COURIER RETENTION

Berlin, Germany 2021

Executive summary

In this collaboration with the logistics team at a major food delivery platform in Germany, we wanted to understand couriers’ reasons for leaving the company. We combined my ethnographic research with 41 couriers (a.k.a. riders) and the company’s courier survey data to try to find out how to reduce logistics fleet turnover.

My role: PhD researcher sharing and comparing my findings with company team members from operations and HR

Skills: Ethnography, contextual inquiry, interviews, thematic analysis, quantitative analysis

Participants

Ethnography: 41 riders
Survey: ~600 riders

Findings: The most common frustrations uncovered across the survey and my ethnography were low salary, communication culture, and equipment quality.

Outcomes

Company leadership made courier retention a top 10 KPI
9 OKRs for improving courier retention
Social media campaign prioritizing couriers
Publications and talks

Problem

Why is courier turnover so high? This was the answer logistics management needed to solve. Many couriers were leaving after just three months. With so many last-mile logistics companies competing for couriers in big cities, rider shortages were a constant worry.

I became aware of this problem in the process of doing my PhD research on app-based food delivery in Germany. One of the logistics managers I interviewed invited me to collaborate on his courier retention project. I was thrilled by the prospect of learning how research is used in a professional context.

Research

I had already recruited 41 courier participants over the course of 5 months of ethnographic research. In design research terms, I was doing contextual inquiry. I conducted interviews and in-depth observations, including working as a rider myself. I wanted to understand not only what riders had to say, but also their low-level, everyday experiences that self-reporting can miss.

Meanwhile, the logistics team had conducted a survey with several hundred responses from Berlin riders. In contrast to my exploratory approach, the survey mostly consisted of targeted questions about specific aspects of the job.

Image: source

Insights

I conducted a thematic analysis of my fieldnotes and interview data to pull out the top 3 themes among 123 courier frustrations:

Order planning (27 complaints, 22%)
Communication with supervisors (19 complaints, 15%)
On-the-road working conditions (15 complaints, 12%).

I presented this to the logistics team in a spreadsheet with some illustrative quotes.

I was surprised and fascinated to learn that, while the survey identified many of the same themes, their frequency didn’t match mine at all. In addition, survey respondents' top issues didn't include on-the-road conditions, app functionality, or career mismatch.

This could have been due to our different research designs. I was hanging out with fellow couriers, gathering data on their habitual frustrations in open-ended encounters. In contrast, the survey’s fixed questions — allowing for quicker data collection from more participants — made it to harder to record some aspects of everyday courier experiences.

In addition, our different positionalities were at play. I was not directly affiliated with the company. But when filling out the survey, riders knew their pressing concerns could heard by people who could change things and may have prioritized their frustrations accordingly.

In any case, I triangulated the top 3 issues across our data sets with my ethnographic data to give some insight into the why and how of couriers’ frustrations. Below, I give the couriers pseudonyms.

Low pay

"There should an incremental pay raise based on how long you worked so that you could eventually earn €11 or €12 hour."

Jeremy, 39

The couriers earned €10 per hour plus a small per-order bonus. I recorded 9 complaints about wanting higher pay.

Communication with management

"The dispatchers never understand our side really. We are in the street, enduring chaos or the streets in disrepair, the cold or the heat."

Luis, 27

"Nothing is clear. They don't tell people how things work."

Hashir, 33

I recorded 19 complaints about communication with management. Most complaints were about negative interactions with supervisors (e.g. dispatchers) (13), followed by nontransparency (6) and lack of face-to-face communication (2).

Poor equipment quality

Ana, 27, shows me the thin black gloves the company recently gave her. Today, it’s 4°C and windy, and they're not warm enough. She shows me where the trim around the wrist is already splitting from the glove: “after one day.”

"One day, I broke three bikes. First I broke the chain. Came back, then I got a flat tire. And with the third bike, the mudguard kept getting stuck in the wheel."

Aadi, 20

I recorded 11 complaints about the equipment provided to couriers. These were about the clothing (5), bicycles (3), and delivery backpacks (3).

Image: Carolin Genz

Results

It’s not easy hearing about riders’ struggles, and it can be hard to invest in improvements when profit margins in app-based logistics are so slim. However, I was glad to advocate for the riders in a team dedicated to making things better.

By the beginning of the next quarter, the team lead shared the project results:

9 OKRs implemented to address riders' frustrations and encourage them to stay
HQ made courier retention a top 10 KPI and launched a project to understand the entire courier lifecycle
Promotional hashtag and social media campaign prioritizing couriers, signalling renewed commitment to courier wellbeing

I also disseminated these findings and my wider research in publications (dissertation, blog post, article) and talks at academic conferences and colloquia.

Reflection

This was a good opportunity to reflect on how my academic perspective can be most helpful in a business context. If I could do this over, I would present my findings to the project team differently. I meandered through my thematic analysis, and when I finished, the team seemed a little confused. More preparation in terms of concretizing the main takeaways, how my research connected to project goals, and perhaps better data visualization could have gone a long way.

I would have also asked if there were opportunities to help workshop the OKRs to get a better sense of the context in which the research was being used.