Google Pay Help & Support

In 2020, as our user base increases, Google spent a huge amount of money annually to support GPay India Support and agents to answer calls daily, with 25% of them abandoned. We need to relook at our support strategy holistically and provide superior help experience to users so that they don't need to call the agents.

_____

THE CHALLENGE

Trust were eroding, and support was overwhelmed

To understand this situation fully, Google is not responsible for any of the stuck transaction, as they partner with a payment rail to enable users to transact. Any transaction stuck is beyond our control, we need to cater the right resolution for our users based on their issue intensity.

Feedback pouring into socal media when transaction fails on GPay India

___

MY ROLE

I led the end to end support strategy and solution design implementation from Mid 2020 ~ End 2022, and handed off the long term to the team to continue the strategy.

In addition, I worked alongside a team of UXers (Content writer, Researcher, Visual Designer) with a group of front-end and back-end engineers, a Support product manager (on agent side) and a product manager (app side).

The end to end strategy was released in phases from late 2020, and fully completed in June 2022.

Phase 3 Brainstorming sesion led by me

_____

KICK OFF

EARLY INSIGHTS

Phone is a divisive channel we want users to use as the last resort. The budget set for 2020 only covers part of the projected incoming volumes of users joining GPay app through acquisition strategy driven by other teams. Having more users abandoning the calls when we are unable to reach the will lead to even more frustrations, thus eroding trust.

We need to prevent more contacts through a strategy to pacify users in the app, and if needed, deflect for issues we can't help even on phone calls.

_____

PHASE 1

Evaluating the entry point

We identify the areas users will need to seek help, and phased it out to focus on the highest contact rate we have for various issues. To ensure there's help available in the app, we need a landing page discoverable for users, called Smart help.

Design principles to guide the design

Accessible

Get help when a problem of a transaction/FOP issue comes up.

Clarity

Users go to Smart help to resolve an issue. Help should be front and centre for them to find what they need.

Control

Always give updates on when the latest of an issue is being updated

Smart Help

Making it discoverable

Based on the highest issue types that users are calling in for, we created video content to help assist users who are looking for self-resolution upon landing. Help center articles that has a high CTR are surfaced here too.

_____

PHASE 2

Automated self resolution

Beyond just surfacing content to find resolutions, we introduced ticketing system for users to stay in context to the payment issues they are having. These tickets can be then used for manual review from agents in the background, sending updates to them for next steps.

In-App Ticketing

Concept proposal

Implemented and launched

Prioritzation of notification/content

If users come in at a state we could predict they have an issue, we want to ensure we give them the right guidance to get help they need.

_____

PHASE 3

Early learnings

From the launch of Phase 1 and 2, we are able to see a 16% contact drop rate, and 2% of MAU created tickets to support their journey of self resolution before reaching out to call centers. As the early phases were focused on pressing issues, our coverage of other app or payment issues needed to be extended.

Smart Convo

Beyond automation

Similar to SmartHelp, a conversation paradigm was introduced with the help of SpeakEasy team integration of automated agent support. These conversations help users to create a ticket, and prompt whether any futher help from an agent is needed.

_____

HOW WE GOT THERE

Constant workshops, testing, metrics collection

To rope in the entire team on a single mission for a phased out vision, constant working sessions were led by me to ensure everyone's opinion and learnings are gathered to improve in every phases. These workshops are done online and collaborated on FigJam.

In Phase 1, we collected ideas from everyone and map it out. Any low hanging fruits were collected for implementation/testing

In Phase 2 (Together with retrospective) , we revisited the different features we have tried in Phase 1, and relook at the main CUJ we have achieved to improve further

_____

NOT JUST ALL ABOUT WORK

In 2022, as COVID measures loosen, the team was able to reunite in person and do a look back on the last 2 phases we have achieved. We reconnected by sharing our great/worst support experiences to remind ourselves why we are still working on what we do everyday.

We see a commonality between all of us: Like our users, we just want to get resolution in the most efficient manner as possible.

Help playbook

For the strategy planning to continue long term, a playbook was created to ensure different XFN teams look at the different solutions available. By understanding what kind of issues issues will face in their feature journey that may impact on call center volumes, we guide teams to ensure product excellence in their work, and be prepared to work closely to ensure the right support is given to their feature release.

As the playbook contains sensitive information that can't be shared, it has been omitted.

_____

THE IMPACT

Reduced the contact rate and kept the abandoned rate at a steady percentage. We gave back Google part of the allocated budget in 2022.
Reduction on contact calls achieved by H1 for the year 2022. EOY target revised to push and drive the target further.
Initial rollout of IAT delivered a significant reduction in contacts. Users who created an IAT ticket did not reach out to support through other channels.
Launched Smart Convo, a AI-based chat to increase support coverage and increase interactivity.
Strategy continues to scale with more solutions to adapt to product growth.

Report abuse