CIOs should learn how to use big data from games developers whose success depends on it, claims an ex-Gartner analyst.
Former Gartner business intelligence (BI) veteran James Richardson told CBR that CIOs are still struggling to make use of "imperfect" big data, and said they should dedicate entire teams to analysing it, rather than "a few boffins in the corner".
Richardson has just joined business discovery firm Qlik, which specialises in presenting big data trends in a user-friendly way, whose software is utilised by King.com, the developer behind hit game Candy Crush that is set to IPO for $500m this month.
The new director of product marketing said King.com’s success is based on how many accurate player trends it can identify from the 22m rows of player data they choose to analyse from 2bn rows stored every day in its Hadoop cluster.
"It’s democratised big data," claimed Richardson. "It’s not just a few boffins in the corner. They use the data to look at where people find the game difficult, where players are leaving them.
"But instead of big data scientists it’s game designers, marketers looking at campaigns to generate more activity, it’s financial analysts."
And he contended that CIOs could adopt a similar approach to make the best of vast swathes of information, much of which he admitted could be inaccurate due to its unstructured nature.
"It’s very true that CIOs are struggling with it – it’s only useful if you’ve got a team that’s going to learn through play," he said. "A lot of the gaming firms are very non-traditional in how they’re set up, and CIOs can learn from that."
Last month Oracle’s UK and Ireland head of business analytics, Paul O’Riordan, warned that companies that were taking big data analytics away from IT were in danger of relying on inaccurate information.
But Richardson said that with the vast majority of big data coming from unstructured sources like social media, CIOs must accept a degree of tolerance is necessary.
"Organisations need to understand that their data isn’t perfect, and say ‘we’re willing to work with the ambiguity around that imperfection.’"
Richardson’s opinions appear to contrast a recent prediction of his former employers that enterprises will employ a chief data officer (CDO) as the demand to utlise big data grows.
The analyst firm said there were 100 CDOs serving in large enterprises in 2013 – more than double the number in 2012.
It also predicted that by 2014, more than 10% of government organisations will have appointed a CDO.
