← All case studies

Case study · Daily Harvest

Loyalty Program

Role
Lead designer: app experience and program design
Timeline
6 months, 2022
Focus
User data and iteration, creative problem solving
+26%ordering rate at launch
+42%in-app engagement at launch
+9%third-box order rate, short term
2.8xreturn on investment, short term

The problem

Daily Harvest's business goal was retention. 40% of customers churned after their second order within 12 weeks, and 45% after their third. Data showed that customers who engaged with the app in their first 12 weeks had 60% higher lifetime value, yet the app was built around function and lacked an emotional connection with the people using it.

Who we were designing for

Finding the right idea

A design workshop started from one question: how might we help customers cultivate a habit of continuously adding fruits and veggies into their diet with a fun and easy experience? Two hypotheses came out of it: a reward program that incentivizes ordering and eating healthy food, and a social community where customers co-create food and improve the flavors of the selection.

I prototyped both and tested them in two co-creation focus groups and five 1:1 interviews. The rewards program stood out as the more exciting motivator. Customers felt it would push them to eat more fruits and vegetables, teach them about their impact, and deepen their relationship with Daily Harvest. Customers early in their lifecycle were open to co-creation, but later-lifecycle customers were harder to excite and needed more incentive.

What success looks like

Two metrics: increase the third-box order rate, which signals habit formation (after three boxes, customers are much more likely to keep ordering), and increase in-app engagement.

The program

Each part of the program had to deliver value to customers and to the business at the same time.

Designing with data, then fixing what the data showed

Content and hierarchy were shaped by both qualitative and quantitative customer insights. A card-flipping interaction came out of a collaborative working session with engineers.

After the first release only 20% of users discovered the new feature, even though 82% of the people who visited the page scrolled past its halfway point. The feature was good and hard to find, so I designed a new entry point on the home screen. Encouraged by the engagement, I then added personalized recommendations to the box, built from user data, company data, and an algorithm.

Testing and iterating

Customers loved the personalized recommendations, especially being told why a reward food was offered. It made choosing easy. Two things needed work: reward progress was too complicated to calculate, and redemption took too much effort. Customers had to add the rewarded item and check out on two separate pages and could not change the reward from the cart. The MVP became a customer journey that guides people to their third box: enter the program with the first box, track consumption progress and personal nutrition, get reward prompts after the second box, then redeem the reward on the third.

Working across the company

The program took PM, data, UX writing, engineering, research, marketing, and creative. It also reached well beyond the digital product: customer service for digital experience and program Q&A, the culinary team for the amount of fruit and vegetables in each SKU, legal for what we could claim, finance for a positive ROI on the redemption model, and supply chain for adding one extra item to the box. I pitched the reward model to C-level leadership and won approval.

Results

The Loyalty Program launched across app and web with seven-plus teams and increased the ordering rate by 26% and in-app engagement by 42%.

The first release showed the early signal in the short term: third-box order rate up 9% and a 2.8x return on investment, with weekly app engagement up 5%, net promoter score up 6%, and 86% engagement with the new feature.

The full story

The full case study as I present it. Use the arrows to flip through the pages.

Not loading? Open it in Figma.