Context

⭐ This product is across all brands (foodpanda in APAC, Yemeksepeti in TR, and foodora in EU). However for this case study, we will focus on the experience in foodpanda’s context.

foodpanda is a Q-commerce app focused on delivering food, groceries and more. It is present in several countries in APAC including Singapore, Malaysia, Hong Kong, Thailand, etc.

In these markets, it runs a free-to-participate loyalty programme called panda rewards where users participate with stamp cards (Read case study here) — a mechanism that encourages users to place certain orders to collect stamps to redeem a final reward which could be points, vouchers, badge of completion and/or a grand prize. Certain stamps also give users mini rewards along the way as they work towards completing the entire stamp card.

---

Problem

To understand the problem, let’s walkthrough the existing user flow to better understand where the challenges and opportunities lie.

Existing user journey: User places an order according to the stamp card tasks. After payment is made, there is a confirmation on the screen that shows what stamp card is related to the order items and the reward (stamp, points, vouchers, badges) the user will earn from this transaction. Additionally:

• This is also an opportunity for users who coincidentally place an order related to a stamp card task, to lock in the reward by ‘Starting’ a stamp card. (Note: Having a ‘Start’ button is a business requirement in order to track participation rate)
• Should the user cancel the order before it is delivered, the reward will be not be issued. This is to minimise any chances of fraud.
• Unfortunately if a user (both aware and not aware) did not place an order that’s valid against a stamp card task, no communication on loyalty will be shown on screen after payment is made.

The user journey above poses these 3 main problems:

1. Discoverability
   • foodpanda’s loyalty programme has suffered in visibility since it’s inception. This is due to the lack of presence in the core journey and the somewhat hidden entry points — users can only browse any information (ie. stamp card tasks) of the loyalty programme through the app’s hamburger menu and a FAB in the list of restaurants (in which through research shows users mistaken the FAB as access to customer service).
   • Without the presence or any information of the programme in the core journey, users are not familiar or reminded about stamp card tasks to complete. This affects the potential engagement of stamp cards, more critically vendor-sponsored ones.
2. Ease of use
   • Due to the entry point of the loyalty programme, to refer back to the stamp card tasks, users would have either:
     1. read the task and remember the order criteria whilst ordering, which results in high cognitive load for the user, OR
     2. constantly exit and enter restaurant and cart page to browse stamp card tasks in the panda rewards page to compare notes, which is a lot of hassle, clicks and deviating users from the main mCVR of Cart and Checkout.
3. Complexity:
   • With the nature of stamp card mechanism and varying order criteria that can happen for each stamp card and stamp, this adds to the complexity of the loyalty programme which can be overwhelming for users and a high barrier of entry. Users can be discouraged to participate.

---

Opportunity

How might we design an intuitive and seamless experience for users to effortlessly engage with and benefit from stamp cards?

---

Goals

Alleviate guesswork
Minimise or eliminate any friction for users to participate in stamp cards. Users should easily know what orders to place according to the stamp card task.

Improved visibility
Minimise any potential engagement of Stamp Cards falling through the cracks. Users should easily discover the programme and thereafter make their own decisions to join.

Minimise redirection
The main goal within the core user journey is to help users go towards placing order. Users should be able to make quick actions and decisions with minimal risk of dropping off and impacting order CVR.

---

Challenges

No control for what Stamp Card tasks will be set up
Tasks are set up region-based by the respective Marketing PIC in the country and product has control on how complex it can be.

Dependencies

In this project, it relies heavily on cross-collaboration with the Design Owner in Cart. As they were going through a redesign and other priorities, the alignment gets recurring uncertainty and varying directions which added unto the project timeline.

---

Process

Desk research

I started off with a quick peek into the competitor’s scene. There isn’t an apple-to-apple comparison available that I’ve found, therefore the aim of this research is to understand patterns used in relation to any rewards communication within their Cart / Checkout pages. This can help establish a familiar layout that can help users adopt this new feature within our Cart page in foodpanda.

Exploration & pair designing

I explored several options keeping in mind the real estate of Cart page and the information we could succinctly display to keep communications straightforward.

Depending on the criteria AND whether or not the user has already started a Stamp Card, the information displayed for the user will be tailored to their individual circumstance in order for them to gain the reward. Here are some scenarios:

• Scenario 1: A user who has not started a stamp card but have met ≥1 Stamp Card order criteria, will be encouraged to Start a stamp card.
• Scenario 2: A user who has started and have not met =1 Stamp Card order criteria, will be encouraged to match the criteria.
• Scenario 3: A user who started and have met ≥1 Stamp Card order criteria, will be encourage to place the order and move toward payments.

With the several uses cases involved as listed above, I worked closely with a UX writer to personalise and optimise each state and thereafter, create an easy to see matrix that helps any stakeholders understand the breadth of the variations that can occur in this mix. In total there can be 9 scenarios:

I worked closely with the Design Owner of Cart to really ensure we minimise any potential negative impact of Cart’s metrics. This include discussions on allowing any CTA that’s rewards related displayed at the forefront which can distract users and potentially confuse them with the main CTA.

Additionally together with Product, we also aligned that in order to reduce information overload in Cart, we would only surface the entry point if the vendor is related to a Stamp Card, and only the next best stamp to earn is displayed as opposed to all stamps within a stamp card. The assumption is, if users are not overwhelmed with choices, faster decision making can be made; Much like ordering at a restaurant with a ton of items on the menu compared to another restaurant with only a handful of options to choose from.

I also gathered other designers who are working on Cart and Checkout in Restaurant and Shops to get feedback on text styling used for these 9 variations.

User testing on Lyssna

I work closely with a dedicated UX Researcher to have this research design vetted through to ensure credibility and effectiveness of this unmoderated experiment. This UT will run as a design survey.

Study scope

Surfacing eligible stamp cards in Cart informs users of their eligible and/or potential rewards in their core ordering journey

Purpose

To gather user’s comprehension on their eligibility and understanding of Rewards they will earn with their order through Stamp cards.

Research questions

1. Can users understand the rewards they will earn with their order?
2. Can users understand what they need to do to earn rewards?
3. What do users expect to see when they click on the entry point?

Hypothesis

It will be easier for users to know if they will potentially earn rewards with their current order when communicating with different copies according to use cases

Participants profile

Factors	Details
👥 Total participants	20
📱 Device	App only
🌏 Markets	SG, PK, SE, TR
🥡 Attributes	Uses food delivery apps / services

Design

Testing steps

As this test aims to gain user’s comprehension of the feature, the test questions are designed with a lot of open-ended questions to capture their raw understanding when asked to do specific tasks. Here are some screenshots of the research design:

Limitations

1. Lack of context
   Users on the test platform may not be actual users of the app. Having context of Stamp Cards may result in different results due to the complexity of Stamp Card mechanism.
2. No avenue to further probe
   With all unmoderated tests, the limitation lies in having limited interaction with testees whom we are unable to reach out to further for further questioning.

User testing results ✨

✅ Most users understood their current order can earn them rewards

✅  Majority of users understood what is required to earn those rewards

🔍 Most users expected to see more information about the rewarded points

✅ 70% of users are able to recall and identify the entry point

---

Impact

Paired alongside the squad’s Product Manager and Product Analyst, the A/B testing was ran to ensure the stability of this feature and to protect the metrics for Cart as it is a high traffic page. The test ran as follows:

• Dates: 21 June - 24 July 2023 (33 days)
• Countries: BD, LA, PK, SG, TH, KH, HU, CR, SE, TR (APAC and Europe)
• Traffic: 50% Control, 50% Variant

Hypothesis

• H0: By having Cart entry point, Participated rate will stay the same.
• H1: By having Cart entry point, Participated rate will increase

🔥 Results
+24.67%▲ Start rate (stat-significant)
+16.75%▲ Participated rate (stat-significant)
No negative impact on guardrail metrics: GMV per user / Incentivised orders per user / Subscription rate per user

---

👩‍💻 Post experiment

With these promising results, this feature was rolled out to all countries on 30 August 2023 🎉

To prepare for roll out, I worked on the web version and tidied up the UI structure to get it ready for development.

As usual, I also prepared a detailed design documentation to handover to engineers, including walking them through the design requirements. Here are some screenshots:

Thanks for reading!