Procurement Frauds Analytics

Procurement Frauds5Almost 59% of the companies are affected by Procurement Frauds, it is showing an upward trend and impacts close to 15% of profits. It is possible to predict and prevent Procurement Frauds with following Analytic approach:

Understand Procurement Business rules and strengthen processes:

  • Processes of vendor qualification
  • Approved vendor list
  • Role and responsibilities of procurement managers
  • Procurement norms,
  • Three way matching process

Typical cases of Frauds are due to:

  • Phantom Vendors, Shell Companies
  • Bid Rigging
  • Supplying sub-standard products
  • False Invoices
  • Unjustified Single Vendor Orders

Big data analytics has now made it possible to detect frauds at early stage and take corrective actions rather than use the data for postmortem analysis. Some of the techniques used to reduce Frauds in Procurement are:

  • Aggregate data from all sources and bring out suspicious pattern like relationship links leading to collusion, irregular transactions and unapproved vendors
  • Analyzing keywords through email unstructured data
  • Establishing relationship between individuals and entities based on common links such as bank account numbers, telephone numbers, addresses, company directors and social media data on relations, profession
  • Look for sudden drastic shift from past pattern to flag it off and highlight it with increased fraud risk score
  • Compare Item cost with past average procurement cost and flag it off in case of abnormal increase

Using Advanced analytics build models to associate activities which are most likely lead to a particular type of fraud. At the same time multidimensional analysis helps to reduce false positives – when link analysis may show relationship but anomaly detection shows transactions to be normal, system wouldn’t flag it off. Similarly algorithms are written to reduce false negatives to reduce the unwanted red flags.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>