How might DQ affect Fraud

Question

One area of interest to them is customers or use cases revolving around FRAUD. I could’nt quite pinpoint the best way to showcase this (other than Source to Target DQ checks)

Is there a good use case or customer working with CDQ in a Fraud Capacity? Or if there isnt one = what’s the best way to make that link// bridge that gap (if we can)

ericgerstner · Answer

I would say that Fraud is just like any other Root Cause for anomalies in Data. Something is amiss, we presume it’s a bad ETL or inadvertent code.

What happens when it’s a mal actor?

What makes CDQ better than your typical rules engine?

Consider the advantages of CDQ amidst typical institutional internal controls for fraud.

Separation of Duties, do you have the sufficient scale to separate your DQ rule-writers from Data Stewards from Data Entry? Wouldn’t it be easier to know that with automated anomaly detection, the rule-writing is agnostic to fraud.
Controlled Information, do you have the access model setup to prevent those who need access to the data from those who manipulate? With CDQ, those manipulations would be caught and shared in our controlled workflows, giving data to those who need it safely.
Real-time monitoring, can you easily scale the timeframe of monitoring to keep up with variances in markets or operations? With CDQ, your scanning interval is changed with a few clicks and the reporting continues seamlessly.

Perhaps a use case to share:

When “Behaviors” notices data drift as rows disappear from our cookie marketing data, our 3rd party vendor may not be sharing everything.
When “Outliers” shows lower BW executions in a regional report for internet connectivity, we might have someone misappropriating BW.

When “Patterns” notices that IP Address X is logging in from location in another continent, we might have a fraudulent login.

When Source does not match the target DB, we could detect malicious changes in previous data.

When Schema notes a brand new column that appears to be joining all PII, we could be seeing a SQL analyst trying to collect organization information.

When Dupes shows rows with fuzzy matches on emails,we could be seeing a robot process meant to overwhelm our login service.

When Shapes see a unique email format, we could be seeing mal actors leveraging a specific domain enter our ISP.

1	Grant Rollerson SUNCORP CORPORATE SERVICES PTY LTD
2	Tom Friesen Algonquin College
3	SB S Benson Plus3 IT
4	S stefan_busch
5	Brian Lloyd USA TODAY Co.
6	D daverichard