iSixSigma

The Final Tollgate: Age Verification Cycle Time

For online betting organization Betfair, based in the United Kingdom, scrupulous identification of customers is vital. Within 72 hours of creating a Betfair account, the customer’s age must be verified to be 18 years or older. Betfair’s verification process is a necessary part of protecting customers and the business, and for complying with relevant U.K. legislation. For the vast majority of customers, this is a non-intrusive background process. In some cases, the customers must contact a Betfair help desk agent to manually complete the process. Slow cycle time and resulting customer complaints regarding this process prompted Betfair to do a rapid process improvement project, based on the DMAIC (Define, Measure, Analyze, Improve, Control) roadmap, on the cycle time of age verification. The project was a finalist in the Largest-breakthrough Improvement Projects for the 2010 iSixSigma Live! Awards.

Define

When a customer creates a Betfair account, Betfair systems automatically run an electronic identity check through third-party software. To verify that the customer is at least 18 years of age, the software simultaneously matches the Betfair account information with independent data sources, including address records, birth registration and financial information.

Customers will not pass the check if they are: a) underage, b) the information on their account is incorrect or false, or c) they have recently moved and only been at their registered address a short time. Customers who are not successfully verified may contact Betfair’s customer service help desk for details and assistance with achieving age verification through a manual process.

In this case, customers must provide additional identity documents, such as passports, utility bills and credit cards, in order for the checks to be run. The software has the unique ability to authenticate identity documents and, again, cross references the additional documents for a clear picture of whether the applicant is a bona fide customer.

Betfair’s customer service help desk operates 24 hours a day, from its U.K. office between 08:00 and 23:30 GMT and overnight from its Malta office between 22:30 and 08:30 GMT. The Betfair help desk deals with calls, emails and web chats. The company established verifications specialists (VS) – trusted operators with access to advanced tools and permissions – in an attempt to handle the manual age verifications quickly. Help desk agents pass on age verification-related customer queries to these specialists, who either verify the account if the customer’s documents are in order, or make a note in the customer’s account that the age could not be verified. A specialist then notifies the help desk agent, who in turn contacts the customer with this information.

Handpicked Content :   HMO Formulary Standardization Six Sigma Breakthrough

Despite the specialists’ efforts, customers complained that they experienced a lengthy delay before their verification queries were resolved. Some of the complaints included:
• “My query couldn’t be handled so the operator promised to ask someone else and email me. Three days later I have no email.”
• “The operator was not much help on how to verify my account.”
• “I faxed my ID three days ago; why is my account still closed? I want to bet on a race today.”

Across Betfair’s websites, the company advertises that customers will receive a reply to their query within 4 hours. From October 2008 to April 2009, the company received 1,425 verification queries; only 39 percent received a response within the service level agreement (SLA) time. The average internal response time between verification specialist and help desk agent was 9 hours and 21 minutes and the average time for the customer to receive a response to their query was 14 hours – 250 percent longer than the SLA.

After creating a high-level process map, the team used a SIPOC (suppliers, inputs, process, outputs, customers) diagram to clarify and define the individual steps of the process. The SIPOC also was used to define the following critical-toquality characteristics (CTQs) for the end-to-end process:

• Accurate and complete resolution of issue in accordance with the published SLA
• Customer receives acknowledgment of email
• Correct performance of software
• Verification specialists and help desk agents understand and adhere to procedures

Handpicked Content :   A Six Sigma Case Study - Tutorial for IT Call Center - Part 3 of 6

Financially, the customer impact was difficult to quantify as it required making assumptions about what customers would have done if they had been age-verified more quickly. To assess this, a random sample of customers handled by the verification specialists in the baseline period was examined to determine how long after verification they placed a bet, and what event, sport or product they bet on next. This analysis showed that 48 percent of customers contacting the help desk for verification could reasonably be assumed to bet immediately, and the annual missed revenue opportunity of verification delays was £46,000. This gave the project an estimated £26,300 value based on reducing the average internal verification response time by 57 percent from 9 hours and 21 minutes to the SLA of 4 hours.

Measure

Using the seven months of historical cycle time data collected in the Define phase, the team conducted normality, capability and stability studies in order to baseline the current process performance.

An Anderson-Darling normality test was carried out in order to better characterize the process. This resulted in a p-value of less than 0.005, indicating non-normal data. Knowing the data was not normally distributed, the project team carried out a capability study against the SLA of response within 4 hours. This showed that the average internal response time was 9.2 hours, with 610,217 defects per million opportunities (DPMO) and a disappointing
benchmark sigma level of -0.28.

Process stability was assessed using a run chart followed by an I-MR control chart, which indicated that the process was not in control, and that staffing to customer demand played a key role in the out-of-control points. To fully understand the process’ current behavior, the project team created a more detailed process map. This gave a graphical depiction of the steps in the process and enabled the team to identify non-value-added steps.

To detect potential root causes of the response delay, the project team conducted a brainstorming session with subject matter experts. With the information gathered during the brainstorming session, the team developed an in-depth fishbone diagram.

Handpicked Content :   Coronary Artery Bypass Grafts: Six Sigma Breakthrough

A number of quick wins were identified. Without further analysis, the team took immediate actions to:

• Notify all managers and help desk agents of the capabilities of the verification specialists.
• Update and correct the notification/education literature around the office.
• Allow the Malta night shift help desk agents to run electronic checks for Betfair customers and update customer accounts.
• Raise a request with Betfair’s technical department for an alert to be sent to the verification specialist when an incoming verification query is received.

Analyze

A failure mode and effects analysis (FMEA) was used to evaluate the risks of the process in more detail, specifically to prioritize root causes, modes and, by extension, solutions based on the risk priority number for each. The FMEA was built using the process steps in the detailed process map and causes in the fishbone diagram.

Subject matter experts rated the severity, occurrence and detection of each failure to establish the highest risk areas within the current process. The FMEA confirmed the conclusion from the detailed process map that there were no current process controls in place at most of the steps in the process. Action plans for the top eight causes were then devised, which also addressed some of the remaining seven failures because many of the problems benefited from the same action. A minority of causes could not be resolved as they were related to a value-added step, or one that needed to remain “as-is” in order to comply with Betfair’s policies.

Improve

Process changes implemented in the Improve phase were based on the FMEA prioritization completed during Analyze. The recommended actions developed in order to eliminate or reduce failures were rolled out in the U.K. and Malta offices.

Key solutions included:
• VS were re-trained and issued formal role guidelines, which provide step-by-step instructions on everyday queries such as how to identify a genuine European passport. The guidelines also include details on more obscure queries, such as how to complete a change in a customer’s name after marriage.
• The training team was advised of up-to-date details regarding the VS, the duties they carry out, and their contact details and hours. In addition, the training team incorporated verification information in the help desk induction training so all new employees start their role knowing the correct process. The training team also is responsible for replenishing educational posters and literature regarding verifications placed prominently around the office.
• Help desk agents were re-trained. The existing help desk agents received training, guidelines and easy-touse templates that detail the information they must obtain from a customer with a verification query, the importance of collecting the correct information so the final outcome is accurate, and how best to forward the information to the VS to ensure the shortest cycle time.
• The scheduling team was made aware of the VS and the number of specialists needed per shift to guarantee that the SLA is always upheld.
• The Malta location night shift agents were trained to handle all aspects of their own verification queries. This enabled them to answer and resolve all verification queries in real time, resulting in increased customer fulfilment and reduced cycle time.
• Standard operating procedures were introduced to monitor the verifications queue on a half-hourly and daily basis.

Handpicked Content :   Building a Business Case for Software Defect Reduction

Using cycle time data collected in the two months that followed the Improve phase, a number of statistical studies were performed.

A capability study was conducted against the SLA of 4 hours. The run chart’s observed performance showed that out of a million queries, 69,383 fell outside of the desired SLA. In the Measure phase, the DPMO was 610,217; thus, there has been an 89 percent reduction in defects. The current process sigma level is 1.48, which is a positive 1.76 sigma shift on the previous level of -0.28.

Process stability was reassessed using the I-MR control chart, and the results show that the difference between the upper and lower control limits has decreased dramatically. Therefore, the process is now far more stable. There are two out-of-control points, which are due to special cause variation. Shift swaps had been approved without taking into consideration the skill sets of the agents, leaving the help desk without a verification specialist for a whole day on two occasions. To address this issue, the scheduling team decided to use a different color to show the VS on their spreadsheet so they can easily see the schedules of the specialists. This will make it clearer when approving shift swaps and annual leave.

Handpicked Content :   Call Center Quality: Satisfaction Over Quantity

Control

A Mood’s median hypothesis test was used to ascertain if the verification cycle time improvement is statistically significant. The result was a p-value of 0.000, which confirmed significance. The end result is a reduction in average internal response time from 9 hours and 21 minutes to 1 hour and 40 minutes. Given the defect reduction achieved, the actual realized financial benefit was £40,100 annually. Beyond the statistical studies completed to validate the improved age verification process, we also received positive comments from customers through an online survey. Some comments received were:

• “I was annoyed about having to email my ID, but it was sorted quickly and I got my bet on Manchester United. Thanks.”
• “The operator was polite and knowledgeable. It was no hassle to get my card unlocked and get in the poker tourney.”
• “The girl called me back within the hour and my account was open. I can bet on Aussie racing now so I’m happy with the service.”

This project has also improved morale on the help desk; after being re-trained, the agents are now more confident in dealing with verifications contacts. The help desk agents know if they receive a verification query it will be dealt with by the verification specialists quickly and efficiently, and the customer will receive great service.

You Might Also Like

Leave a Reply