Expedia PM interview: two-sided marketplace, metasearch, and the AI disruption question

Expedia Group PM interviews have a single most common rejection point: candidates who think of Expedia as a consumer travel app. The company operates a two-sided marketplace serving both travelers and 75,000 B2B partners through a platform that processes 21 billion API calls daily. Interviewers explicitly probe whether candidates understand that what is good for a traveler can directly conflict with what is viable for a hotel partner, and a product answer that ignores that tension misses the bar regardless of how clean the framework is. The second most common failure: treating AI’s impact on travel as a future hypothetical. It is not. It is a live product problem that some interviewers ask about directly.

The interview loop

Recruiter screen (30 min). Motivation and background. The recruiter is checking that you can speak specifically about Expedia Group’s business, not travel generically. Know the brand portfolio (Hotels.com, Vrbo, trivago, Orbitz, Travelocity) and be ready to say which product area interests you and why.

Hiring manager screen (45-60 min). Conversational product and behavioral. The HM will probe your experience with data-driven decisions and experimentation. Expedia runs thousands of A/B tests annually, and “experimentation fluency” is specifically flagged by interviewers as a differentiator. If you’ve never described a test you designed, owned, and drew a conclusion from, this round is where that gap surfaces.

Product case study (45-60 min). A core product sense round. You will get an open-ended product problem (improve search, design a feature, diagnose a metric drop) and will be expected to structure a response that names a specific user segment, identifies their real friction, grounds improvement in a metric Expedia would actually move, and acknowledges the two-sided constraint. More on what that looks like below.

Technical acumen round (45 min). Not a coding screen. This round tests whether you can engage credibly with engineering tradeoffs, understand API architecture (relevant given the Rapid API platform scale), and scope work with an engineer in the room. For AI-adjacent roles, expect questions about when to use retrieval versus fine-tuning, how to evaluate model outputs, or how to think about latency and trust in a booking context.

Cross-functional panel (2-3 hrs). Two to three interviews with design, data science, and engineering stakeholders. Each interviewer is evaluating a different dimension. The data person cares about metric definition, sample sizing, and whether you know the difference between a leading and lagging indicator. The design person cares about whether you have a real UX opinion or just defers to “test and learn.” The engineer cares about whether you understand build complexity and can scope sensibly. Preparing for the panel as one audience is a mistake.

Leadership interview (45-60 min). Behavioral and strategic. This round scores directly against Expedia’s five named behaviors (below). Some roles compress the loop to four rounds; the order above reflects the full senior-role version.

The five behaviors and what they score

Expedia publishes these on their careers site. They are the behavioral scoring rubric, not marketing copy.

Traveler First does not mean “ignore partners.” It means that when tradeoffs arise between traveler experience and internal convenience, traveler experience wins. In an interview, this surfaces as: can you describe a decision where you held firm on the user even when the business pressure pointed elsewhere? A story where the traveler first call was actually easy does not score well.

Think Big tests whether you engage with structural questions, not just feature optimizations. In 2026, this means interviewers want to see that you understand what it means when AI assistants can surface a bookable hotel result directly in a chat conversation, bypassing the Expedia funnel entirely. A candidate who thinks “Think Big” means “bold roadmap” will give a worse answer than one who engages with the structural disintermediation question.

Operate with Excellence is the experimentation and rigor behavior. Specific: sample size, test duration, holdout group, primary metric with a guardrail. Vague: “run an A/B test to validate.”

Ownership Mindset is about span. Did you write the brief, or did you also run the analysis, work with legal, and brief the partner team? The best stories here cross functional lines.

Succeed Together is the cross-functional and partner alignment behavior. Given that Expedia’s B2B partner network is central to inventory quality, a strong “Succeed Together” story often involves a situation where you had to align a hotel partner or distribution partner, not just an internal engineering team.

The two-sided marketplace test

Expedia is simultaneously a B2C marketplace and a B2B platform. When you propose a product change, interviewers listen for whether you’ve considered what happens on both sides.

A candidate who proposes improving Expedia’s hotel search by surfacing “vibe” signals (pet-friendly, chill atmosphere, good for solo travelers) has a smart traveler-side idea. A candidate who also notes that hotel ranking on Expedia is influenced by partner bid prices and that adding qualitative signals to ranking changes the value proposition for hotel partners, requiring either a new signal purchase model or an explicit tradeoff discussion with the partner team, has passed the two-sided test.

The same principle applies to any feature that touches search ranking, pricing display, review credibility, or cancellation policy. Every one of these is a negotiation between traveler interest and partner commercial interest.

The metasearch distinction

Expedia Group owns trivago, which operates on a cost-per-click model rather than an OTA conversion model. The PM surface is different: trivago’s ranking algorithm balances advertiser ROI (cost per click, conversion rate by advertiser) with traveler relevance, and the business model is ad revenue, not transaction margin. If you are interviewing for a trivago-adjacent role, or if you want to show breadth, understand that “improve hotel search” means something different in a CPC metasearch context than in a direct-booking OTA context. Candidates who conflate the two signal they’ve done surface-level research.

What questions actually get asked

• “How would you improve Expedia’s search experience?” (Most common product sense question. See what a strong answer looks like below.)
• “Our search-to-booking conversion dropped 8% last week. Walk me through how you investigate.”
• “Design a product for Expedia’s B2B partners that helps them improve their ranking without requiring an increase in bid price.”
• “How do you think AI will impact travel?” (This is not a thought experiment in 2026. It is a live strategy question.)
• “Tell me about a time you made a decision with incomplete data and were wrong.” (Ownership Mindset and Operate with Excellence.)
• “How would you measure the success of Expedia’s natural language trip planning feature?”

Strong vs. weak on the search improvement question

The AI question in 2026

Interviewers at Expedia ask “how do you think AI will impact travel?” and a generic answer about personalization or chatbots will not land. The honest answer acknowledges two things simultaneously.

First, Expedia is structurally threatened: when travelers can ask a Claude or ChatGPT instance for a hotel recommendation and receive a bookable result without visiting Expedia.com, the discovery layer that Expedia has owned for two decades moves to someone else’s interface. Expedia’s search traffic and affiliate referral model are directly exposed to this shift.

Second, Expedia is better positioned than most to survive it: they have booking infrastructure, supplier relationships, trust at the transaction layer, and now direct integrations in the AI channels. As of Explore 2026, AI search visibility (bookings initiated through ChatGPT and Claude integrations) is the fastest-growing marketing channel for the group. More than 30% of self-serve customer support is now AI-handled. Natural language trip planning is live on both Expedia and Vrbo: a traveler can describe a “pet-friendly lake house near Austin for a laid-back friends getaway” and receive bookable results.

The PM insight interviewers want to hear: feasibility is not the constraint. The constraint is getting a traveler to complete a $2,000 booking through an AI interface they haven’t learned to trust yet. That is a viable/lovable problem. Viable means the business captures enough margin when discovery is owned by an AI intermediary. Lovable means the AI experience earns trust fast enough to close, not just to recommend.

The 2025-26 product context worth knowing

Hotels.com and Vrbo returned to growth after completing a multi-year platform migration in 2025. If you want to show you’ve engaged with real PM challenges rather than generic marketplace thinking, reference what a painful feature-parity and brand-differentiation problem looks like post-migration. Expedia’s B2B AI Toolkit, announced in 2026, includes composable AI components for partners and an AI-powered merchandising API. The Responsible AI Council reviews high-risk AI deployments before they scale, which is relevant to any product design question involving AI features in a high-stakes (booking, financial) context.

Expedia also acquired Tiqets (activities and experiences) and CarTrawler (B2B car rental) in 2025-2026, expanding the product surface well beyond lodging and flights. A candidate who positions Expedia only as a hotel booking site is working from an outdated map.

Compensation (2026)

PM II in Austin: approximately $150K base, $12K sign-on, $12K relocation, $17.5K RSU. Senior PM total compensation varies significantly by location and team. Full context at PM salary by level.

What clears the bar

Name both sides of the marketplace before naming a solution. Know the difference between the OTA conversion funnel and trivago’s CPC metasearch model. Show experimentation fluency with a real test you owned: hypothesis, metric, sample, result. Engage with the AI disruption question as a current strategy problem with Expedia on both sides of the threat. And map your behavioral stories to the five published behaviors before you walk in, because those are the scoring rubric, not background context.

For the 2026 AI PM lens that sharpens every product answer here, see feasibility is free, lovable, not just usable, and proving viability.