Kount LogoBLog Against Fraud

Optimize Manual Reviews with Agent-Based Reporting

posted on: Tue Jul 18 2017

For virtually all eCommerce operations, manual reviews make up the largest percentage of the fraud mitigation budget. Fortunately, there are a number of best practices you can employ to make your human agents more efficient and reduce the costs of conducting manual reviews. Review

Salaries, benefits, and tax impacts of agents are typically the biggest contributors to overall review department expenses. If you employ the 7 best practices listed below, your agents will not only increase efficiency, but will also increase sales through reduced false positives while they catch more fraud.

  1. Analyst Screening & Training
  2. Event Notification System (ENS)
  3. Custom Weblink Integration
  4. Custom Reason Codes
  5. AI & Machine Learning
  6. Agent Automation
  7. Feedback Loop with Agent-Based Reports

In fact, eCommerce operations that have helped their reviewers become manual review rock stars report huge improvements:

  • Agents able to review up to 2X more orders
  • Up to 75% improvement in accuracy

This blog post will focus on how a feedback loop with agent-based reports can help your agents double their productivity while dramatically improving results.

As Peter Drucker, renowned productivity expert, has famously said: “If you can’t measure it, you can’t manage it.”

From a departmentwide perspective, it’s important to know how agents overall are performing when it comes to manual reviews. But on an individual level, agent-based reports are essential to helping your agents become manual review rock stars.

An underlying assumption is that you are using a fraud prevention system that tracks activity on an agent-by-agent basis, including specific task and goal metrics. For example, these might incorporate:

  • Time spent per review
  • Number of resources accessed during review
  • Reason codes invoked (by percentage)
  • Decline rate as a percentage of reviewed orders
  • Chargeback rate as a percentage of reviewed orders

From an overall department standpoint, the type of insight you get from agent-based reporting helps you make the most out of your team’s strength. To begin with, you can more easily identify your strongest agents and route the most critical transactions to them. For example, if your review queue is prioritized by urgency (e.g., orders shipping next day) or by ticket amount (e.g., with the largest transactions at the top of the queue), you can auto-route these orders to your best performing agents to ensure maximum productivity and accuracy.

Further, if you see agents repeatedly taking certain steps on orders with similar data profiles or invoking the same reason codes over and over, you could set up a process that triggers an automated review routine (if your fraud prevention system possesses this capability). For example, if agents are consistently confirming delivery addresses on orders where the billing and shipping addresses differ, you could set a rule to automatically query third-party data services and perform these address checks without agent involvement. You could then set rules to pass on this automatically-collected information to an agent, saving them a step in their review and accelerating their efficiency. Or you could set rules for the system to auto-decline if the shipping address turns out to be problematic, bypassing a manual review agent entirely.  

What about individual agent improvements? One of the other best practices involves agent training. Agent-based reporting can help you make the most out of this training. For instance, if you conduct weekly “lunch-and-learn” sessions, you’ll know in which areas to focus attention (based on agent-based reporting results). Thus, if you’re seeing a high number of chargebacks with certain types of reviewed transactions, you can spend more time during the “lunch-and-learn” on tips, tricks, and techniques to do better on these type of transactions.

And when it comes to training on an individual, one-on-one basis, agent-based reporting can help you zero in on the areas where a specific agent may need more help. For example, if an agent demonstrates a blind spot around geographic or regional warning signs, you could follow up to make sure they fully understand how to use tools like custom weblinks to Google Map data.

Learn how to approve more orders and optimize the biggest spend in your fraud prevention budget in our eBook "7 Best Practices to Become a Manual Review Rock Star".

New Call-to-action
Topics
Share
TOP