Decisions, decisions … will ‘Big Data’ have ‘Big’ impact?
Corey Yulinsky, The Financial Times, February 24, 2012
A phone book with a billion pages would reach about 80 miles into the sky. ‘Big Data’ is about utilising computer databases containing that scale of information - and more. This new management buzzword is clearly here to stay.
Such the attention is well-deserved. The potential advantage to companies able to take trillions of bytes of information, mine relevant data, convert it into insights, and make it useful for customers and employees is enormous. Information abundance and technology advances have taken us from a long era of information scarcity and dropped us into the deep end of information overload. The real question facing companies is how to ensure that the big potential of Big Data is actually realised.
The answer is to recognise that the most powerful benefits of Big Data come from changing the core elements of an organisation’s operating model as much if not more so than its technology. The first step is to understand “what’s so big about Big Data?”
Data becomes Big Data with the confluence of four factors
1. Information ubiquity:The breadth, volume and timeliness of available data, structured and unstructured;
2. Speed: The ability to store, process, and retrieve massive quantities in dramatically reduced timeframes, often near real-time;
3. Machine intelligence: Processing power driving algorithms that access vast data sources and drive decisions – amplified by feedback loops enabling continuous learning and improvement;
4.Economics: The cost of all of this data collection, storage, processing and delivery has dropped radically.
Big Data only has an impact when and organisation can capitalise on these attributes to change the decisions it makes or the way decisions are made. Don’t forget that the predecessor of Big Data was often called “decision support.” Otherwise, it is simply an interesting but low-value diversion of resources. Organisations that consciously and explicitly adapt decision-making processes to the opportunities afforded by Big Data analytics will reap the benefits. Others will experience the frustration and bottlenecks we saw with CRM (customer relationship management) and similar data-driven initiatives.
An approach to revolutionising decision-making starts with recognising four broad categories of decision-making and how they involve Big Data:
1. Episodic decisions: Decisions made infrequently, sometimes on a regular cycle and sometimes in response to triggering events. Often strategic in nature (which businesses/markets to enter/exit, which customer sets to pursue) these decisions are highly discretionary in nature, and require synthesis of multiple information sources. Increasingly, simulation models (combining descriptive and predictive approaches) are used to drive scenario analyses making tradeoffs more clear.
2.Fixed cycle decisions: All businesses have an ongoing cadence for key processes (demand forecasting, sales calling, marketing campaigns, pricing, inventory replenishment). Historically made as discretionary decisions backed by “rear view mirror” performance data, they are now often supported by predictive analytics leveraging Big Data flows.
3.Embedded algorithmic decisions: These “machine-made” continuous decisions driven by optimisation algorithms built-in to consumer (or frontline) facing interfaces are most often the face of Big Data decision making. They include dynamic online offers, real-time credit and pricing decisions, automated underwriting and call routing. These decisions driving “personalised” outcomes are the type most often changed by the data explosion.
4.“Controller” decisions: Evaluation, iteration and refinement of decision-making is usually focused on the effectiveness and ROI of algorithmic decisions. Machine learning ensures that the algorithms constantly get better, but there also has to be a rigorous layer of analysis to make sure these algorithms stay aligned with the company’s objectives and targets.
When viewed through this decision lens, it becomes clear that there are three major ways to create winning impact with Big Data.
1.The Battle of Business Rules – Who will create the most compelling and effective algorithms? Companies who have the capability to understand which data and which analytic insights will drive desired outcomes will create the foundation for cumulative learning and increasing advantage. Machines still need to be “taught” foundational business rules and the smartest teachers will win.
2.Decision Migration—Big Data’s analytics and processing aspects create the opportunity to transform episodic or fixed cycle decisions into algorithmic decisions, with potentially lucrative impacts for disrupters. In effect, this happened in segments of the advertising business where a large portion of traditional negotiated price (episodic/fixed cycle) ads were replaced by online ads driven by real-time auction pricing (algorithmic continuous). Another example can be found in the stock market with the advent of high frequency trading. Identifying which decisions can be migrated and how to do so will be another competitive arena.
3.Information Transformation – More rarely, we expect to see instances where new ways to play emerge in existing industries driven by companies that reset industry economics through by applying new analysis driven by Big Data capabilities. Featuring capabilities and cultures that are very different from their competitors, their strategic decisions are made differently and manifest themselves across the organisation. Two early examples are Harrah’s (now Caesars) in the gaming industry and Capital One in credit cards. In both cases, a CEO saw the transformative power of intensely data-driven decision processes to catalyse their businesses. The nature of this kind of transformation is complex relative to the other ways to win, but the pay-offs can be terabyte-big.
For Big Data to go from buzzword to providing value to your bottom line, the quickest route is to identify, manage and evolve the way your company uses data to make decisions and then focus on using Big Data to transform the key decisions that drive your business model.
Corey Yulinsky, a New York-based Partner in Booz $ Company’s Financial Services practice
