Saving over $1m dollars for a bank by setting up their self-help centre

Helping a bank establish and scale a self-service support website and save over a million dollars by lowering inbound call volumes to its contact centre.

04 Jul 2019


A leading Asian bank had a large call centre fielding all types of questions from customers. The centre ran like a well-oiled machine. They even had voice analytics that transcribed voice calls into text and tried to categorise the calls.

The problem, however, was that the customer effort in calling the bank was high. This was also the time when research reports across different industries showed that customers, especially millennials, prefer self-service options. One report even found that millennials prefer to “clean the toilet than to contact customer service”.

PebbleRoad was commissioned to set up a self-service support website to help customers quickly get answers to their queries (to lower customer effort). We were given a target to lower inbound call volumes by 7%!


As always, we started with research. We sat with the call centre representatives (reps) and listened to the calls. We spoke with the long-serving reps and asked them about the types of customers and types of calls. We studied existing analytics on the operations.

To make sense of all of this data we created a tool called the “Call Query Cube” which analysed the call data across 3 dimensions:

  1. General or specific
  2. Simple or complex
  3. New or known

We used this framework to analyse 6 month’s worth of calls. We got valuable insights. We found that 15% of the calls were self-serviceable. These were general, simple and known calls. A typical query would be “what is the SWIFT code for the bank?”.

After getting our target content, we then mapped the self-service journey. The journey started from a Google search all the way to issue resolution. We even put in a strategy for measurement and escalation.

With the content and the journey mapped out, we took to writing the self-service content.

Writing self-service content is very different from writing articles or FAQs. Self-service content first has to be found, then designed to offer a solution right at the outset and finally guide the customer to resolve the issue. We tried and tested several versions of this type of content before defining a winning version.

With the content written, we had a few more things to do:

  • Design the information architecture of the website so that customers can easily locate their issue.
  • Integrate the content to other relevant pages on the website.
  • Offer an enterprise-grade search engine to not only find the support articles but also to give instant answers.

We helped launch the self-service support website with 120 selected support articles for the personal banking section first. This was our proof-of-concept. With the measurements in place, we were all set to launch.


Within three months of the launch of the self-service support website, the bank saw a 5% reduction in call volume. The reduction went up to 15% twelve months later as improvements were made and more content was added to the website. The overall cost savings from this reduction has crossed over US2 million per year.

To help the team continue building the self-service support website and scale the concept to different countries, we created a self-service support toolkit. This toolkit included the vision, principles, practices and templates. It also included the roles and responsibilities for the self-service support team. The toolkit has proved to be a valuable resource for the team in scaling the self-service support concept and ensuring that the vision and thinking were aligned.

Kickstart your transformation journey.

Be the first to know when we share insights and host events with international experts.