Designing for better user expectations and improving user satisfaction on Postmates

Isabelle Su
7 min readDec 16, 2020

In the time of COVID-19, many new users have started using on-demand food delivery. I wanted to take on this project to understand how I could better their user experiences now that so many are turning to this service. I chose to redesign Postmates because I found when viewing engagement metrics on similarweb.com of top on-demand food delivery services, Postmates was not a successful as its competitors. I sought to learn why and on the way, I discovered that the biggest issue in on-demand food delivery users face is setting user expectations, which led to low user satisfaction. My solution is to give users more information up front before making their order, so that they can make better decisions on where to order and what to order and in turn be more satisfied with their decisions.

Getting to know the users

The very first step in my process was to interview users and understand their behaviors, goals, and pain points when it came to ordering on-demand food delivery. I conducted 3 user interviews with users who didn’t use Postmates and 3 user interviews with Postmates users and was able to distill my findings into two user personas: the Brand Loyal user, and the Brand Agnostic user.

Personas

Uncovering my problem

Through speaking with users, I found that the two had both overlapping and differing goals and pain points. I chose to focus on their overlapping pain points, so that I could solve an problem that affected both personas. The common pain points I uncovered were when users would receive the wrong orders or missing items, facing extended wait times, and confusing menus. Then I had my aha moment. I realized that all of these pain points were tied back to the users’ expectations! There was a disconnect between what the users expected out of their experience versus what they actually received. Therefore, I decided to focus my project on solving this problem.

How might we set user expectations for their on-demand food delivery experience to increase user satisfaction?

I had a vision that adding ratings/reviews to the Postmates flow could help solve this problem, so created a sprint map of the user experience to help me locate the area of opportunity.

Sprint map

My responsibilities and constraints

The project was conducted in an 8-week design class where I was responsible for all of the tasks. This also means that the work was constrained to the time and resources available to me. For further validation, I would like to consult with a developer, product manager, and possibly other stakeholders.

Initial round of ideation… The first of many

To work through the problem, I started with a Lightning Demo ideation session and evaluated websites and apps that I thought did a good job of setting user expectations.

First, I went to Yelp because all of my users had said that they went here to get more information on the restaurants and their food. Yelp utilizes user generated content, such as ratings, reviews, and photos. The ratings and reviews give users detailed stories and information. The photos show users exactly what the food is going to look like, a clear way to set user expectation.

Next, I went to Zillow to find how they present the listing information in a way that users can make a judgement on which one could fit their needs. I found that they give users a clear list of facts and features on each listing. Users can rely on the fact that they can go to Zillow and compare every listing based on these facts and that these facts clearly set their expectations for what the listing is like. In the end, I chose to redesign the Postmates mobile app with a remixed version of all of these features.

Lightning Demo and Crazy 8's

With my potential solution in mind, I started wireframing what the new features would look like in the existing Postmates app. I mapped out two task flows that could solve my problem.

Task 1

You are interested in ordering from Full House Café, but you have never had their food before. Navigate through the app to read reviews on the restaurant.

Features added in Task 1

  • Ratings in restaurant profile
  • User generated photos in restaurant menu
  • User generated reviews
  • Graphs depicting percentages of users who thought the food “Arrived as expected” and was a “Good value”

Task 2

Imagine you have just received your food and after opening it, you’ve discovered that the food did not meet your expectations. Give the restaurant a rating and review.

Features added in Task 2

  • Rating, review, and add photo experience that starts from the users’ recent orders

Finding out what went right and what went wrong

With my wireframes done, I redesigned them in high-fidelity, created an InVision prototype, and took my designs to test with 6 users who fit my user persona. My main goal was to see if my problem of setting user expectations had been solved by my solution. The key questions I asked during testing that gave me the most insights were:

  1. Would you leave a review if you felt that you had a negative experience?
  2. Did you find the ratings, reviews, and photos beneficial in giving you and understanding of what you would receive if you ordered from this restaurant?

The Good

  1. 6/6 users said that they liked seeing the reviews and photos uploaded by other users. They thought it was helpful in giving them more context into what the orders would look like and what other people thought about the food. The photos would help them make the decision on whether to order from that restaurant or a different one.
  2. The rating and review screens were easy to follow.

The Bad

  1. 6/6 of the users made close guesses on what the labels “Arrived as expected” and “Good value” meant. However, it would have been a better experience for them if the definitions were clear. One user gave me a great piece of insight and said that she thought it would be helpful to know what percentage of users would order from that restaurant again.
  2. 3/6 of the users had difficulty locating where to start their review. They expected to start from the reviews tab. When asked why they thought it was there, they referenced how Yelp’s experience had the user start their reviews from the restaurant’s profile.
  3. 3/6 of the users said the tag menu item screen’s + and - icons were confusing and felt that a checkbox would be a more commonly understood affordance. They also thought that to reduce the amount of items listed, they should only see the items that they ordered.
  4. 2/6 of the users expected to see a confirmation screen after submitting their review. They also expected to see a history of their completed reviews.
  5. 3/6 of the users noted that they would probably not leave a poor review unless the food was truly terrible.

My second chance

I took the usability test findings and made changes accordingly to my design. Below are the key changes I made to better solve my problem.

  1. Users said that they would not leave a review if they had a negative experience, so I removed the public reviews. Without consistent use, the reviews would be skewed.
  2. Users said that photos were more important to them in helping their decision making, so I still allowed users to add photos and I created a “Photos” tab under the restaurant profile.
  3. I changed the “Arrived as expected” and “Good value” ratings to “Would you order again?” This makes sense because this reflects the users’ true mental model. If they are not happy with the service or food they receive, they do not support that business again.
  4. Instead of starting reviews from recent order, I chose to take the user straight into leaving a review the first time they open the app after receiving their order. I predict that this will push more users to use this feature.
  5. I redesigned the ratings screen to give users a set of options depending on the star rating that they give. I believe that the quick options will prompt more users to give feedback. Users still have the option to leave a comment for Postmates to view.

What I’d like to do in the future

Some features that I would like to test in the future that I omitted from the second iteration are:

  • The ability to view a history of your added photos as well as a way to remove the photos
  • Leaving a rating from the restaurant profile

I would love to test my second iteration with users to find out if my design solves the issue of setting user expectations. It would be interesting to do a study that compares user satisfaction of the current Postmates app with user satisfaction of my redesign. I think it would also be very interesting to explore and dig deeper into research on the psychology of human behavior around expectations to better improve my designs and increase user satisfaction.

--

--