How I Turned Agoda’s -5 NPS Into a Conversion Machine
A case study on fixing user frustration with one simple, data-driven design decision
A deep dive into transforming negative customer sentiment into measurable business growth through user-centered product design
It was a sleepy Wednesday afternoon, and Priya’s mind was already on the upcoming long weekend. With the kids out of school, a last-minute staycation in Bali seemed like the perfect escape. She opened up Agoda.com, her go-to for quick bookings, and typed in “Bali” and the number of guests: two adults, two kids.
The screen filled with hundreds of options, a dizzying array of sprawling resorts, romantic villas, and business hotels. She found herself clicking the filter button, her fingers automatically searching for “swimming pool,” “family-friendly,” and “kids’ club.” This felt less like planning and more like work.
“Again?” she thought. “Why doesn’t this site just know I’m with my family?”
Priya’s frustration wasn’t just another user complaint. It was a symptom of a much bigger problem that was quietly bleeding millions in revenue and destroying customer trust.
The Wake-Up Call: When Numbers Tell a Brutal Story
When I did some research, the data painted a stark picture that no one could ignore:
Net Promoter Score: -5
For every customer willing to recommend Agoda, there was at least one actively discouraging others from using the platform. In the hyper-competitive travel industry, this wasn’t just bad, it was existential.
Source: https://www.comparably.com/brands/agoda-company
Customer Satisfaction: Below Industry Average
Post-booking surveys consistently returned scores that made the company uncomfortable. The feedback was brutal but consistent: “frustrating,” “difficult,” “time-consuming.”
But here’s what really kept me up at night: users weren’t just failing to convert, they were leaving with negative impressions that could permanently damage their relationship with the Agoda brand.
The Competitive Reality Check
When I benchmarked Agoda’s NPS against competitors, the gap was devastating:
- Booking.com: +20
- Expedia: +15
- Industry Average: +12
- Agoda: -5
They weren’t just losing; they were in a completely different league. And in the wrong direction.
Going Deep: The Detective Work Behind the Numbers
Numbers tell you what is happening. But as a product designer, I needed to understand why. So I rolled up my sleeves and went hunting for the real story.
The 500-Review Deep Dive
I spent weeks analyzing over 500 app store reviews. The pattern was unmistakable: 40% of negative feedback mentioned the repetitive, laborious nature of filtering.
“Every time I search, I have to set the same filters over and over.”
“Doesn’t remember I always travel with kids.”
“Why can’t it just show me family hotels first?”
The Heuristic Evaluation Reality Check
Using Nielsen Norman’s 10 Usability Heuristics, I audited the existing search flow. The results were damning:
- “Flexibility and efficiency of use”: Failed. No shortcuts for power users.
- “User control and freedom”: Failed. Users felt trapped in endless filter loops.
Competitive Intelligence: What Others Got Right
I dove deep into 5 major competitors. While none had perfect “smart” systems, the leaders were already surfacing relevant amenities on main search results. They understood something we didn’t: context matters more than choice overload.
The insight hit me like a lightning bolt: Agoda’s search treated every user like a blank slate, every single time. This willful ignorance of user context was the primary source of friction.
The Users: Putting Faces to the Frustration
Through user interviews and journey mapping, two personas emerged as the most underserved by their current system:
The Strategic Approach: Impact vs. Effort Matrix
Before diving into solutions, I mapped potential fixes on an Impact vs. Effort matrix. This isn’t just good practice, it’s survival in a resource-constrained environment.
High Impact, Low Effort (THE SWEET SPOT):
- Smart filter presets based on guest composition
High Impact, High Effort:
- Full AI-powered recommendation engine
- Complete search redesign
The choice was clear. We needed to start with the highest-impact, lowest-effort solution that could deliver results fast.
The Solution: Elegantly Simple, Powerfully Effective
The final design had two critical components:
The Logic: Transparent and Extensible
IF kids > 0
THEN apply "Family Preset"
(auto-select: Swimming Pool, Family Friendly, Kids' Activities)IF adults == 1 AND kids == 0
THEN apply "Solo Traveler Preset"
(auto-select: Free Wi-Fi, Fitness Center, Business Center)Simple rules. Transparent logic. Immediate value.
The Interface: Helpful, Not Intrusive
The magic was in the execution. A single-line, dismissible banner appeared below the search filters:
“ We’ve applied filters for ‘Swimming Pool’ and ‘Kids’ Activities’ because you’re traveling with family.”
Two clear controls:
- [Change Filters] → Opens full filter menu with pre-selections visible
- [Reset] → Immediately removes all smart filters
The microcopy was crucial. Not just what I did, but why I did it. Transparency builds trust.
Validation: The Hypothesis-Driven Approach
Every great product decision starts with a testable hypothesis:
Get Mayank Sabharwal’s stories in your inbox
Join Medium for free to get updates from this writer.
“By intelligently pre-selecting relevant filters based on user context, I will reduce cognitive load, decrease time-to-relevance, and thereby increase user satisfaction and conversion.”
Primary KPIs
- CSAT (Search Experience): +10% lift target
- Search-to-Booking Conversion: +1.0pp lift target
Secondary KPIs
- Search Abandonment Rate: -10% reduction target
- Time-to-Magic: -30 seconds target (time from results load to first hotel click)
Guardrail Metrics
- Page Load Time: No increase in latency
- Filter Usage Rate: Monitor behavioral changes
The Results: When Design Meets Business Impact
The A/B test results exceeded our most optimistic projections:
Customer Satisfaction Transformation
- Search Experience CSAT: +15% lift (vs. +10% target)
- Overall NPS improvement: -5 to +8 in the test cohort
Business Impact
- Search-to-Booking Conversion: +1.35pp lift
- Revenue per Visitor: +12% increase
- Search Abandonment: -18% reduction
Projected Annual Revenue Impact
With 20M monthly searches and average booking value of $120: $26M additional annual revenue from this single feature.
The ROI was undeniable. Users were happier. The business was growing.
The Rollout: Methodical and Risk-Aware
Success in testing doesn’t guarantee success at scale. We rolled out methodically:
- Internal Testing: Rigorous QA with internal teams
- 1% Release: Monitor guardrail metrics closely
- 10% → 50%: Gradual expansion with continuous analysis
- 100% Release: Full rollout after validation
The Future: Building on Success
This wasn’t just a feature, it was a foundation. The roadmap ahead included:
Phase 2: Learning Preferences
When users deselect smart filters, we learn and refine their profile.
Phase 3: Expanding Personas
Business travelers, couples, groups, each with tailored presets.
Phase 4: Natural Language Search
“A quiet hotel near the beach with a great pool for kids” → perfectly filtered results.
The Deeper Lesson: Context Is King
The most profound insight from this project wasn’t technical, it was philosophical.
In our rush to provide infinite choice, we had forgotten that users don’t want more options. They want the right options, faster.
Priya didn’t want to see every hotel in Bali. She wanted to see family-friendly hotels in Bali. David didn’t want budget hostels in his business search results. He wanted efficiency and relevance.
The system that understands context isn’t just more useful, it’s more trustworthy. And trust, as we learned from our NPS journey, is the foundation of sustainable business growth.
Key Takeaways for Product Teams
- Negative metrics are gifts. They tell you exactly where to focus your efforts.
- Simple solutions often have the biggest impact. Don’t over-engineer when elegance will do.
- Context beats choice. Users prefer relevant options over comprehensive options.
- Transparency builds trust. Show your work. Explain your logic.
- Measure everything. Hypothesis-driven design isn’t just best practice, it’s survival.
The Bottom Line
Sometimes the most significant user experience improvements don’t come from massive redesigns or flashy new features. They come from identifying and removing a single, critical point of friction.
In Priya’s case, that friction is the exhausting ritual of re-filtering every single search. By acknowledging her context (traveling with family) Agoda could transform her experience from frustrating to delightful.
And the business impact? $24M in additional annual revenue potential from understanding that families need pools and business travelers need WiFi.
Not bad for a Wednesday afternoon insight that could change everything.
This is my unsolicited design proposal for Agoda. Sometimes the best ideas come from outside observers who can see the forest for the trees. What do you think, should they implement this?
About the Author
I’m a product designer who believes great user experiences drive great business outcomes. I love analyzing products and imagining how small changes could create massive impact. Connect with me to discuss more ideas like this.
