Companies are increasingly turning to data analytics to detect and stop fraudulent schemes, a recent Deloitte poll found.

Nearly 35% of companies are using some type of data analytics to keep eyes on their supply chains, according to the poll of more than 3,200 professionals from a variety of industries.

That’s an increase of nearly 10 percentage points from 2014, and a sign that more companies, large and small, are adopting complex technologies to closely examine their supply chains.

Companies need to constantly look for fraudulent practices in the manufacturing and shipping of goods, with those looking to steal often as crafty and innovative as the technologies that stop them, said Guido van Drunen, a principal in KPMG’s forensic advisory group in Silicon Valley.

“Fraud goes where there is the least resistance,” he said, comparing fraud detection to a game of Whac-A-Mole. Nip one scheme in the bud and another one pops up elsewhere.

A major advantage of using data analytics to examine supply chain transactions is that detection can happen much faster, said Larry Kivett, CPA, a Houston-based Deloitte Advisory partner in the firm’s forensics and investigations division.

Schemes that might have gone on for years, such as a factory operator setting aside a pallet of electronics for black market sales, can now be found out shortly after the theft begins.

Now, complex algorithms can sift through thousands of incoming invoices and work orders from third-party suppliers to find instances of double-billing, overcharging, and theft.

That saves companies money, especially those with complex, global supply chains with hundreds of vendors and suppliers to keep track of, Kivett said. In many cases, the fraud can be detected before an invoice is paid, instead of having audits years down the road reveal the problem.

Advanced algorithms can also match bills of lading, forms filled out at ports that reflect cargo shipments and times, to ensure that goods are leaving when and where suppliers say they are with the correct number of manufactured products, Kivett said.

“If you can leverage data analysis to help out, you’re getting much earlier visibility” on fraud, he said.

Even DNA has a role. Instead of taking a swipe of a criminal suspect’s cheek, DNA swabs from a crate of clothing can tell a company where a product was produced, and ensure a contractor isn’t surreptitiously having goods made in another factory by underage workers, van Drunen said.

Here are tips from Kivett and van Drunen on how to shore up supply chains to detect and prevent fraud:

Consider the risk. Ignoring problems of graft and theft at factories and subsidiaries isn’t an option, van Drunen said, calling the supply chain the “lifeblood of the company”.

Violations of child labour and corruption statutes could have major implications for a company, with exposure opening up the chance of a public relations disaster and other fines and sanctions. The US Foreign Corrupt Practices Act could also shut down a production line, for example, if a company’s representatives are linked to bribery or other corrupt practices.

That’s why companies, especially those that produce and manufacture goods in areas of the world with looser regulations, need to be diligent in looking for problems.

Go in-house or contract? It all depends on your company’s strengths and weaknesses. Companies with robust technology departments and data scientists may do well to develop their own data analytics and internal controls.

But companies that don’t have a team of analytic experts may want to bring in outside help to examine their processes and make recommendations of how to best monitor their supply chain, Kivett said. “You really don’t understand where your susceptibilities are,” he said.

Experts in fraud can help point out weaknesses and conduct an audit of current controls by using experience of dealing with widespread cases of fraud.

Realise that no one is immune from fraud. Supply chain fraud risk is fairly consistent, regardless of industry and size of the global operation, Kivett said, citing Deloitte research. It’s an indication that companies need strong internal controls and need to periodically examine the controls and see if additional measures are need.

“For organisations, there’s the tendency to say that wouldn’t happen here; we’ve got good people,” Kivett said.

That’s a foolhardy stance to take in today’s environment, especially in a business that depends on cross-border trade and can be infiltrated through cyberattacks or by other means.

WHERE TO LOOK

There’s not a single technology or area to look for fraud, Kivett said. Instead, companies should examine all their processes and try to move as much information to electronic sources as possible.

With data such as timekeeping systems, financial reports, and other information available in a searchable way, companies can periodically examine operations to look for suspect patterns.

“You can see what the baseline patterns are like and identify any anomalies,” he said.

Similarly, companies should make sure new acquisitions and new vendors are using the same internal control and data analysis procedures.