SmartShift, a Mahindra & Mahindra Ltd. subsidiary, helps small business owners in India rent delivery trucks on demand. After only a year, SmartShift was facilitating more than 200 delivery jobs per day across Mumbai and Hyderabad.
The SmartShift app, however, was struggling to gain traction with users. In order to improve the company’s scalability, the app’s process for booking a truck for a delivery needed to be easier than phoning the call center. I redesigned their app to make it easier and more intuitive for their growing customer base.
The SmartShift app had been developed in a hurry. In order to support their growing business, the company contracted with an agency who didn’t have the time or the resources to adequately understand the needs of the user base. As a result, the app had severe usability issues.
As the only in-house member of the product team, I was tasked with redesigning the app in order to increase the app’s share of delivery bookings (as opposed to those of the company's call center).
Defining the Problem
The app needed to be simpler. My first step was to get to know the product by drawing it, screen by screen. This reinforced what I already knew: the existing booking flow was unnecessarily complex. Ten items of information were collected in order to book a single truck. Before I could be sure of what I could cut out, I needed to get to know our users better, so I wrote a set of interview questions, and set out with SmartShift's field staff to interview our customers on-site at their shops and warehouses.
My interviews helped me understand our customers and our areas of opportunity. Our customers were small business owners, mostly self-made, often without formal education. Some supplied plywood or cement to construction sites, others ran small factories for office furniture or sold mattresses and home furnishings. Many had never used a smartphone to access the internet, though they did own them for making calls and using Whatsapp. Below is an example of one of my user personas.
HCI papers on non-literate users helped inspire my solutions. Most of my users had difficulty reading. Simply changing the reading level of the copy, however, wasn't enough. In addition to difficulty reading, non-literate users also struggle to understand abstract representations of a concept and to build a mental model of an interface. These cognitive differences are compounded by cultural factors. Many illiterate users have a fear of "breaking" the device if not used "correctly" and feel generally intimidated by technology as a result of their social status. Reading HCI research helped me to develop a design that empowered even my least capable users.
Some of the most helpful papers included:
Knoche, H., & Huang, J. (2012, May). Text is not the enemy-How illiterates use their mobile phones. In NUIs for new worlds: new interaction forms and interfaces for mobile applications in developing countries-CHI 2012 workshop.
Medhi, I., Cutrell, E., & Toyama, K. (2010). It’s not just illiteracy. India HCI/Interaction Design & International Development, Mumbai.
Medhi, I., Sagar, A., & Toyama, K. (2006, May). Text-free user interfaces for illiterate and semi-literate users. In Information and Communication Technologies and Development, 2006. ICTD'06. International Conference on(pp. 72-82). IEEE.
Uber’s booking flow did not fit our use case. While Uber's design had the simplicity we needed, its first screen (the map with pin) did not make sense to our users, since our users did not have a mental model of how a digital map worked. Our users did not have experience with the spatial reasoning required to evaluate their physical location on the map.
Note: When I started this project, I was teaching myself how to prototype. I experimented throughout with different levels of fidelity, so more detailed prototypes do not necessarily mean later versions.
Simulating a chat bot created more problems than it solved. Since so many of SmartShift's users were familiar with Whatsapp, I designed a prototype that duplicated Whatsapp’s UI and combined it with the interactivity of a Messenger chat bot. The new layout would be familiar to our users, but because booking a delivery truck was so different than sending a message to family, the input methods (typing an address, selecting a truck) became confusing instead of reassuring.
A highly visible navigation made users feel confident. Our users were uncomfortable with smartphones. They needed to feel in control. So instead of following the conventional wisdom that a content-focused layout drives engagement, I put the users' comfort first by making a “command center” home page, in which each button had only a single function. This 1:1 mental model had very little functionality buried within the navigation heirarchy. By exposing all the key actions on the home screen, users could better understand the app’s functionality at a single glance. This approach enabled SmartShift to better communicate the advantages of booking with their service and also made users more comfortable exploring the interface.
Booking became a breeze. By storing the user’s past booking preferences, the user now only needed to input two pieces of information before booking a truck: the location of the drop off and the desired truck size. Setting the pick up location using a map and pin was unnecessary because users and drivers preferred to explain their exact location over a phone call (plus, pick up locations were almost always the same as the GPS location of the device). The user could also duplicate a previous booking or add additional drop off points to his delivery route, making the booking action both easier and more powerful.
Exposing the navigation addressed customer concerns. Our platform had several advantages that should have made users excited about SmartShift, but were often lost in tabs or the nav drawer of the earlier app. Now with our new, more informative home screen the user could see key features like truck location tracking and our rate card. The ability to track the truck’s location made users feel safer about sending expensive goods with our trucks, while the rate card allowed them to easily compare our price against that of their local drivers.
The app redesign is still pending. As the company has grown, other technical priorities have have kept its engineering team busy. The app redesign is projected to go online by the end of 2017. My design for the truck booking process required six pieces of information for a new user, and only two pieces of information for a repeat user, an efficiency improvement of 40% and 80% respectively.
I have several ideas about what I could have done better on this project. One mistake was not translating my prototypes into Hindi and Marathi. It is a challenge to test usability if your customers barely know the language of your prototype. I also wish that I had started with lower fidelity prototypes to explore different directions more quickly. If I had done that, I might have spent more time refining my proposed solution.
Note: I worked on several smaller projects for SmartShift during my time with them, including a style guide for their web and mobile presence.
I am very grateful to the entire SmartShift team. Thanks particularly to Kausalya and Pranshu for giving me my first professional opportunity in interaction design. I would also like to thank our hard-working and knowledgeable field team for being so generous with their time during research and testing.