With equal parts humor and frustration, someone has likely asked you quantify the unquantifiable and you’ve offered the reply, “How long is a piece of string?” Searching and solving unknowns is never easy for IT and big data initiatives are no different. CIOs will have to work through the ambiguity of the term "big data" and identify specific tools, architectures, and technical pain points the business faces. Unfortunately, when all that IT owns is a [ERP] hammer, all problems look like nails.
Myth - One-Size-Fits-All
Considering the urgency upon monetizing data, the unique work of data scientists and the requirements for flexible systems of insight, CIOs mistakenly believe consolidation of OLTP and OLAP fit into ERP architectures is a path to big data. IT justified ERP with enterprise agreements (EA) and one deployment. That one-size-fits-all ERP and in-memory database approach was intended to reduce IT license fees, vendors, and consolidate data management. My previous blog substantiates that ERP is not the one-size-fits-all for data.
Implementing one-size-fits-all, monolithic ERP architectures and business intelligence applications, was intended for IT savings and consolidation mandates but overlooks agility, simplicity, and data independence expected by big data users. In the distant future, we’ll see ‘data war’ studies showing every $4-$5 saved through ERP consolidation required CMOs spend $2-$3 to by-pass IT processes and compensate for lost functionality.
Marketing and CMOs have become the driver of big data projects, and according to advisory group CEB, big data now accounts for 37% of their tech spend. In the Accenture study titled the CMO-CIO disconnect 1,100 senior marketing and IT executives found, “... 40% of CMOs believe their company’s IT team doesn’t understand the urgency of integrating new data sources into campaigns....” Even gaining access to virtualized capacity in many organizations can be a multi-week process followed by days or weeks of software installation, configuration, and tuning. One might argue these are limits of IT, others might see it as limits of operating within ERP.
ERP vendors might as well be asking Georges Seurat to paint only ‘within their lines.’ We find big data solutions require systems and user freedoms to operate ‘outside the lines’ of ERP using purpose-built appliances, software, and in-memory storage. IBM Business Intelligence Pattern with BLU Acceleration is built upon IBM PureApplication System hardware designed and tuned exclusively for big data tasks.