A Primary Matters Total Cost of Ownership/ROI Analysis . . .
The Impact of a Next Generation Voice Recognition Solution on Credit Union and Banking IVR Applications
August 10, 2002
Primary Matters, Inc.
Table of Contents
Voice Recognition solutions continue to evolve in quality and ease of use. This analysis focuses on the approach offered by the next generation of voice recognition applications, which may be developed using the emerging application development platforms from a variety of companies. Typically, the companies developing these platforms are beginning to offer XML-based scripting languages ready for building packaged applications. They also use underlying voice technology from companies such as Nuance and Speechworks.
The goal of these platforms is to make it easier for an organization to include voice recognition in a company’s portfolio of customer service options. The goal being a significant impact on both cost savings and increased customer satisfaction.
Just as Touch-Tone based Integrated Voice Response (IVR) has revolutionized the interface to customer service centers over the last 10 years, voice recognition solutions are likely to do the same. Those companies, who embrace the technology making it a seamless part of their service environment, will find their costs per customer decreased. These companies will also find greater customer loyalty and increased new customer attraction. This will increase their market share.
This analysis describes the financial and customer relationship impact of replacing a classic banking Touch-Tone IVR application with voice recognition. The usage data describing the Touch-Tone application is based on actual data. The post-voice recognition changes are based on analyzing the impact that voice recognition would have on the use of touch-tone applications. In this analysis the bank receives 100,000 calls a month from a base of 200,000 customers. The impact of voice recognition can be summarized as:
Impact on the Customer Base and Financial Performance of a Small Financial Service’s Retail Banking Business
The following table and graph summarize the impact of voice recognition. The customer base grows, and revenues and profits increase significantly. The product costs increase due to the larger number of customers remaining with the bank longer.
This graph illustrates the impact on the major financial measures of the bank.
Impact of Voice Recognition on Bank’s Financial Performance
The information in this document is based on a small financial services company offering retail banking, loans and credit card services to its customers. A Touch-Tone based IVR application is used as the first point of contact to all callers. They can choose to go directly to Contact Center Agents, or to use the IVR application to receive the information and service they need.
The basis of the data and estimates in this document are:
The Call Model
The following table illustrates the major flow of calls into the organization (based on actual data). It also provides estimates of the impact of voice recognition on call flow.
BENEFIT 1: Shorter period of time using IVR due to simpler, faster entry of information using verbal responses instead of Touch-Tone responses
These changes in the IVR usage time and the reduction in call transfers to Agents are quantified in the next several pages of this document.
BENEFIT 2: Higher Use of IVR Scripts for Customer Service and Fewer Transfers to Customer Service Agents During IVR Usage
Two major themes are present in the increased IVR productivity from voice recognition technologies when compared to Touch-Tone based IVR applications:
BENEFIT 3: Higher Customer Satisfaction reducing churn and increasing new customer acquisition rates
Quantifying Benefit 1: Time Savings due to simpler, faster entry of information using voice recognition
The following tables outline the usage of the most important of the IVR menus, and compare a Touch-Tone IVR Application environment with a Voice Recognition Application environment. The duration of a session using voice recognition is 54% shorter, 52 seconds less per caller, than using Touch Tone. This benefit takes about 16 months to phase in, as the Caller’s become accustomed to Voice Recognition and begin to use it more. The impact of upgrading to voice recognition is summarized in the following table.
This analysis highlights the impact of voice recognition on the four menu choices that account for 70% to 80% of the typical IVR application usage in a credit union or small bank. These choices are the ‘Greeting Menu’, ‘Caller Authentication’, ‘Retail Banking Menu’, and ‘Checking Account Status’ menu.
Usage Patterns and Durations, by Prompt, for Touch-Tone and Voice Recognition-based IVR Application
Analysis of Calls Transferring to Agents Comparing a Touch-Tone Interface with a Voice Recognition User Interface
Guide to the IVR Application Menu Analysis Tables
The following terms define the table headers that follow.
The Initial Greeting Menu
Current IVR Application:
This example IVR menu typically greets each caller to a financial services company when their main number is called. Of the new call arrivals 75% of the callers go through this menu and 25% immediately make their choice, either by dialing ‘0’ to go directly to an agent or by choosing another option going directly to the next IVR application menu.
With a Touch-Tone User Interface:
Under the Touch-Tone IVR Applications approach, the average time spent on this menu is 43.8 seconds.
With a Voice Recognition User Interface:
Most callers to a small bank or credit union are reasonably familiar with the IVR application content and know why they are calling. These callers will be able to state “I want to talk to an Agent.” or “I want my checking balance.” and the application will pass them to the next step, for instance, requesting a User Authentication, saving time.
The following chart provides an example of the new timing associated with a menu using Voice Recognition
With Voice Recognition, the average time per contact for using this menu is 27.2 seconds a savings of about 12.8 seconds per caller.
The number of callers transferring to Agents declines, since more callers will use the Voice Recognition interface. They don’t have to remember their Account ID and PIN, and the interface is easier when driving and using cell phones.
The number of callers listening to this menu declines significantly with voice response, since a caller can state, “Checking Balance” and bypass this menu without having to know what to key into the telephone set. The estimates of use of this menu are:
The User Authentication Menu
Whenever a user chooses to go for the IVR Banking application menu or they immediately choose to talk to an Agent, they are requested to enter their Account Number and PIN for Identification. Of the calls arriving at the bank’s main number, this typically represents about 75% of the callers.
Current IVR Application:
With Voice Recognition
This Voice Recognition solution is based on the emerging voice authentication recognition standards that can provide highly secure caller identification by requesting the caller’s name, and then asking them to repeat three words. This approach has been proven to be as, or more, accurate in authenticating callers than the current approach used with Touch-Tone.
There will be a slightly higher usage rate since those that cannot remember their account number and PIN will be able to fully identify themselves, and the time required to accomplish the identification is about half of that required by the traditional touch-tone user interface.
The IVR Retail Banking Main Menu
After successfully being identified by the system, this example menu is typical of the IVR menu that customers use to acquire account information or do other simple services, such as activating banking cards or re-setting their Personal Identification Numbers (PINs):
Current IVR Application:
With Voice Recognition
With the Voice Recognition solution users will no longer be restrained by a structured menu approach. For instance, after the caller’s authentication is passed, they will state their interest (as described above) bypassing this menu entirely.
The Checking Account Status Menu
This example is typical of the menu of the IVR applications that customers use to acquire account information or do other simple services:
Current IVR Application:
With Voice Recognition
Two factors affect the calls transferred to Agents:
These impacts are detailed in the tables presented under Quantifying Benefit 1.
A financial institution that increases its IVR platforms ease of use and shortens the time it takes for a user to receive benefits will increase customer satisfaction. Changes to customer churn, especially in the credit card industry, may have a material impact on financial results.
In this document, the initial customer life cycle is assumed to be 48 months. We assume that the average customer life cycle lengthens by 20% to 60 months due to the major improvements in communicating with the financial institution. This phases-in over a sixteen-month period, as the users become familiar with the voice recognition user interface, and then become dependent on voice recognition!
The voice recognition solution is presented as an ASP service, as opposed to a customer-installed option. Voice technologies are evolving quickly, and such an approach enables the solution provider to continuously upgrade to further improve customer service.
This is an expensive approach costing $400,000 per year. (See the detailed table below.) The return due to this solution justifies such a major increase in costs.
This graph shows the Total Cost of Ownership (investment, expenses, operations and personnel costs) associated with implementing the voice recognition solution compared to the cost savings obtained by the financial services company. Break even is reached in three quarters, leading to a positive Return on Investment in the first year.
The increase in revenues, not accounted for in this cash flow and breakeven analysis, is the most important benefit. However, it is not necessary to include this benefit to justify the investment in a Voice-Recognition Solution.
This chart shows that the cumulative Total Cost of Ownership of the voice recognition solution, and the cost savings to the small retail bank cross into positive territory in the fifth quarter after deployment.
The break-even is based on a 16-month phase-in for the bank’s callers to become familiar with the voice recognition solution and change their usage patterns. If this change is faster, the financial benefits to the bank will be accrued more quickly.
Financial Implications: Return on Investment, Based on Internal Cost Savings Only
The Return on Investment for voice recognition applications is positive in the first year, increasing significantly in years two and three. This chart does not include the benefits of the revenue increases, which provides a much more compelling case for voice recognition solution.
This chart shows the annual profit increase due to voice recognition’s impact on the customer life cycle. A dominant part of the increase is available in the later years as the impact of lengthening the average customer life cycle begins to increase revenues.
Financial Implications: Increase in Profits Due to Higher Customer Satisfaction
This chart shows the annual profit increase due to voice recognition’s impact on the customer life cycle. A dominant part of the increase is available in the later years as the impact of lengthening the average customer life cycle begins to increase profits.
Summary of Assumptions and Methods
The model used is of a generic, smaller financial institution. Although specific banks may differ, such differences can easily be accounted for to better understand the specific impact of Voice Recognition on an individual bank.
In providing this Business Impact Analysis, Primary Matters has:
The financial institution has customers using Credit Card and Retail banking services. In this study, the voice recognition solution affects these businesses in several ways. Details of the calls, volumes and changes in these due to the solution are provided later in this document.
For the revenue and product cost analysis, it is assumed that the improvements in customer satisfaction from better service will lead to:
Contact Center Agents Compensation and Benefits
Customer service agents receive an average compensation of $13.00 per hour and work a 40-hour workweek. They are provided full benefits, two weeks of vacation, 10 holiday days and require three weeks of new agent training at the start of their employment. There is one week per year of ongoing agent training. The average job duration is 24 months.
Contact Center Supervisors Compensation and Benefits
Supervisors receive an average compensation of $17.00 per hour and work a 40-hour workweek. They are provided full benefits, two weeks of vacation, 10 holiday days and require three weeks of new agent training at the start of their employment. There is one week per year of ongoing agent training. The average job duration is 30 months.
Operations Personnel Compensation and Benefits
Operations personnel are those who manage, maintain and run the technology for the bank or credit union. Operations personnel receive an average compensation of $55,000 per year and work a 40-hour workweek. They are provided full benefits, two weeks of vacation, 10 holiday days and require three weeks of new agent training at the start of their employment. There is one week per year of ongoing agent training. The average job duration is 30 months.
A simple, installed base and per customer revenue model is used in this analysis. Each customer has a Retail Banking account, offering checking, savings and money market services, and a credit card account. The average revenue and costs of supplying each of these accounts, excluding the costs of the customer contact center and IVR applications follows:
The details underlying the revenues and costs per customer are detailed below.
New Customer Acquisition Costs and Average Active Customer Life Cycle
The average cost to obtain a new customer is $120. This cost includes marketing and sales-related costs as well as new account set-up costs. The average customer life cycle - the time that the individual is a customer and that their accounts are actively used (thus creating revenues and profits for the bank) is 48 months.
Detailed Assumptions: Credit Card Account Revenue & Product Costs
Detailed Assumptions: Summary of Banking Account and Loan Revenues and Product Costs
This budget comparison shows the pre- and post- voice recognition contact center budget. There is a material impact on the cost side of a bank’s major lines of business.
Scenario (After Voice Recognition Solution) Operations Budget
This chart shows the ROI for the voice recognition solution based only on the cost savings in the Contact Center. The investment is justified solely on the cost savings, even though the most material benefits are the revenue and profitability increases to the Bank.
NOTE: For ROI Analyses a Cash Budget is used. The annual and other budget figures differ from other charts in this report since this is a cash-based budget, and the others capitalize the investment.
System Investment, Systems Integration, Maintenance and Operations Personnel Expense Details
This Full Time Equivalent Agent Headcount comparison shows the pre- and post- voice recognition Contact Center staffing for Agents, Supervisors, and Operations Personnel.