top of page

real-time shopping personalisation optimizing ROI and boosting NPS with Qoqa

About QoQa

QoQa is a Swiss e-commerce platform that stands out by offering unique and time-limited deals on a wide range of products and services. Founded in 2005, QoQa has grown into a beloved marketplace known for its dynamic and engaging approach to online shopping. The platform provides its community with daily deals that span various categories, including electronics, travel, home goods, and lifestyle products, all curated to deliver exceptional value.

What sets QoQa apart is its commitment to creating a vibrant and interactive shopping experience. The platform not only focuses on quality and affordability but also fosters a sense of community among its users through engaging content, social media interaction, and exclusive events. With a strong emphasis on customer satisfaction and innovation, QoQa continues to redefine the e-commerce landscape in Switzerland, making shopping a fun and rewarding experience.

User Story

As a business, QoQa proposes exclusive offers to their users on a daily basis. Over time the amount of offers grew from one to around 10-15 per day. While this led to business growth it also impacted users’ satisfaction (expressed through NPS scores) as they felt overwhelmed by the amount of offers proposed. This led to a paradox. In order to overcome this challenge, the QoQa team identified the following user case:

“As QoQa I want to deploy a mechanism to determine which offers are shown to users, based on their past behavior in order to personalize their experience and improve their satisfaction score.”

As QoQa’s product offering is divided into specific shops (Kids, Alcohol, Sports, etc.) we decided to turn their general use case into the following specific, actionable and measurable user story: 

“As QoQa I want to de-prioritise or hide alcohol related product offerings to QoQasians (clients) that have been identified as non-alcohol drinkers in order to improve QoQasians satisfaction without impacting business growth on our alcohol vertical”



Step I: Define and create the “non alcohol drinkers” audience

What might sound as the easiest step was actually not since what we are looking for is the absence of an action rather than the action of buying alcohol itself. What if we make a mistake in the way we define the audience and therefore expose the client(s) to a wrong experience. This is an example of a question QoQa had to answer. After careful reflection, the team decided to build an audience using the following criteria: 

  • 10+ orders (on your lifetime)

  • 1+ order in the last 360 days

  • Opened the website or app in the last 90 days

  • Not subscribed to notifications or newlester on the alcohol vertical

  • No purchase on the alcohol vertical in the last 720 days

  • Not navigated on the alcohol pages on the site or app in the last 30 days

Step II: Design and validate the results of the personalisation

Several scenarios were possible to fulfill the identified user stories: 

  1. Pushing down - meaning positioning the alcohol offering (for the people identified as non-alcohol drinkers) at the very bottom of the page(s) (website & app)

  2. Collapsing - meaning collapsing the alcohol offering at the bottom of the page(s) with an explicit message asking they were interested in seeing those offers 

  3. Hiding - meaning fully removing the alcohol offering from the page(s) (website & app)

In the end, the team decided that fully removing (hiding) it was the best option to overcome the feedback shared by the clients. 

Step III: Design & implementing the technical solution

After considering several options, we decided to implement the following components:

  • Twilio Segment’s Audiences

  • Twilio Segment’s Profile Sync

  • QoQa’s internal app personalization logic

From there, the implementation was straightforward. Once the audience was built in Segment, it was automatically picked up by subsequent Profile Sync runs, ensuring the information was available in the data warehouse. With this information, QoQa was able to identify users in the audience within their database, which also contains the product catalog. 

As it was a first test of personalisation of the app and they wanted to be able to measure the impact, QoQa split the users from the audience into two groups: exposure and control. For the exposed group, they implemented a mechanism to not show any alcohol products in the app. 

Products that users see in the app are based on offer lists they’re eligible to. QoQa hid the offers tagged as alcohol for the exposed group, resulting in these users not being shown these offers anymore when navigating the app. On the other hand, the control group, also part of the ‘non-alcohol drinkers’ audience, was shown the unaltered lists of offers - including alcohol. 


In order to statistically assess the results of the user story, the QoQa team deployed an A/B testing which exposed 50% of the eligible audience to the personalisation (= the hiding) and 50% of the eligible audience as the control group not exposed to the personalisation. The results were very positive as the following was measured: 

  • +5% growth on the CSAT of the group that has been exposed to the personalisation

  • +5% growth on the the feeling of relevance of the offers 

  • +15% on the frequency of opening the app on a daily basis for the group that had been exposed to the personalisation


As a next step, the QoQa team started re-using the same logic and applying it to other vertical. Example: removing/hiding kids related product to people flagged as not interested.

Interested? Feel free to reach out!

Recent Posts

See All


bottom of page