As the leader in voice biometric rollouts, we spend a lot of time talking about how successful our clients are and the advantages they gain using biometrics. (Check out Year of Biometrics, News about the Australian Taxation Office, and Barclays) Today we are taking a slight detour to discuss our lessons learned from the rare – unsuccessful rollouts.
What makes an unsuccessful rollout?
First, you need to understand what deems a rollout to be successful. For success, we typically look to increase the security of the current authentication solution and/or reduce agent handle time. Success can be measured in ROI, increase in CSAT scores or, in some cases, a lower turnover rate of agents.
While it is said that success has many parents but failures are orphans, it is important to identify and acknowledge what makes rollouts unsuccessful. The way we see it, it’s broken down to three essential issues:
- Low Adoption Rate (Customers choosing not to opt into the system)
- High False Rejection Rate: Bad User Experience (Customers failing to authenticate)
- High False Accept Rate: Low Security Rate (Not secure enough)
Everyone that has rolled out a system is aware of these problems, but often the policies and tradeoffs they make in creating the system have a dramatic effect on the success. So how do we identify these risks? Here are four essential factors found in unsuccessful biometric roll-outs:
Factor 1: Missing agent training leads to low adoption rate
Agent training is essential to a successful rollout because your agents will determine whether or not to recommend this new service to your customers. Agent training should take place as soon as they have a prototype of the biometric solution and at least 30 days before launch. Any new agents should receive the same training. Check out Manulife’s infographic used to train agents.
Agents need to understand how biometrics is going to benefit the caller from both a user experience and security perspective. For many users, this will be their first exposure to biometrics so agents need to know how to answer FAQs from users. Agent-led enrollments are also key to improving adoption rates for any biometric solution. If the agents stop pitching enrollment, the adoption rate will decline, thereby diminishing the value and ROI of the solution.
Factor 2: Missing security assessment leading to bad user experience
The security assessment teams need to be engaged before the design cycle of the biometric solution is completed. The earlier the correct teams are engaged, the less of a risk there is to redesign the enrollment and authentication process. Running a security and risk assessment will address all concerns and provide mitigation tactics without affecting overall user experience.
Security teams tend to lean more towards the “increased security” aspect of biometric solutions. This means that they are more likely to set the system to a very strict authentication setting for all transactions, causing false rejections, thereby affecting user experience. The users lose faith in the biometric solution based on their “bad experience”.
Another common trend is adding too many steps to the enrollment process. Enrollment should be quick, easy and secure at the same time. If a user must jump through hoops to enroll, that creates an overall bad user experience and eventually a low adoption rate. By providing the security team with all the information and biometric rates training early on, we can work on finding that optimal setting where the system is secure enough and friendly enough to accept a true caller’s authentication.
Factor 3: Missing system optimization leads to bad user experience & low security rate
System optimization is highly recommended on a monthly basis during the first year of launch. After that, a yearly optimization is recommended for any security solution to ensure the system is still operating at the desired thresholds.
There are several factors post-launch that can affect the operating points of a biometric system such as: new caller population, change in infrastructure, adding a new language, adding new transactions, ramp up with more speakers and more. The benefits of optimizing a system include:
- Ensuring that the solutions continue running at peak performance: driving consistent security and business benefits
- Preventing against slow erosion in performance that becomes more effort-intensive to fix later on
- Triggering optional optimization activities based on what’s uncovered (such as enrollment strategy changes, or agent scripting changes)
One of the biggest risks, as in any security system, is to lower the security rate – simply by not continually optimizing the performance. This may invite potential fraudsters to break into the ‘neglected’ system. The other symptom of poor system optimization would be the increase of false rejections causing users to lose faith in the system due to “bad experience”.
Factor 4: Missing communication plan always leads to low adoption rate and often bad user experience
Internal communications need to begin at least three months before launch, and an external communication plan absolutely needs to be lined up at least one month prior to launch. If the end user doesn’t know to expect a biometrics solution, this will result in low adoption rates as well as bad user experience because they are caught off-guard when being asked to enroll.
Communicating the benefits of using biometrics (see this page from Virginia Area Credit Union) will help increase the enrollment rate and create awareness around the solution. Defining the rollout strategy is a big part of communications and will help control the uptake of users enrolling. Executive management needs to be aware of the solution so they are equipped with the knowledge to explain it. Campaign materials, surveys and media hype are extremely beneficial when rolling out any new solution to drive adoption. If the user is not aware that there is a new solution out there for them to use, and that it benefits them, they will be hesitant to use the system, and worse, they will refuse to enroll if they do not understand it.
Each of these taken separately can doom an otherwise successful rollout, but taken together they practically guarantee failure.