Voice Search in Mobile Apps: Gaana

This blog is an attempt to perform in-depth analysis of voice search functionality in Gaana application and review their overall Voice Search Experience.

Welcome to a new blog series by Slang Labs, called "Voice Search in Mobile Apps". Here we tear down the voice assistant and search functionalities that have been added by businesses to their mobile applications and discuss them in detail. In this edition, we breakdown the voice search feature in 'Gaana' Android app owned by Times Internet.

We believe that it's essential to recognise the trendsetters and show how they are adding voice assistants inside their applications and what is the result of it. This analysis will also help developers understand voice search usage and the importance of voice assistants in applications.

Logo of gaana.com
Logo of Gaana.com

This article will talk about one of the most extensive music streaming services in India, Gaana. It is one of the crown jewels in Times Internet's kingdom with more than 100 million monthly active users. It has witnessed a nearly 100% year on year growth.

Gaana Home Screen with Voice Search icon
When you are building for Bharat, regional languages are critical. Language selection screen is present during onboarding.

App For Bharat

Gaana offers unlimited and free legal access to 45 million songs across Bollywood, International and 30 Indic languages including Hindi, Tamil, Telugu, Marathi, Kannada, Punjabi, Malayalam, English, Bhojpuri, Rajasthani, Bengali, Assamese & Oriya. In 2016, about 75% of the total music was of Bollywood genre. Now, it's about 50%. Regional music has grown to 35%, and finally, English music contributes about 15%.

These numbers are not surprising seeing a lot of Gaana's growth comes from Bharat (Bharat refers to Tier 2 and Tier 3 cities). Gaana's usage growth was the fastest in tier-two cities (96%), followed by tier-one (84%) and then tier-three cities (78%).

Voice Search

According to Prashanth Agarwal, CEO of Gaana,

Today, 22% (now increased to 25%) of our users overall have started using voice search but if you look at the data for [just] new users, that [percentage] is much, much higher. It could be because of convenience.”
“If you look at teenagers on WhatsApp, there's been a big shift in people [sending] voice messages and not typing them out. For two big user segments, the tier II and tier III customers and the teenagers coming online, the voice will be a very big thing.”

Gaana’s version of Voice Search

Functional Breakdown

Even if the album name is said, Gaana figures out and plays a song from that album.

Tap and Play

Perhaps the most significant difference in search via touch and voice is that the song plays directly when you search via speech. Whereas in touch, you get a list of choices where you select the song you want to play. When searched via voice, the application automatically matches the utterance to the closest song and plays it. If the utterance doesn't correspond to a particular song, the application plays the closest possible match. Voice search is optimised for voice to action, and it prioritises taking action than being conservative.

What works

Currently, the only function, that works is "play".  A user can try different utterances to play a song, such as 'Play Arijit Singh songs' or a particular genre or mood or the song/album's title. If any other action verb is used, the system either ignores it, or it breaks.

Visual Breakdown

Gaana Voice Search Demo
Screenshots of the Voice Search feature inside the Gaana App

Position of the mic button

Gaana has placed the voice search button on the home screen and bottom in the application, making it accessible and easy to reach. This centric position on the main screen improves the discovery of the mic button by the user.

Training the user

On clicking the button, are ever-present 'coach marks' like, Try saying 'Play BLOW', which guides the user to tell the utterance in a particular manner. These suggestions are also personalised to the user.

Traditional Search bar

A traditional search bar is present in the voice search window along with one at the top of the app. In the author's opinion, this is done for two reasons: First, making sure people who clicked this button after reading 'search', can still find desired results by typing. And second, to give the user an alternate option if voice search did not provide the correct result.

We have done a detailed analysis of how YouTube integrated voice assistant in their app to improve user experience!

Things that can be improved:

Multilingual support

Ability to speak in multiple languages will give better results primarily to cater to those 30 languages they serve.

More Voice-enabled features

Ability to do more actions in the app using speech. E.g., "Take me to my downloaded songs."

Reducing the number of taps

Today it takes two clicks to search via speech. Reducing it to one-click will surely improve the user experience.

It is evident from this analysis that there is a clear need for robust and effective voice assistants inside mobile applications to reduce taps. Voice Assistants save a few taps and make the applications easy to use by non-conventional users.

However, Gaana has implemented a minimal use case of searching for songs via speech; there is a considerable scope of improvement. Let's hope Gaana improves its voice search experience for its customers, especially when that feature is seeing such high adoption.

Slang allows you to add voice search in multiple Indian languages in a fraction of the time with inbuilt analytics to gauge how users are using it. It also gives businesses the ability to add more voice-enabled functions quickly to the app. Make your app ready for Bharat.