Case study for trivago Discovery

Inès Mir
6 min readMay 17, 2021

Objective: Trivago aim to create a product that captures more users in the earlier part of the travel funnel; where they get inspiration for their next trip. Focus made on regional/nearby trips for the user.

Product: https://discover.trivago.com/

Value Hypothesis

Task 1.1.

Do a walkthrough of the discover product. Identify an opportunity for improvement. What makes you think this is worth to be addressed?

Verification of the Value Hypothesis

To identify an opportunity for improvement we need to validate the value hypothesis first. The question is if trivago Discovery is helping people to find inspiration for the trip well enough. To answer that, I’ve conducted a quick usability testing.

Instruments used

  1. Guerilla testing
    I asked family members of different age how did they choose a destination for the last trip, what helped them. Briefly walked through the Discovery feature, but, unfortunately, it has no practical use on Russian market.
  2. Unmoderated remote usability testing
    I then questioned two friends living in Germany (where I believe trivago works the best) about their last travel and asked them to complete tasks in trivago Discovery and send me an audio file with their perspectives.

Results

  1. All respondents took local trips in the last year.
  2. Most of the participants chose the destination based on a friend suggestion (conversation or instagram).
  3. Few picked up a destination based on a unique property available at Airbnb (tree houses, farms, etc.)
  4. The majority of respondents didn’t find the listed destinations at trivago Discovery inspiring, as there is no information on why visit.
  5. Some of them marked that the feature is lacking customisation and personalisation, such as traveling with dogs, choosing property type (house, apartment, etc), map navigation, public commutes, weather at the destination, etc.
  6. All wished there were some curated property selections, as some of the respondents admit that they sometimes look for an escape at a cosy place and don’t care much about surroundings.
  7. Few other small discoveries had been made such as an inability to see a public commute, shortage of a maximum radius as in Russia 300 km is too small, etc. This insights will partly be sketched in design and partly omitted, to focus on a core findings.

To validate this insights I would interview 5–6 people of different age, gender and social groups in a controlled environment. It’s important to question people from a similar country or region to eliminate situation like one I had in Moscow, where the product features 10% of destination available.

Summary

The majority of respondents didn’t find trivago Discovery useful for the earlier part of the travel funnel. As it’s not providing any information on why to choose a certain destination. It’s also lacking customisation and personalisation.

If this insights are appear to be true, then the value hypothesis is false and people either don’t use the product or use it in an unplanned way. That is a core issue of the product and should be addressed at the first place, as it evokes the question if we should spend a resources on the product.

Task 1.2.

Please map 1–2 solution ideas for the opportunity. Together with a visual representation of the idea.

Solution 1

I believe, that the product should focus on turning the Discovery page from listing of local destinations to a curated collections of best deals.

Possible collections

  • Near you
  • Weekend plans
  • Best stays
  • Seasonal offerings
  • Properties grouped by activities and the type of users (for a big company, family, couples, etc)
  • etc.

The destinations could be customised according to the collection and be more visual.

Presenting destinations from different angles

In the series of experiments the best combination and strategic could be found by comparing CTR of different collections for different users. In time the position and content of the collections could be defined with ML technologies, which would help trivago be more personalised.

Solution 2

Make destination’s landing more informative and attractive would help a user to decide. Adding the COVID-19 latest information would create trust to the product, leaving user more secure of the choice.

Other small improvements

Incremental improvements in style (rounding, shadows) and added features like weather conditions and a map aim to make the product more friendly and responsive to human needs.

Smooth and modern UI would make an impression that the product works smoothly, predictably and is maintained by the team everyday.

Task 1.3.

If you had more time and resources, how would you validate if your solution is worth building?

Lab usability testing

In order to prove my solution is worth building I would repeat the usability testing from the research stage on the prototypes, to see if we have any change in a people opinion. 5–6 different people of different ages, social status, nationality and gender, from the same country or region we chose before.

If the respondents opinion improved

MVP

We could launched an MVP on one region featuring all new ideas. We could cut costs by building it on default components, with graphic images selected manually and for a small amount of destinations.

As we validate the core value hypothesis we should immediately see how our improvements influence the funnel and retention (I think, in trivago’s case we should consider rolling retention).

if not

In-Depth Interview

In this case we should search for more insights from users in a form of intensive individual interviews with a small number of respondents (up to 5) to explore their perspectives on a travel industry, their travel experience and find new ideas for the product.

Then repeat the ideation step.

Task 2.

How do you manage speed vs. quality?

I see the design process as a progressive jpeg loading on a screen.

As soon as I have an idea or direction I can deliver a sketch or prototype of a quality that directly proportional of the time I have.

Of course, at some point the grows of quality and definition starts to slow down to a point where adding more time won’t affect it.

I figure the optimal time/quality point on a graph by defining and managing the expectations of the team and stakeholders. I pay special attention to extreme points.

Possible extreme points

  1. Shortage of time, huge scope
    • Asking what is the vital in the scope and what can be postponed
    • Discussing quick prototyping (paper sketches, low-definition wireframes, etc)
    • Considering the usage of commercially available off-the-shelf products
    • Delegating to other designers
    • Building an MVP on existing components
    • Presenting a correct estimation of the scope and negotiate on a deadline
    • etc
  2. No deadline, undefined scope
    • Finding a responsible person who can define the priority
    • Defining the scope by researching
    • Breaking the task down into pieces each one with a certain scope and deadline
    • Setting the deadline myself
    • Understanding if the task is worth making
    • etc

By questioning and negotiating I’m able to deliver the expected quality in a given time.

Thank you for your time,
Inessa.

--

--

Inès Mir

Principal product designer at Zalando and Instagram influencer @ines.ux