Measuring the Results of Interactive Voice Messaging in the Contact Center

By Matt Edmunds, VP, Collections and Financial Services

Maximizing your agents’ time and productivity is key to a successfully run contact center. Different elements contribute to its success or failure, and it is important to evaluate the strategies being implemented to maximize efficiency. Integrating new technologies such as Interactive Voice Messaging (IVM) into the contact center will boost agent success through more right party contacts, more warm inbound calls, improved agent morale and motivation and an overall better experience for both the agent and the consumer on the other end of the phone. New developments in speech technology are also contributing to the effectiveness and overall call experience for both parties.

Developments in speech technology have done away with mechanical-sounding voice recording and replaced it with a friendly, human voice. Contact centers now have the choice of “who” they want speaking to their customer base: male or female, young or old, choice of accent, etc. This degree of personalization now available with interactive voice messaging technology is a major push forward.

Each step toward a less robotic, more personalized phone call through IVM directly impacts the experience of the consumer receiving the phone call, and in turn, the agent experience once the consumer is directly connected to the contact center. A consumer’s experience will be at a higher level if the call received is from a professional, non-invasive voice, that correctly pronounces his/her name, calls at an appropriate time, validates that he/she is the correct party to be speaking with and ultimately provides the option to speak with a live agent. At this point, the agent is being put in contact with a willing, satisfied consumer to complete the process for which the initial call was placed – a win-win for both parties.

Of course, taking advantage of the new technologies will benefit the contact center, but not without a careful evaluation of its value before implementation. In business, we’re often challenged to move so quickly that we can’t take time to measure the value of decisions and strategies. Unfortunately, this dynamic inhibits our ability to deploy the most optimal strategies. Managers of contact centers need to create an approach to ensure they’re measuring the value of the different channels they’re utilizing to communicate with customers. There are two critical common denominators of those companies who have truly adopted Interactive Voice Messaging (IVM) as a dominant part of their call strategies. First, these companies took a rigorous approach to measuring value. Then, they measured the value based on profit rather than on cost. This is critical.

Businesses can take a simple 4-step approach to measuring and evaluating the value of voice messaging to ensure they’re getting the optimum results from it. The process begins with defining your objective. It sounds easy, and it is, but you’d be surprised how voice messaging technology is sometimes misunderstood, and thus, measured and used incorrectly. You have two choices for objectives: either drive more revenue, sales or responses, or at least maintain your current revenue, sales or responses while decreasing cost – thus, increasing profitability. If you desire to increase your revenue, sales, or responses solely, then you should utilize the Supplement strategy. If you desire more profit, then you should utilize the Substitute strategy which, quite frankly, is the most powerful application of the product when used correctly.

In the Supplement strategy, a contact center simply adds voice messaging to the existing contact strategy to drive more revenue, sales or responses. Use this strategy on portfolios or campaigns where performance is below expectations. The reason being, you may be performing poorly due to a lack of regular and well-timed penetration. Voice messaging will increase your penetration in total and during more varied times, which generates more right party conversations. However, if you’re already performing at a successful level, then adding voice messaging to your strategy may bring in more right parties, but the return-on-investment (ROI) might not be overly compelling as the cost to collect or upsell the small increment of “able” and interested consumers still out there is typically quite high. It is important to remember your primary goal with the supplement strategy is to produce more absolute revenue, sales or responses with a potential for greater profitability.,

On the flip side, if you are looking for increased profitability and a high ROI, then you may be better suited for the Substitute strategy. In the Substitute strategy, a contact center would add voice messaging and remove dialer and/or manual outbound calling, . This is an effective strategy because you will be substituting all the wrong party talk time that occurs today in your outbound call center with right party contacts. Your penetration will also be maintained because of the unlimited calling capacity you gain when integrating voice messaging. This strategy is most effective on portfolios or campaigns that are performing at or above expectations in terms of total revenue, sales or responses, yet are in need of a higher profit. On these portfolios, you’ll measure profit, profit margin, and/or ROI. The most common pitfall when measuring the value of IVM is that users will deploy a Supplement strategy to drive revenue, sales or responses but measure results of a Substitute strategy (profit or ROI). Know your objective, strategy, and which measure(s) are associated with achieving that objective.

After establishing which strategy is best suited for your campaign or contact center needs, it’s time for step 2: Set up Your Test Design. There are two general approaches, one “Ideal” and one more “Simplified”. In addition, there are a few very important points to consider in order to ensure your results are valid. The “Ideal” design is the most thorough and accurate as to eliminate a number of backend calculations and assumptions. The “Ideal” design requires well thought-out planning and perhaps programming, yet it eliminates the need for back-end work. Essentially, this approach begins with the creation of two groups within your targeted portfolio. The “Control Group” is simply whatever your strategy is today. The “Test Group” is your proposed strategy based on your objective from step 1.

There are really four critical considerations for this test. First, keep accounts separate throughout the test phase. You’ll need to setup separate codes or classifications, but this will help you track contacts and revenue over time – particularly from callbacks. Second, ensure the size of the test and control groups are large enough to eliminate such variables as random variation, full time employee talent, etc. By dedicating at least two teams to both groups, you alleviate this issue. Third, ensure the test strategy incorporates voice messaging calling during prime and non-prime times and days. This will ensure you’re simulating a rollout production environment, and this will test your strategy at the best and toughest of times. Finally, it is important to have dedicated teams to the opposing strategies. Teams can switch sides to eliminate the risk of uneven talent, but when you have dedicated people on the voice messaging campaign, you’ll gather greater user feedback and really get a feel for how many people to dedicate to the accounts in order to reach your desired contact rate or penetration. This is critical to help determine the optimal usage within your operation. The most common mistakes in setting up this design are that accounts cross-pollinate between test/control which skews results measurements, and sometimes the test group size is not large enough or truly random which hurts the credibility of the results. Be rigorous in your planning process.

Some contact centers do not have the time, resources, or infrastructure to create a true apples-to-apples test design. If so, it is possible to draw conclusions with a more simple design. You must capture historical business-as-usual results/metrics of the portfolio in which you’ll test. The key metrics to capture are simply determined by the objective discussed earlier. In addition, you may want to track metrics such as right party contacts per hour and/or cost per right party contact. However, the real key is to know the profitability and daily performance of that portfolio or campaign. When capturing historical data, be sure to use a bare minimum of three months worth of aggregated data to normalize for performance variation, anomalies, and seasonality.

Then, kick off your proposed new strategy. Some contact centers have run tests for as little as a few days, though a multiple-week or month long test is optimum to increase validity and control for variations.

Be disciplined in running your strategy. Implement checks and balances so that you can determine that the designed strategy was implemented just as you designed. The trick now is normalizing results for the test period. Effectiveness can vary by month, by time of the month, by day of the week, and by time of day. Therefore, if you ran a 2-day test on Thursday and Friday, your results would look very different than running a 2-week test that captured weekends and all days of the week. If you test on Thursday and Friday during mid-month, you should really compare test results to historical Thursdays and Fridays during the middle of past months. This creates more calculations and assumptions for you which is why the “ideal” design is preferred. Most critically, if your objective is to test the Substitute strategy on a marginally profitable business, be sure to raise your account-to-agent ratio before you begin. This is the ultimate value proposition were you do the same or more with fewer agents. Every situation is different, but a 20-30 percent increase is a good recommended starting point.

Finally, you will come to steps 3 and 4: Managing the Test and Measuring the Results. The most successful and valid tests tend to occur when there’s one person within an organization that is directly responsible for administering a fair and valid test. Ensure this person is evaluated on the validity of the test, not the results or effectiveness of the test. Challenge this person to create tools such as daily/weekly/monthly progress reports along with call templates to ensure the calling strategy was executed as intended every day. Within these reports should be quantitative results/metrics that relate to the test objective.

The final piece to discuss is what precisely should be measured. Measuring profit seems simple, but there are pieces to the cost equation that are often overlooked. First, remember that your control group has telephony costs. Be sure to include today’s telephony costs into the control. Second, and often the most overlooked when employing the Substitute strategy, is to include FTE costs. If you’ve increased your account to agent ratio, then you’ve saved on labor costs. You absolutely must include labor costs into the comparison between test and control. For shorter tests, you can track the labor hours dedicated to complete the calling, and then simply extrapolate that over your entire portfolio to understand long-term P&L impact. Finally, remember all the other “savings”, some quantitative and some qualitative, of utilizing an On-Demand ASP solution in your test: unlimited capacity, flexibility, enterprise control, outsourcing much of your IT/Dialer personnel, etc.

Here is a very simple example of a third party collections agency test that was run over a 3-week period. Notice that it is a very straightforward P&L within a Substitute test.

Don’t overcomplicate your measurement criteria. If you want more revenue, sales or responses only, then add voice messaging to business as usual and track these metrics or total right party contacts. If you want the same or more revenue, sales or responses with lower cost, then add IVM , reduce dialer and/or manual calling, remove some labor, then measure profit (i.e. value). As you look to make IVM a larger percentage of your contact strategy, slowly peel off that band-aid called accounts to agents ratio. You can either reallocate FTE or simply let attrition occur without backfilling it over time. You’ll replace the lost labor with increased IVM dialing productivity and track daily, weekly or monthly results. If performance is maintained or improved, then let the accounts to agents ratio go a little higher. Continue with this strategy until you find the optimal point on your portfolio. Remember to put things into a true P&L perspective or calculation, and in making final decisions around value, be sure to consider all the other advantages associated with full-scale production of an ASP OnDemand product. The key to maximizing the efficiency and profitability of a contact center includes the implementation of key technologies, as well as implementing the appropriate testing and value measurement strategies to ensure the best outcome. Hosted voice messaging solutions such as SoundBite’s OnDemand Customer Communications platform offer the flexibility, capacity and ease of use that can immediately result in a positive impact on your customer communications campaigns.