Has your business experienced problems with fraud? If so, you’re not alone — and if not, it’s possible that you just haven’t looked hard enough.
A 2020 study from PwC found that 47% of businesses had at least one incidence of fraud in the previous two years, with an average of six instances per company. Customer fraud was the most common, followed by cybercrime, asset misappropriation, bribery and corruption. The incidents were almost evenly split between internal and external perpetrators. Altogether, the losses for the 5,000+ businesses surveyed amounted to $42 billion.
With the destabilising nature of the COVID-19 pandemic, the likelihood of fraud is even more prevalent. A report from the Association of Certified Fraud Examiners (ACFE) found that 77% of companies surveyed as of August 2020 experienced an increase in fraudulent activity since the start of the pandemic, with 92% expecting fraud to increase over the next 12 months.
Companies are rushing to fast-track the onboarding of new vendors, sometimes without proper vetting. And with many businesses flying blind in their newfound remote work environments, it can be easy to overlook the financial checks and balances that might occur in the corporate office. But if you’re not looking closely, it can be easy to miss situations where your ledger doesn’t balance out correctly.
Fortunately, data analytics can spot warning signs that humans sometimes miss. Here’s a look at some key types of data inconsistencies that data analytics solutions can spot from insider and outsider forces.
General Ledger Entries
Ledger entries should be scrutinised closely for potential fraud or errors. For instance:
1. Identify and Search For Suspicious Keywords
Identify suspicious journal entry descriptions using keywords that may indicate unauthorised or invalid entries.
2. Stratify General Ledger Accounts
Look at the normal range of values to spotlight entries outside that range — if your average ledger entry for payroll is $2M and one shows up as $500K, the entry may warrant further scrutiny.
3. Hunt for Outlier Journal Entry Amounts
Use your analytics solution to flag entries more than two standard deviations away from typical entries, which may indicate an accounting error or potential fraud.
Business Expenses
Expenses in areas such as travel and entertainment are often where unscrupulous employees may fudge numbers through multiple methods. Here are a few ways to monitor for fraud in expense tracking:
4. Analyze the Average Spend by Department
Your sales department likely has much higher travel costs than your programming team. Monitor department spending over time to understand the average range for each division and set up an alert to trigger if the department deviates from that range. Sometimes, a specific seasonal event, like a conference, may throw off the balance, so do a line-by-line analysis to confirm whether everything is on track.
5. Check for Suspicious Reimbursement Claims
Fuel and mileage reimbursement are common areas where employees may (often inadvertently) double-dip. If they’re already submitting mileage for reimbursement, fuel charges should be flagged as a duplicate item.
6. Identify Split Purchases
Employees may skirt around purchase limits by breaking one large purchase into two different line items — your analytics solution can alert you to such “split purchases.”
7. Watch for High Group Dining Expenses
Historical records give you a good baseline level for average group dining expenses. A record should be flagged for review if it is out of the normal range.
8. Find Round Amounts
See an expense for a round amount, like a $200 charge? These charges aren’t likely to happen by chance, so it may mean that the employee is using their company card to take out cash advances.
9. Identify Duplicate Claims
Some employees may submit expense reports for purchases they’ve already charged to their corporate accounts, getting reimbursed twice for the same expense.
10. Deactivate Dormant Company Cards
If an employee leaves your company but their card is still active, it may be susceptible to fraud. Make sure these credit cards get shut down immediately.
Payroll
11. Flag Suspicious Salary Increases
Look for multiple small salary increases for the same employee within a single calendar year and contact HR to verify that they are approved.
12. Identify Phantom Employees Still on the Payroll
Are employees who have left the business still showing up on the payroll? This may sometimes mean another employee has modified the payroll to their bank account details to conduct financial fraud.
Vendor Payments
Contractor payments are rife for fraudulent behaviour, as either the vendor or an unscrupulous employee can funnel money through fake or inflated invoices. Here are a few tips for avoiding vendor fraud and billing oversights:
13. Segregate Vendor Creation and Billing
Ensure the employee creating vendor profiles isn’t responsible for generating and approving invoices. A dishonest employee can use this role to create fake invoices from fictitious vendors, as can an outsider who’s gotten access to system credentials.
14. Flag Duplicate Invoices
Vendors may submit the same invoice multiple times, either by accident or to follow up on unpaid bills. If you don’t have a system for tracking and flagging duplicates, you may pay the same invoice twice.
15. Watch for Frequent Changes to Vendor Details
For example, if a master table includes frequent changes to a vendor address, this is a potential indicator of fraud.
16. Identify Matches Between Vendor and Employee Details in Your Master Table
Pinpoint matching bank details, address or other data between a vendor file and employee payroll details. This may indicate that the employee has set up a fraudulent vendor account.
17. Find Non-PO Purchases Above the Dollar Limit
Most companies associate purchases above a certain threshold with a purchase order. If you notice large purchases that aren’t, they may indicate fraud or an accounting or vendor error.
18. Track Blanket Receipts
Blanket receipts involve invoices for multiple services that the provider has not yet rendered. Make sure that, if vendors are issuing payment in advance of service or receipt of goods, you have a way to track the delivery of their work.
Customers
Fraud and accounting errors can also cause problems when working with your customers. Here are a couple of ways to verify accurate accounting with your customers:
19. Validate Customer Credit Limits
Ensure all your customers have consistent credit limits in line with their business credit profile and company size, flagging those with outsize credit limits. Those customers may not be able to pay back outstanding loans.
20. Identify Sanctioned Customers
Run your customer accounts through a database that flags sanctioned customers from public records. This helps you identify customers who need further validation or should be cut off.
Next Steps: Preventing Fraud With Data Analytics
There are many ways to identify fraud and accounting errors within your business, but you’re likely to miss a lot if you only review transactions manually. Errors, unauthorised charges and signs of fraud can slip past your accounts payable team’s sight, potentially leading to millions of dollars of losses over a year.
To accurately detect and quickly remediate fraud before it significantly impacts your organisation, it’s essential to implement best-in-class analytics that can analyse 100% of your financial data (instead of sample testing) to identify red flags automatically.
Your toolset should integrate with internal and external datasets, including your ERP, payroll system and government datasets such as census data and other public records that can be used to verify your data’s accuracy. A continuous delivery system, which will instantly provide alerts regarding potential violations, can also use custom rules and triggers designed for your industry. By deploying advanced machine learning, your accounting, compliance, and IT teams will have vital intelligence available to help them keep your organisation secure.
Download our Detecting & Preventing Fraud With Data Analytics eBook to learn more.