Portfolio

Explore

Portfolio

Portfolio

Jenny Wang

Built a better matching matrix for pets owners and sitters

Helped Rover.com to come up with a creative solution that can potentially improves pets owners' requests acceptance rate

Overview

Project Summary

Rover.com is an online marketplace for connecting 5-star sitters and dog walkers near you who offer dog boarding, dog walking, house sitting, or doggy daycare service to busy pet owners over 14,000 cities.

Rover.com came to School of Visual Concepts with a challenge they were experiencing with their app. I worked along 3 designers to help them to come up with a design solution that potentially improves the pets owners’ requests acceptance rate by increasing the enjoyment of using the app and pets’ hosting experience.

My Role: UX designer

Client: Rover.com

Duration: 10 weeks

The Problem

“But, your calendar says you’re available...”

Imagine you're leaving for a vacation in two weeks and you need to find a sitter for your dog as soon as possible, you start searching on Rover app and hope that you'll find a qualified sitter for your dog when you are gone. You find a sitter and contact her/him. However, you got a message from her/him saying, "Sorry, I'm not available." You probably wonder, "But, your calendar says you are available." Rover has been frustrated that pet owners get denied by service providers due to unavailability. Pet owners contact Rover and think it's Rover app's fault.

But, why?

I had the same question when I found out about the problem. From pet owner’s perspective, if you present yourself as available on your calendar, why do you say you're not? Why do you deny requests? What’s with the inconsistency.

The Goal

My Process

User research - the first thing our team did was to start user research to understand more about the problems, and I interviewed both service providers and pet owners.

Ideate - Based on research findings, we started brainstorming and sketching out possible ideas for the problem.

Prototype - Another prototyper and me headed down and made a working prototype to showcase our idea.

Test

User Research

What Service Providers Said:

With all the questions in mind, I started seeking answers by conducting user interviews.

⁠

Quotes from service providers

Quotes from service providers

⁠

Data on reasons for denying requests from service providers

Data on reasons for denying requests from service providers

⁠

The statistics from interviews show 80% of people didn’t deny requests due to unavailability. From that group of people, 46% of them, which was the majority of the group, said they deny because of personal preferences, particularly the type of dogs and characteristics of dogs. Only 18% of them said they forgot to update their calendar, which was the problem Rover has been facing.

What Pet Owners Said:

Rover's mission is to help pet owners find qualified sitters, so interviewing pet owners and understanding the good and bad in their experience are very important. I asked them about their experience with booking a sitter, and here are some of things they said:

⁠

Quotes from pet owners

Quotes from pet owners

⁠

What a Pet Owner Struggled The Most:

I asked Emily to walk me through her booking a service process from beginning to end. I detected a few pain points.

⁠

Journey map: booking a dog boarding service

Journey map: booking a dog boarding service

⁠

Search: Her dog has behavior issues, so there is a lot to looking for in a sitter in order to find the best fit.

Send Requests: She usually contacts 3 - 5 sitters on top of her search list. She said contacting more than one sitter will increase the likelihood of getting accepted by one.

Communication: Conversations usually last for a couple of days. She would talk about details about her dog's characteristics and care routine.

Synthesize

From all the feedback and insights from users, besides some usability issues, we realized that service providers actually use unavailability as an excuse to deny requests. The real problem is mismatching. Service providers don't want to sit dogs that don't match their personal preferences, and the inaccurate matching creates the pain points pet owner Emily experienced in the booking process.

Ideate

Once I knew the right problem to solve for, I started to brainstorm for potential solutions. I started to look for inspirations by doing market research. I found dating websites like eHarmony use machine learning to match people, and there are existing matching surveys for dog adoption online to help people to figure out what type of dog parents they are.

So, as a team, we decided to onboard service providers and pet owners with matching services that can use to generate matching score algorithm.

⁠

Matching algorithm

⁠

Matching algorithm logic

Matching algorithm logic

⁠

This is the framework I came up with. First, for the 1st tier data, pet owners and sitters all have to take a survey that creates generic matching based on dog breeds. For the 2nd tier data, they will have a questionnaire that creates more accurate matching based on the nuances they write. With the two tiers of data, we can train algorithm via successful matching experience that are based on post booking ratings from two parties. In the end, that can generate the matching score.

1st Tier Data

ASPCA (American Society for the Prevention of Cruelty to Animals) Adoption Program takes the mystery out of the selection process by identifying people’s preferences through an online survey, and assessing how a dog is likely to behave upon arrival in your home - all to increase the likelihood that you and the dog find your best fit.

Even though dog adoption is different from finding a dog sitter, the goal is the same, which is to increase the chance of finding the best match between dogs and dog sitters. So, I decided to base off the ASPCA's survey for the 1st tier data.

⁠

2nd Tier Data

Every dog has its own unique personality. This questionnaire will give pet owners the chance to write all the nuances about their dogs, like care routine and special conditions.

Prototype

During our design process, we followed Rover's brand guidelines. We used crazy 8's method to sketch our ideas on the paper first. Everyone talked about their sketches and reasons behind.

⁠

Team members and my sketches

Team members and my sketches

⁠

Here are the wireframes for pet owners. An introduction page will inform pet owners about match survey and what it is used for. From the survey and 'tell us more', we can not only generate matching scores but also automatically generate dog profiles for pet owners.

⁠

⁠

On the other side, service providers would get a similar introduction page to the matching survey. It will inform them what the survey is about and the result will tell them the type of dog sitters they are. In the end, all the answer they input in the survey about their preferences and experiences will generate a profile automatically.

⁠

Test

We were on the 8th week of our project. We didn't test our prototypes in the early stage of the project because we weren't confident about our solution and designs. As we were approaching the deadline, we realized we should have done it earlier because we didn't have time to test the prototype with as many people as we wanted to. Even though it was no metric to prove the solution work, we did get some positive feedback.

"Matching score makes it easier for me to pick a sitter."

"It's very convenient that I don't have to spend a long time to write a polished profile."

"The survey helps me understand more about the type of sitter I am."

⁠

Even though we couldn’t test on that, I personally did further research on the potential impact of matching score. There is a study on the relationship between personality match and pet satisfaction among dog owners from the department of psychology at Oklahoma State University. The result suggests that future dog owners may want to consider adopting dogs who match their personality. Matching the personality of dogs and their owners on a few characteristics may predict owner satisfaction. This is important as it may reduce the number of dogs that are relinquished by owners as well as the number of dogs left homeless each year.

⁠

The result of this study can reflect how matching can also be effective for Rover to reduce requests deny rate by increasing the enjoyment for service providers.

What’s Next?

Test and iterate at an early stage of the project

I learned that doing usability testing at an early stage of the design is very important even if we didn’t feel confident in our ideas. Therefore, I want to do more usability testing and come up with a metric that can prove that matching score and automation in messaging work.

Explore alternative solutions

Because the solution takes time and money to implement, I want to explore other solutions that take less time and effort to improve the app. For example, from the research, 9% of people said they denied because of usability issues. That could be a start for the alternative solution.

⁠