One of the most important aspects of creating a big data analytics strategy is empowering leaders, according to Jean-Michel Franco, director of product marketing for data governance at Talend.
Franco was addressing delegates at Computing's Big Data & Analytics Summit 2015, where he said data quality and not quantity was key to maximising insight using big data analytics.
But he said a huge part of this was empowering stakeholders within the organisation.
He said a typical data-driven organisation would have a data architect, a line of business user, an "integration cloud citizen", an app developer and a data curator - these would be the "go to" people in the organisation, he said.
"It's about picking people in the organisation ... you make them officially accountable [for the overarching big data strategy] and bring them the right tools in order to bring value to the organisation," he said.
Franco said ensuring data quality has to be at the forefront of any data strategy, adding that providing easy access to data using tools such as Tableau and QlikView has questionable worth if the data being viewed is unreliable.
"There is a problem with the data in the system," he said.
Talend estimates that about 25 per cent of customer contact data is inaccurate. Many organisations fix this with a data quality project, but Franco said that a year later, many organisations have the same problem again.
Another issue, said Franco, is that organisations struggle to get a complete view of their customers.
"If you ask IT if they have it they may say yes but if you ask a line of business they say they don't have the data," Franco claimed.
Not having ownership of data creates issues with compliance, said Franco, who alluded to the attacks on Sony and suggested that a lot of data that had been compromised was on Excel spreadsheets that weren't controlled.
Businesses experiencing problems with the quality and depth of their data and with compliance should take a more collaborative approach to managing their information, he said.
"It's not just the CIO, it is people in marketing, billing, it is a collaborative effort to manage the ongoing data quality," he suggested.
Finally, Franco advised firms to take on board his "three pillars of information refinery".
The first is a data lake - to ingest the data as it comes, and then classify and prioritise it based on relevance. The second is the data lab - discover the data, create and test the models like a clinical trial phase. And the third is to turn information into a trusted and shared asset.
"This is where you turn information into a trusted asset with a lot of governance behind it. It means working with someone from the ‘data lab' to turn it into insight, so you're not just sharing raw data but taking meaningful data out of it," he said.
---
Autor(en)/Author(s): Sooraj Shah
Quelle/Source: computing, 26.03.2015
Bitte besuchen Sie/Please visit:
