Connect. Chat redesign
Restructuring existing in-app chat module, updating UI kit and implementing new features
Problem
The Connect is an internal B2B application for an international company. It helps to inform its employees from all over the world about company news and educate them about new products and sales information through gamification. Chat is a communication channel in the Connect app. Through this module users receive content, different notifications and have an opportunity to contact tech support.
Problem 1.
Content and notifications engagement via chat was reduced.
Problem 2.
Users complained about the slow work of tech support, and tech support cannot speed up the processing.
Why did this happen?
Сhat was developed at the very beginning at the product launch stage and since then it has not been updated to meet the new business requirements and user needs. Content, notifications and communication with tech support occur through one single screen, and since the amount of content has increased, users lose some info in the feed and do not interact with them.
How I found a solution
Analyzed the "content" problem.
Collected statistics on user interaction with content over the past 6 months. Filled in the factors: the time and day of the week of publication, the amount of time between publications, the presence of a cover image, the presence of a button, the presence of a call to action in the text. And correlated this data with activity.
Analyzed the "tech support" problem.
Studied how long users wait for their problems to be solved through technical support. Discussed with support managers the most common questions that can be answered without involving managers. Found out from the support team what data they are usually requesting. Reviewed the current script and decided which questions from the manager can be automated or grouped.
What was done
Content and communication with tech support are different in functionality, so we decided to split these features into separate communication channels ("Notifications" and "Support"), but leave them in the same section of the application. Considering how the product will scale we added the possibility of adding new channels for business needs.
"Notifications" channel visually made like a feed, because this is a more common pattern of consuming a large amount of content. The UI was updated for each type of content: we updated the display of images, added an unread message marker, added the ability to design a title, added buttons and attached new types of actions to them (playing a video, redirecting to another section of the application, launching a mini-game, opening articles, etc.).
The content plan was updated based on user activity: added rules for publications (time, day of the week, frequency), added headings and prepared cover image templates for each type.
"Support" channel was made more convenient for both users and support managers. Questions from users were divided into categories, had their own id, name, status. After updates communication with tech support started from a chatbot that collects general info about the problem: device info, type of problem, section of the application and so on. If users had popular questions, the bot offered to study one of the guides that we wrote for each problem. If users could not resolve the issue on his own or the problem was unknown, a manager communicated with the user live.
Results
The number of "skipped" content decreased by 31%. Users consume content evenly and the average number of active users interacting with content reached 87-91% due to the updated content plan, new content types and actions, unread message marker, cover image template and call to action header.
The sessions number increased by 9% due to the fact that regular content is published on a schedule and users open the application more often, knowing that there will be a certain publication on a certain day. And for the same reason, the average session duration decreased by 3%, as users consume content and almost immediately exit the application.
The number of technical support questions increased by 40% due to the creation of a separate "Support" channel and the ability to add multiple tickets at the same time. Of these, users solve some of the questions on their own through a chat bot without involving a manager, thanks to guides.
At the same time, the speed of processing the questions decreased from 72 hours to 8 hours and the time spent on processing by the support team decreased by about 50% thanks to the chat bot, guides and tickets system. This is also made it possible to reduce the load on the team and not expand the staff.
The most important thing for us was that this update allowed us to build a regular system for collecting statistics on questions from users. And it became an additional way to collect data and insights for future updates.