Selected works / 02 of 04 Fast Track Ancillary

The outcome

+€180k/mo

in recovered and incremental flight revenue, after a -0.9% conversion failure was diagnosed and rebuilt into a +2.05% win.

Fast Track Ancillary

A first-of-its-kind security-skip product, sold inside a flight booking funnel that was already conversion-optimized. And how a failed launch turned into the diagnostic process I'm proudest of.

1 · hero-context Cinematic shot of V1 Fast Track on the Booking.com Extras page · 1200×750 · 16:10

01 / The setup An OTA-first ancillary, built against the clock.

Booking.com Flights generates revenue through ancillaries: seats, baggage, insurance. In 2024, leadership set a clear priority: expand the ancillary portfolio to drive incremental revenue. Fast Track was the target. Airport security queues are one of the most common pain points in air travel, and no online travel agency was selling Fast Track inside a flight booking. Booking.com would be the first.

Being first meant no playbook. No competitor data to benchmark against. No proven UX patterns. And leadership committed to a 5-month timeline to capture summer 2024 booking season.

The constraints I was designing into:

Constraint Why it mattered
Zero OTA precedent Every design decision was a bet without reference data.
5-month timeline Eliminated traditional staged rollout. One shot to get V1 right, or learn fast.
Limited airport availability Fast Track was only at certain airports. The offer couldn't be shown to every customer.
Unproven supplier integration We relied on a third-party for QR code delivery. New relationship, untested at scale.
Optimized funnel at risk The booking funnel was already conversion-optimized. Adding anything new was inherently risky.

02 / Discovery Customers told me what Fast Track meant. Then I asked the wrong follow-up.

Before designing anything, I needed to answer two questions. What does Fast Track actually mean to customers? And where in their journey do they expect to buy it?

I ran a quantitative study to understand customer perception and purchase intent. Customers see Fast Track as more than a utility. They described it as a way to treat themselves at the start of a trip; a solution for arriving and finding the queue longer than expected; peace of mind when running late.

2 · customer-quotes 3–4 anonymized customer-research quotes pulled from quant study · 1200×750 · 16:10
What Fast Track means to customers, in their own words.
3 · purchase-moment-chart Bar chart: % of customers preferring each purchase touchpoint · 1200×750 · 16:10
Where customers expect to buy Fast Track.

How others were selling Fast Track

I led a market positioning workshop with my Product Marketing Manager to map how Fast Track was being offered across travel: OTAs, airlines, airports, adjacent services.

Provider type Example How they sell it
Bank card perks MasterCard Bundled as a card benefit; no active purchase decision.
Transport bundles Heathrow Express, Glasgow × Fast Park Paired with parking or transport for a combined deal.
Airline branded fares Norwegian Airlines Included in premium fare tiers, not sold separately.
Standalone products Fasttrack.com Dedicated website where customers seek it out.
Direct airport sales Bristol, Birmingham airports Sold on airport websites, often last-minute.

03 / V1, what I built An Extras-page bet, with a redemption flow built for things going wrong.

I led a customer journey workshop with 26 participants from product, engineering, marketing, and research. We mapped the full journey from booking through arriving at the airport, identifying opportunities at each step. The workshop used COMBAT statements and How Might We questions, then I used RICE scoring to turn the messy brainstorm into a ranked roadmap.

The placement decision

The biggest question: where in the booking funnel? Research showed customers wanted to buy during booking. That pointed to the Extras page, where customers were already making ancillary decisions.

Option Reasoning Decision
Extras page Research-validated. Groups Fast Track with similar 'peace of mind' ancillaries. Selected.
Dedicated standalone page More space to communicate value. Rejected. Adds a step, higher abandonment risk.
Post-booking upsell only No risk to flight conversion. Rejected. Customers preferred buying during booking.

The bet I was making. Customers could evaluate Fast Track's value while also deciding about baggage, insurance, and seats. This turned out to be wrong, but I didn't know that yet.

4 · v1-extras-page V1 Extras page in browser frame, showing radio-button placement · 1200×750 · 16:10
V1: Fast Track as a radio-button option on the Extras page, alongside baggage, insurance, and flexibility.

Designing for things going wrong

The supplier relationship was new. I couldn't assume things would always work. So I designed for five status states, not just the happy path. For each failed or cancelled state, I wrote dynamic UX copy that was honest about what happened and told the customer exactly what to do next.

5 · status-state-matrix 4-up grid of phone/desktop mockups: Confirmed / Pending / Failed / Cancelled · 1200×750 · 16:10
The four real status states, each with its own UX response.

I also added a 'How Fast Track Works' modal to educate unfamiliar users. Many customers had never heard of Fast Track, so asking them to pay for something they didn't understand was a barrier.

04 / V1, what happened A 1.3% attach rate. A -0.9% drop in conversion. A €270,000 hole.

The experiment ran as a 50/50 A/B split on desktop and mobile web, targeting a 5-week runtime during summer booking season.

6 · v1-failure-dashboard Dramatic results viz: line chart of -0.9% conversion drop OR scorecard with the three numbers · 1200×750 · 16:10
V1 results across the 5-week experiment.

Fast Track hurt conversion more than it helped revenue. The experiment was stopped after 5 weeks.

-0.9% Flight conversion
1.3% Attach rate
-€270k Net cost / 5 weeks

Fast Track generated roughly €500/day from 22 purchases, but cost roughly €2,300/day in lost flight bookings. Net: approximately -€54,000 per month, totalling -€270,000 over the 5-week run.

05 / Diagnosing Was this UX, or was it product-market fit?

The question was not whether V1 had failed. It was why. If the failure was a UX problem, V2 could fix it with design changes at relatively low cost. If it was a fundamental product-market fit problem (meaning Fast Track simply did not belong inside a flight booking funnel), then we should kill it.

I needed to find out which one before recommending next steps to leadership.

I led a cross-functional brainstorm with PM, Engineering, Data, and Marketing to generate failure hypotheses. The goal was to separate UX issues from PMF issues before any iterate-or-kill recommendation.

7 · hypothesis-grid 2×2 plotting 5 failure hypotheses by confidence × type · 1200×750 · 16:10 · the staff-signal artifact
Five failure hypotheses, plotted by confidence and type. The top of the chart is UX, not PMF.
Confidence Hypothesis Type
High Page length pushed Next button below fold. Users abandoned. UX
High Decision fatigue from too many ancillaries on one screen. UX
Medium €23 perceived as too high without context. Value prop
Medium Fast Track was unfamiliar; perceived value was low. Education
Low Wrong customer segment was seeing the offer. Targeting

Validating with usability testing

I ran an unmoderated usability test with 16 participants. 8 EU and 8 US, ages 20 to 65, all on desktop web to match V1 conditions. Two flows: purchase (book a flight, decide on Fast Track) and redemption (find your QR code).

8 · usability-findings 3–4 think-aloud quotes from participants OR heatmap of friction points · 1200×750 · 16:10
What the testing revealed about V1.
Finding How many What it meant
Could successfully add Fast Track and complete the flow. 14 / 16 The core UX worked. Not a usability breakdown.
Wanted to decide about Fast Track later, not alongside baggage and insurance. 5 / 16 Decision fatigue confirmed.
Were not sure if €23 was worth it. 6 / 16 Value proposition wasn't landing. Price without context felt expensive.
Older users were confused by 'security lanes'. 3 / 16 UX copy needed clarity for less frequent travelers.

06 / Building the case A two-month sprint, with a break-even under three weeks.

I presented leadership with a clear recommendation: iterate, do not kill. The case had three parts.

  • Diagnosis. The problem was UX execution, not product-market fit. Customers wanted Fast Track. We just made it too hard to buy.
  • Prioritized roadmap. Fixes ranked by conversion-recovery potential. Primary changes targeted the biggest conversion drivers; secondary changes strengthened the value proposition.
  • Business case. A 2-month sprint at roughly €40,000. Recovering the 0.9% loss meant 23 additional bookings per day. About €69,000/month. Break-even in under 3 weeks.

07 / V2, the recovery Three changes. +2.05% conversion. +€180,000 per month.

V1

9a · v1-extras-page-detail
V1 Extras page: radio-button pattern, full ancillary stack, Next button pushed below fold

Forced choice. Long page. Ambiguous price.

V2

9b · v2-extras-page-detail
V2 Extras page: checkbox pattern, baggage moved off, value/redemption messaging separated

Optional. Shorter. Value before redemption details.

Metric V1 result V2 result Change
Flight conversion -0.9% +2.05% +2.95 pp
Fast Track attach rate 1.3% 2.5% +1.2 pp
Funnel drop-off after Fast Track screen +1% -0.3% -1.3 pp
10 · v2-recovery-chart Recovery line: V1 endpoint at -0.9% rising through V2 launch to +2.05% · 1200×750 · 16:10
The recovery curve from V1 endpoint to V2 outcome.
+2.05% Flight conversion
2.5% Attach rate
+€180k Combined monthly

+2.05% meant roughly 50 additional bookings per day, approximately €150,000 per month in recovered and incremental flight revenue. The 2.5% attach rate added ~43 Fast Track purchases per day. Approximately €30,000 per month in ancillary revenue.

08 / What I'd carry forward Diagnosing failure systematically matters more than avoiding it.

Speed without validation is costly. V1 taught me how to bring the right people into the room, generate and prioritize hypotheses, distinguish a UX problem from a product-market fit problem, and build a business case for iteration instead of giving up.

The V2 outcome speaks for itself. But I'm more proud of the diagnostic process that made V2 possible than the final numbers.

If V1 had succeeded, I would have shipped a product. Because V1 failed, I learned how to save one.