Picture the desk of a manufacturing or supply chain manager on any given morning: a stack of disconnected Excel files, raw reports pulled from the ERP system, and printouts from the warehouse management tool. Each source tells a different story, and none of them talk to each other. At that moment, ‘big data’ stops being an abstract technology discussion and becomes a very tangible operational problem. The question is not whether the data exists — it does, in abundance — but where to begin making sense of it.
Big data is not solely a volume problem. The definition encompasses variety and velocity as well: data produced in different formats across disconnected systems, failing to reach decision-makers in time to matter. In a mid-sized Turkish manufacturing or logistics company, all three dimensions create simultaneous friction. The ERP system holds stock and order data; accounting runs on a separate platform; the sales team maintains its own tracking spreadsheets; warehouse movements sometimes stay on paper. In this environment, the chronic condition is ‘data rich, information poor,’ and managers end up relying on intuition precisely when structured analysis would serve them better.
Before constructing the 90-day framework, one prerequisite needs to be made explicit: a big data initiative is not an IT project, it is a business decision problem. Technical infrastructure comes second; the first task is defining which question needs answering. For this reason, the first thirty days should be dedicated entirely to data inventory and problem definition. Which systems produce which data? How much of that data is regularly updated, and how much goes stale? At which decision points do managers find themselves thinking ‘if only I had known that’? The answers to these questions should not feed into a spreadsheet but into a priority matrix that ranks problems by business impact and data availability.
The second thirty days are allocated to selecting and structuring the pilot use case. A well-chosen pilot has two characteristics: it must create a measurable business impact, and it must be workable with data that already exists. In supply chain management, excess inventory and shortage analysis typically meet both criteria. Questions such as which SKUs are tying up unnecessary working capital, or how variance in supplier lead times translates into production stoppages, are both concrete and answerable using existing ERP and purchasing data. A business intelligence tool or a data warehouse solution can be brought in at this stage, but over-investment must be avoided. No licensing or consulting expenditure should be committed without a total cost of ownership (TCO) calculation that includes implementation, training, and ongoing maintenance.
The third thirty days focus on generating the first insights and building organizational learning. Reports and analyses produced within the pilot scope are presented to decision-makers, but the primary objective at this stage is not to make the right decision — it is to learn how to ask the right question. In a team where data literacy is still developing, presenting ‘insights’ rarely produces the expected response. Forming a small working group to discuss and interpret findings is therefore essential. The manager’s role here is not that of a technical project owner but of a facilitator who helps the organization build meaning from data.
The most common obstacle encountered during the 90-day process is data quality. When data entered into systems turns out to be inconsistent, incomplete, or simply wrong, the project loses momentum. The way to handle this is not to treat data cleansing as a separate project that must be completed first, but to accept it as part of the pilot analysis itself. Which fields are reliable, and which require manual correction, should be documented — and that documentation should serve as an input for future system investment decisions. Poor data quality is not a technical obstacle; it is a reflection of process discipline gaps in day-to-day operations, and communicating this clearly to senior management is part of the manager’s responsibility.
At the end of 90 days, three concrete outputs should be in hand: a documented data inventory, at least one insight from the pilot project with a measured business impact, and a realistic assessment of the organization’s data maturity. These three outputs form the foundation of the ROI analysis that will determine whether to scale the big data investment further or recalibrate the approach. Understanding whether the organization is genuinely ready to use a technology before purchasing it is the most critical managerial judgment in this entire process — and it is one that no vendor briefing will ever provide for you.
This article was originally written in Turkish by Gökhan MERCANOĞLU on February 28, 2011 and has been automatically translated into English and other languages using machine translation.