A retail chain’s IT director walks into the general manager’s office with a budget request for a big data infrastructure. The presentation is polished: references to competitors’ data warehouse projects, excerpts from consulting reports, bold claims about future competitive advantage. The general manager listens, takes notes, then asks a single question: ‘Where are the numbers specific to us?’ The IT director has no concrete answer. The meeting is postponed. This scene captures the most persistent obstacle facing big data investments: the gap between technical enthusiasm and financial justification.
Big data refers to the point at which data volume and variety exceed the processing capacity of conventional database tools. Mid-sized and large enterprises in Turkey are generating growing volumes of data from customer transaction logs, supply chain records, and call centre histories. Raw data carries no inherent value; it creates value only when analysis improves the quality of decisions. The real challenge is translating that potential into a language the board of directors can evaluate.
A sound business case rests on three distinct pillars: revenue growth, cost reduction, and risk mitigation. On the revenue side, customer segmentation and cross-sell analysis can be positioned as a measurable sales efficiency target. On the cost side, reducing excess inventory or optimising logistics routes produces concrete savings line items. On the risk side, framing customer churn prediction as an early-warning system allows the projected loss per churned customer to be directly linked to the investment. Modelling each pillar separately is what transforms a business case from a ‘visionary narrative’ into a ‘financial proposal.’
Avoiding the trap of unmeasurable benefit promises is critical throughout this process. Phrases like ‘we will make better decisions’ or ‘we will build a data-driven culture’ generate scepticism rather than confidence in the boardroom. Instead, lay out a clear total cost of ownership (TCO) analysis: licensing or infrastructure costs, internal resource requirements, training burden, and ongoing maintenance. Against this, place the expected returns from the three pillars defined above. Build the ROI calculation in two scenarios: conservative and optimistic. The conservative scenario earns you credibility; the optimistic scenario shows where the project can go.
A pilot-based proof approach breaks a large budget request into small but measurable steps, which significantly simplifies the approval process. Rather than proposing an organisation-wide data warehouse, suggest a three-to-six-month pilot focused on a single business unit or a single business question. A narrow initiative that analyses only customer return data with the goal of reducing the return rate on a specific product line, for example, both tests the technical team’s capacity and delivers a concrete proof point to the board. The real numbers produced at the end of the pilot form the strongest possible justification for the next phase.
The most common practical difficulty encountered in this process is that existing data quality turns out to be well below expectations. Before a big data project begins, a data audit is needed to assess how much of the data to be processed is clean, consistent, and reliable. In many Turkish businesses, different departments record the same customer under different codes, and sales and accounting systems maintain records that do not align with each other. If this problem is ignored in the business case, the data-cleansing costs that surface during the pilot will blow both the budget and the timeline, damaging the project’s credibility. Explicitly including data quality improvement costs and timelines in the business case prevents the disappointment that otherwise arrives later.
As a manager considering a big data investment approval, apply this criterion: does your business case clearly show what will be lost if the pilot fails? Proposals that present only the success scenario create unease among decision-makers. Including the risk and the fallback plan in the case demonstrates the maturity of the proposal. Keep the pilot scope narrow, define measurement criteria upfront, and re-price the next phase using the real data you have at the six-month mark. A big data investment business case is not a one-time presentation document; it is a living financial argument, updated at each stage and fed by evidence.
This article was originally written in Turkish by Gökhan MERCANOĞLU on June 27, 2011 and has been automatically translated into English and other languages using machine translation.