The term “garbage in, garbage out” is as relevant to your data today as it was 20 years ago. If your company’s data is in chaos – incomplete, fragmented, trapped in siloes, hidden in legacy systems, poorly identifiable, duplicated – anything you try to do with that data will be, unfortunately, garbage.
Four Ways Your Data Chaos Will Hurt You – Badly
The pandemic shutdown exposed or exacerbated problems with cost optimization, cost efficiency, business continuity and disaster recovery. As companies reopen and retool post-lockdown, it’s more important than ever that you have clean and complete data – everything starts with having a handle on your data. If it’s in chaos, it will hurt you.
1) The drip, drip, drip of legacy systems
It’s kind of a “Catch-22” scenario – your data situation is in chaos because of your legacy systems, but you don’t think you can move from the legacy systems because of how deeply ingrained your data is. But legacy systems cost more to maintain, things take longer, and opportunities are missed. Data is growing exponentially at a time when you need speed and efficiency – but that new data is going into chaos. Whether moving to the cloud or staying on-prem, it’s time for application modernization and integration. You don’t need to do it all at once – but you better start soon, or the drips will keep on coming.
2) The data privacy movement
The EU’s General Data Protection Regulation (GDPR) and California’s Consumer Privacy Act (CCPA) and laws like it give individuals a level of control over how companies use their collected data. This includes the “right to be forgotten” – a company must delete a requestor’s information in its systems as mandated. Fines can add up – in CCPA it’s $2,500 per compliance violation, going up to $7,500 if the violation is deemed intentional. If you don’t know what data matches to which customers, or all the nooks and crannies where that data is and how to get to it, the penalties could add up fast.
3) Entity resolution
“Wizard of Data” Jeff Jonas defines entity resolution as “who is who and who is related to who” in data. The common terms identity resolution; record linking, matching and deduplication; and “merge-purge” are forms of ER. Entity resolution means you have a single version of something that your business must deal with, like a customer or a vendor, or you understand the relationship between somethings, like which salesman belongs with which vendor. If your data is garbage, and entity resolution is difficult, you’re setting yourself up for problems from bad customer experiences and cost inefficiencies to outright fraud.
4) Advanced and predictive analytics
Companies today are using artificial intelligence and machine learning (AI/ML) tools to automatically detect sophisticated data patterns to predict potential outcomes, giving them better decision-making insight and abilities. For example, a financial organization using AI/ML can analyze its entire historical loan portfolio on an ongoing basis, learning and recognizing patterns of factors in real time that lead to a much higher rate of default. If a new loan application has those factors, the AI/ML program would flag it for denial. AI/ML is for “what’s next” questions, like which current customers will likely buy your new product or when will a piece of equipment break down. These insights can have you crushing your competition – unless your data is in chaos. Then the competition will be crushing you.
Get a handle on your data
Coffee with Talend
Join data experts Brian Kordelski (Prolifics) and Rolf Heimes (Talend) for a discussion on how companies are striving to be leaner and more efficient when it comes to housing and managing data. They’ll also look at how migrating to an opensource framework can be cost-effective and quickly deployed. More information and registration here.