold colors and soft edges made it feel trendy and tech-forward, which built trust with young travelers. The data-backed price prediction and “wait or buy” recommendations used calm copy that built confidence rather than pressure, which made them more effective conversion levers.
Animations, illustrations, and personable copy also softened the stress of booking with a boost of playfulness. Loading and success states became moments of reassurance and celebration rather than dead time, helping the app feel more human and rewarding.
The Las Vegas effect
Their branding was top-notch; the problem was their Atmosphere. Using Hopper felt like walking through a loud, crowded casino in the heart of Las Vegas. I was overstimulated, on high alert, and looking for a way out.
High visual density was overwhelming. With so many competing calls-to-action, I didn’t know where to look. Nearly half the home, search, and detail screens promoted upsells — price freeze, insurance, rewards, “Secret Deals”. Primary and secondary actions had no clear hierarchy, and too many badges and colors further made everything blur together. One user I interviewed hadn’t even noticed the “View Rooms” button because the price freeze banners were so dominant.
Inconsistent interaction patterns eroded confidence. Core actions changed styling and behavior screen-to-screen. Opt-in/out actions sometimes appeared as radio buttons, other times toggles or buttons. Primary actions shifted between full-width buttons, half-width buttons, links, and hidden states that only appeared after a selection was made. Each variation forced me to pause, second-guess, and relearn the interface mid-flow.
Dark patterns trained distrust. A banner assured me prices were final, but the total jumped at checkout. The “Secret Deals” CTA implied the platform wasn’t showing me its best prices upfront. The insurance opt-out was hidden below the fold; the only visible path forward was to accept. A tip was auto-enabled just before payment.
Unclear structure caused disorientation. I couldn’t tell where I was in the product. Multiple navigation menus meant there was no obvious starting point, and five nearly identical insurance upsell screens appeared back-to-back appear in the booking flow - I wasn’t sure if I was moving forward or looping.
Interruptions derailed the flow. Pop-ups disrupted my focus mid-task and multiple backdoors pulled me out of the booking flow. The room selection screen had both an “Unlock 15% off” banner that led to an unrelated support chatbot and a push notification prompt that exited the app entirely. Even without the detours, getting from room selection to payment required about ten taps (a competitor did it in two).
Context-blindness felt transactional. Hopper knew I was booking a same-day hotel at 11pm, yet showed me the full upsell stack anyway. Trip cancellation insurance and price freezing were irrelevant in my situation, and showing them anyway only added frustration.
With each interaction, the app was training me to distrust it. I was tired and time-crunched. I expected to feel taken care of — not on edge, suspicious, and cognitively overloaded. I completed the booking, but only because I was doing research for the interview. The experience eroded my trust in Hopper and that feeling stuck. I haven’t used the app since.
Shipping the org chart
Many of these friction points trace back to how Hopper was organized. Teams were split into self-sufficient pods to maximize speed and autonomy, but sacrificed coherence in the process.
Through interviews, I learned that revenue was Hopper’s north star, not user outcomes, and that team funding and performance reviews were tied to how much money each feature generated. Their culture was intensely data-driven, and the absence of shared metrics or attention to harder-to-measure things like trust and anxiety created blindspots. Teams competed for real estate and user attention to move their own numbers alone, which is how you end up with four back-to-back insurance upsells and too many banners and badges on each screen.
This is what ‘shipping the org chart’ looks like. No one owned the end-to-end experience, and customer lifetime value suffered as a result.
Three lessons from Hopper
1. Atmosphere is one of the most underinvested retention levers.
Atmosphere is a business asset. The emotional memory it leaves behind and the trust it builds or breaks shapes how users feel about a product long after they use it. A single negative experience may not always register as churn in the moment, but feelings are sticky. A user who feels anxious or manipulated may still complete the transaction, but they won’t come back. It’s a monetization win masking a retention problem.
Atmosphere can be both a positive and a negative business lever — depending on how it serves the task at hand. A gaming app built on intensity and overstimulation retains users because that’s the experience they came for. Hopper engineered the same intensity in a product where users needed reassurance and it built distrust instead. Same lever, opposite impact on customer lifetime value.
2. Atmosphere is the emotional output of every signal a product sends, not just aesthetics.
An app can be visually beautiful and still feel incoherent. Atmosphere isn’t only shaped by visuals or branding — it’s the emotional sum of every layer of the product experience. How a product directs attention and surfaces actions (Cue) shapes whether users feel confident or uneasy about a decision. How it orients users across screens and within flows (Route) shapes whether they feel in control or lost. Designing it well means examining the emotional impact of every layer. Hopper’s brand was great, but the Atmosphere it created
3. Atmosphere must meet users where they are
We have more data available to us than we act on. A same-day, late-night booking signals urgency, fatigue, and stress. Browsing flights months in advance signals openness and exploration. This might be the same user in different moments — and they shouldn’t encounter the same Atmosphere.
The right emotional conditions aren’t fixed — they’re contextual. Hopper knew I was checking in within the hour and showed me the full upsell stack regardless. It’s a signal to the user that the product either isn’t paying attention or doesn’t care. Instead, Atmosphere should account for the emotional state users arrive in.
Hopper’s winter redesign
Hopper’s recent redesign started moving the interface toward a calmer baseline. The home screen is more focused. There is more white space, a clearer hierarchy that guides eye movement linearly, and a single obvious starting point (the search bar). Navigation is simplified and affordances are more standardized.
As a whole, the product feels less like it’s competing with itself for attention. There are still rough edges — excessive pop-ups, dead-ends, price jumps at checkout, misused UI affordances — but the Atmosphere has clearly shifted toward more confidence and control. For a product whose brand promise is to save users money, this shift isn’t just good UX; it’s core to growing the business.
Coming up next: Route
ARCH gives us a shared vocabulary for these kinds of conversations: not just that a product feels “off,” but why, and what to do about it. Next: Route — how products guide users through space, and what goes wrong when the path isn’t clear.
I wrote a deeper case study on Hopper’s UI design in 2023, which helped me land the offer. If you’d like to read it, drop a comment below and I’ll send it over.
This one genuinely moved me.
Not because of the Hopper analysis, impressive as it is. But because of what it represents: a level of thinking that no AI can replicate.
AI can write copy, generate UI patterns, optimize flows. But it works in templates. When you want to take your product to the next level, you need someone who feels the experience, names what's wrong with it, and builds a new vocabulary around it.
That's exactly what you did here with Atmosphere. You shed light on something that was always there but impossible to articulate.
This is the work that actually moves products forward.
Would love to see the version of Hopper's homepage if you had taken that job ;)