Helping travelers spot hotel deals on Google Search
I designed how hotel deals appeared across Google Search, helping travelers notice unusually good prices while keeping the experience consistent with the rest of Search.
In six weeks, I grouped 20+ formats into five patterns and defined how deal indicators adapted to each.
Some examples are simplified reconstructions to explain the design decisions.
- Company
- Role
- Senior User Experience Designer
- Systems thinking · Visual language · Experimentation
- Team
- Product, Engineering, User Research, UX Writing, and Ads and Search component teams
- Project length
- 6 weeks
- Projected incremental annual Travel Ads revenue
- $16.9M
- Travel Ads revenue lift in the early experiment
- +0.60%
- Peak Travel Ads click lift versus control
- +0.24%
Step 1 of 4: Search for hotels
TL;DR
A hotel price can represent unusually good value without being a discount. I made that distinction visible across Google’s hotel Travel Ads, using signals that fit the surrounding Search experience.
- In six weeks, grouped 20+ formats into five patterns, and defined how a deal badge, green pricing, and supporting copy adapt to each while keeping the same meaning.
- Research confirmed travelers understood the price comparison.
- The treatment with the deal badge and green price did well in the experiment, with a +0.60% revenue lift, and launched to production.
Key decisions
- Follow Ads conventions or use the page’s existing deal language?
- I recommended adopting the green treatment already used in organic results.
- How much explanation did a green price need?
- We tested green pricing with and without the supporting copy, alongside the deal badge.
- How often could the badge appear before it lost meaning?
- With platform partners, we set a cap of ten deal treatments in a slate of roughly 25 to 30 results.
- Projected incremental annual Travel Ads revenue
- $16.9M
- Travel Ads revenue lift in the early experiment
- +0.60%
- Peak Travel Ads click lift versus control
- +0.24%
The problem
A deal needed to mean something specific
Finding a good deal is one of the biggest factors in how people book hotels, but Travel Ads had no consistent way to show that a price was good value. The goal was to help travelers spot those prices and book with more confidence.
For this project, “deal” meant Google had determined that a hotel’s price was unusually good for that property and those dates.
That ruled out familiar retail treatments such as a crossed-out previous price. Those patterns could imply a markdown that had never happened.
The design needed to help travelers make a good price noticeable while staying honest about why it counted as a deal. It also had to sit alongside organic hotel results, which already used green pricing and deal labels.
The challenge
20+ formats across Search
The Travel Ads system had 20+ formats. Some had image slots that could hold a badge. Others had compact price areas or existing containers that limited what we could add. Supporting text also needed to work across different widths and languages.
Designing each format on its own wasn’t realistic. I mapped every format and grouped them by where a deal signal could fit. They fell into five patterns, and every format in a pattern followed the same rules.
Tall card
Pattern: Image slotDeal badge in image slot
Hotel nav SRP unit
Pattern: Text heavyDeal badge + green price + supporting copy
Multi-entity
Pattern: Price tileDeal badge on price tile
Deal treatments
Three treatments, from most explicit to most subtle
I defined a set of treatments we could apply according to those constraints: an explicit “deal” badge where the component supported it, green pricing where space was more limited, and supporting copy where there was room to explain the price comparison. The UX writer and I worked out some of the combinations together.
Sponsored results
⋮Hampton Inn
https://www.hamptoninn.com › boston ⋮
Hampton Inn Boston Seaport District
4.2(1.2K)
Deal$198
Comfortable 1920s hotel with dining & a bar
Most explicit signal. Used where the format supports a badge.
What I wrestled with
Tension
An early neutral treatment fit existing Ads conventions. But alongside organic hotel results, it introduced a second visual language for the same idea of value.
Decision
I recommended adopting the green treatment already used in organic results. That gave travelers a consistent reference point as they moved through the page.
Trade-off
I prioritized consistency across the whole page over the conventions of the individual Ads component. The team adopted the green language, and it shipped across the page.
Neutral treatment
- Aligns more closely with Ads conventions
- Creates a separate deal language beside Organic
Green deal signal
Recommended- Connects to existing hotel deal language
- Makes the signal more recognizable across the page
A neutral badge that follows Ads conventions, versus a green signal that matches organic results’ deal language Tension
I was concerned that color alone wouldn’t explain why a price represented good value. And “cheaper than usual” described the price, not the hotel. Placed with amenities, it could read as one more hotel feature.
Decision
Working with the UX writer, I explored supporting copy such as “cheaper than usual,” kept close to the price where the format allowed. Some formats couldn’t, and wrapping and localization added further constraints.
Result
So we tested green pricing with and without the supporting copy, alongside the deal badge.
Three versions of the same sponsored hotel card. With the green price alone, the color is the only signal. With “Cheaper than usual” right under the price, the copy sits by what it describes. In a narrower card, the same copy wraps onto a second line.
Green price alone, the copy by the price, and the same copy wrapping in a narrower format Tension
A proposal to show the deal badge on every treatment in a carousel raised a concern: if everything looked like a deal, the badge would stop standing out.
Decision
I worked with the systems team to find internal precedent and brought that evidence into the discussion. With platform partners, we set a cap of ten deal treatments in a slate of roughly 25 to 30 results.
Trade-off
The cap drew on what related Google products had learned, plus our judgment about the page. It gave the team a consistent rule as the treatment scaled.
The same four sponsored hotels twice. In the first, every card has a deal badge and a green price. In the second, only two do. Under it, a slate of 28 results in which ten carry a deal: the cap of ten deal treatments in a slate of roughly 25 to 30.
Every card badged, versus a selective carousel within the cap
Research
Investigate what travelers understood
Alongside the live experiment, our UX researcher ran moderated sessions to see how travelers interpreted the deal signals. I built the prototypes and helped write the questions.
The research validated the design: travelers understood the price comparison. Their feedback also showed where the details could be clearer, and I refined those.
| What we checked | What travelers told us | What was refined or validated |
|---|---|---|
| Whether travelers understood the price comparison | Participants understood the deal meant a lower price than usual for that hotel. | No change needed. |
| What “deal” suggested on its own | Some expected a sale, a previous price, or an expiration date. | Describe the comparison precisely and avoid markdown imagery. |
| Whether green alone carried meaning | Participants noticed the green price, and a badge or short explanation made the meaning clear. | Pair color with a badge or explanation where space allows. |
| What “Cheaper than usual” compared against | Some asked: other hotels, other dates, or this hotel’s usual price? | Make the basis of comparison clear. |
| Where the explanation sits | Grouped with amenities, it could read as a hotel feature. | Keep the explanation next to the price where the format allows. |
| Whether the deal crowded out other details | Participants still weighed location, reviews, cancellation terms and total cost. | Keep those details visible, without the badge dominating the card. |
| How sponsored deals read | Some asked whether advertisers paid for the label. | Keep the deal’s meaning distinct from sponsorship. |
Results
Early results supported further investment
After roughly two weeks at limited exposure, the treatment with the deal badge and green price performed well and became the version that launched to production. The research also showed that travelers trusted a comparison with the hotel’s past prices.
Outcome
The deal badge and green price, after roughly two weeks at limited exposure
- Projected incremental annual Travel Ads revenue, using the optimized trigger threshold
- $16.9M
- Travel Ads revenue lift
- +0.60%
- Peak Travel Ads click lift versus control
- +0.24%
Guardrails held: no regression to whole-page performance, a healthy deal trigger rate, balanced badge distribution across formats and regions, and legal compliance with pricing accuracy above 90%.
Search results for “hotels in Boston”: deal badges on the sponsored carousel, beside organic results’ own green deal language
01 / 06Reflection
Define the rules that let a design scale
Grouping 20+ formats into five patterns gave the team one set of rules instead of a decision per format, and let the work land in six weeks.
The decisions that mattered most were made with teams outside my own: with the Ads component owners on the green language, and with the platform teams on the badge cap. In both, evidence moved the discussion further than preference.
I now bring Engineering and platform partners into that mapping at the start, so we agree early on what stays consistent, where variation is allowed, and which assumptions need testing.
Next case study
Uber: Driver Loyalty
A two-week trial that let Uber drivers use Uber Pro benefits before earning them.