Usability Testing of Google Allo, an artificially intelligent messaging app

Evolution of messaging apps

The fact that Google, Facebook, and Apple are all focusing so intensely on messaging apps isn’t just a fad. Rather, it’s an evolution that is aligned to the requirement of the people today. This usability testing also goes on to prove that the busy users of today require such artificial intelligence.

To start with SMS and MMS was around for decades. It went strongly for years at a stretch. The second era of messaging was brought on by apps, such as WhatsApp, Hike, WeChat etc.. People share messages, chat with each other or broadcast to groups of friends. This is where we are currently.

Going a step ahead of this generation of messaging apps, Google came up with the next generation of messaging apps — Google Allo. The idea behind this smart messaging app is : messaging anyway consumes a massive share of the time we spend on mobile phones, so why not make messaging into the portal for all the things we do on our phone? Which is an excellent thought!

About Google Allo

Google Allo, is a smart messaging platform to chat with your friends and family. It helps you make plans, find information, and express yourself more easily in chat by predicting what you could say. There are plenty of other exciting features such as playful stickers, doodling on an image, varying the text size etc that make the app quite innovative and fun to use.

Objective of the usability testing

Predictive smart replies are one the most highlighted features of Google Allo. The objective of the test was to identify any usability issues in the app while a conversation and understand how effective and useful the reply suggestions are.

User research

User persona

Prior to conducting usability tests, I developed a user persona to better understand the target users of Allo’s Android app. This process helped me get into the mindset of the users, thinking in terms of their contexts, needs, and goals. Largely, the user persona was the existing users of Whatsapp and similar messaging apps.

So meet Marcus!

Platform used for experiment

I have used the CanvasFlip online tool for creating the prototypes and for UX insights such as session replay (the user videos). conversion funnel and heat maps.

No of users in the test

I tested the app prototype with a larger section of the audience just to be sure if there was any issue with the interface and also understand the average time one takes to have a chat of a standard length. So I tested it with 32 users having android mobile devices.

Task scenario, given to the users

I mentioned the most common scenario for a messaging app as the task — “Update your friend about this amazing trip you went on. Send the images and give him details of the trip.”

Usability Testing

I chose un-moderated remote usability testing because I wanted to capture the experience of the users in their natural habitat, that is while lying on their bed or hiding the phone under a desk in class or waiting at the bus stop. Because, messages on these messaging apps are prone to be sent in their natural surrounding rather than in a room or sitting in front of me in a cafeteria.

So here’s the prototype I built for the test.

Prototype tested on :

(Task Scenario for the usability testing : Update your friend about this amazing trip you went on. Send the images and give him details of the trip.)

Open the prototype in a new tab to experience it.

Insights on the Google Allo prototype:

The UX flow for a chat is very simple and clean. It is not overloaded with lots of options to choose from. Most importantly, the design team at Google has kept in mind that the user does not have to relearn the chatting app. With its differences clearly standing out, the users did not face any issue with figuring out the flow of the app. The majority of the users of Allo will belong to the community of Whatsapp users. So, the smileys, the position of send icon, the voice message icon position are not played with.

1.Minimal friction in the flow

If we check out these session recordings of users who used the prototype, we do not find any friction in the flow…

2. Usefulness of the predictive replies

I have two flows in my prototype — 1. When the user types the message and sends 2. User opts for the predictive replies. The analytics show that about 78% of the users opt for the predictive replies. This goes on to say that users are more comfortable with suggestions and more than that willing to use it. I am all the more positive about this because with time the suggestions are bound to get better and better.

There are multiple other features in Google Allo that I am looking forward to testing. But for now, the chatting feature (THE primary feature in a messaging app) works like a beauty!

Final Words

Clearly, Google Allo is in its early days and every reply suggestion will not speak your mind. But it does manage to reduce at least 80% of your typing load. It’s a big leap in the user experience of the messaging app domain. The usability test also shows that users are not struggling to discover the flow of the app. With few other features such as increasing the font size and getting the most out of the Google Assistant might be a bit challenging. My next tests will focus on these aspects of the app.

Jason Cornwell, leading the UX for Allo says —

“The job of designers like me is to make UI go away… we’re just getting to the point where natural language processing and AI is good enough that you don’t have to create UI artifice anymore.”

Related Posts
Monika Adarsh