How Reliable is Big Data in a Constantly Changing, Unpredictable World?

Data’s Backward-Looking Lens – Usefulness Versus Reliance

In 2011, ailing US retailer JCPenney recruited Ron Johnson as CEO, the previous president of Apple’s retail operations, who is credited with pioneering the idea of the Apple Store.

Johnson arrived at JCPenney intent on reinventing the model and boosting gross sales. He applied a broad refurbishment program, creating a steadier pricing system by eradicating coupons and clearance gadgets and presenting the shops as modern vacation spot boutiques inside malls. Johnson lasted lower than two years on the firm as JCPenney’s gross sales collapsed: same-store gross sales decreased 25% – a discount in gross sales of $4.3 billion – and the group ended with near $1 billion in internet annual losses.

Two of Johnson’s missteps for JCPenney had been noteworthy:

• Assumptions made – Many: Johnson assumed that what labored for Apple Stores can be a profitable recipe to comply with for giant malls, even these with worth delicate prospects who sometimes understand worth via promotions (reductions and coupons).
• Data used – None: Instead of testing the concepts with sure shops, gathering information and insights, prototyping and iterating, the brand new CEO assumed that the total overhaul of all malls would work. Johnson did not validate his assumptions earlier than executing the intensive and costly retailer revamps. Likewise, Johnson assumed the adjustments in pricing construction would obtain vital development in gross sales and profitability with out testing these concepts as a part of the decision-making course of.

The technique was fairly daring (or harmful): a easy strategy of specializing in a fastened, predetermined endpoint. The decision-making was equally binary: full reliance on untested assumptions with no seek for extra info, information, or insights to tell this linear course of.

Apple’s tradition and technique is one which usually doesn’t take a look at previous to launches. JCPenney has a wholly completely different proposition and buyer base. The new CEO was supported by activist investor Bill Ackman, and their concepts had been modern – they believed that reinventing JCPenney in this manner can be compelling. Of course, making assumptions is a regular a part of technique. The shortcoming right here was relying fully on a dangerous and costly strategic plan which assumed a singular attainable final result. The plan of ending markdowns and turning shops into locations may have benefited from groundwork and testing, with insights evaluated from pattern information to tell the choice.

Although information has limitations, these don’t equate ignoring it outright. Data will be highly effective when used to check assumptions. Analysis can rework uncooked information into insights to tell emergent decision-making. These insights and suggestions loops provide a clue to a multitude of attainable futures, however information’s usefulness doesn’t imply we must always rely solely on it.

We crystallize beneath six key takeaways on information in our liminal and unpredictable world which provides us a palette of shades between “reliance,” “limitations,” and “usefulness”:

Facts versus assumptions

A key profit of knowledge lies in its capability to offer empirical proof to substantiate subjective opinions and assumptions. Testing tacit and express assumptions can present validation, and whereas details are higher than assumptions, they nonetheless solely present data of the current state (versus the longer term). Validating assumptions concerning the previous or current is a steady loop. In an updating world, assumptions must be reevaluated and frequently examined.

Big information

Massive quantities of networked datasets can present ever-deeper insights via sample recognition at scale. Machine studying permits us to find non-intuitive dynamic connections, whereas pure language processing is efficient for unstructured extraction. We can acquire new learnings and insights from the info now we have as we speak, and might use it to tell decision-making and actions. Business technique is more and more reliant on large information, which is additionally used to coach and enhance AI purposes. A key distinction between outright “information” and “large information” is also known as the three Vs. Big information is characterised by quantity (giant dimension), velocity (rising quick), and selection (numerous sources, together with social media, databases, and purposes, each bodily and digital). Whether we name it information or large information, if we search to use it to the longer term, the issues stay the identical. Data doesn’t predict something past the modeled assumptions of a system with stabilized parameters. Such predictive analytics will be invaluable in controllable, particular domains, the place machine studying and sample recognition will be utilized to almost infinite simulations. However, advanced environments are dominated by unknown variables, and are usually unpredictable. Correlation can solely be established retrospectively with information evaluation in these environments, and causality will be tough to deduce.

“Relevance-driven” beats “data-driven”

Relevance is decided when assumptions and information confront the true world. To keep related, you can’t lose sight of understanding buyer conduct. Clayton Christensen’s Jobs To Be Done (JTBD) reminds us that the explanation individuals purchase and use any services or products is to get a particular job performed. It is no coincidence that Amazon’s first management precept is Customer Obsession (“Leaders begin with the shopper and work backwards… Although leaders take note of rivals, they obsess over prospects“). Given the extent of competitors, best benefit is fleeting. Yahoo had the first-mover benefit, however Google turned the dominant search engine. Google’s obsession with prospects is to remain related. This drives their mantra to concentrate on the consumer by creating new, shocking, and radically higher merchandise.

True innovation, similar to the longer term, is not measurable at its inception

However beneficial information will be to tell decision-making, the problem lies in measuring the unmeasurable. Breakthrough innovation is a really novel act of discontinuous creation (not merely an enchancment to an already-extant object). New and shocking have a tendency to not be conducive to ex ante information.

The worth of knowledge is to tell related decision-making, to not be prescriptive

By testing what will be examined and measuring what will be measured, dynamic intelligence is generated over time. Analysis can reveal insights for a particular vary of quantifiable information units. Machine studying provides real-time suggestions loops, creating an evolutionary course of in which the outputs are reused as future inputs. This helps smarter choices with up to date interpretations of the outcomes from our each day experiments. These insights of the previous and the emergent current can inform decision-making as we speak and tomorrow, regardless of being anchored in the previous.

Counterintuitively, limiting reliance on information releases its superpowers

While insights derived from information will be highly effective, understanding the restrictions of knowledge releases its true superpowers. When you combine that at any level in time, information on the longer term is nonexistent, you retain an open thoughts to the limitless prospects. You can use suggestions loops in decision-making to assist anticipate shifts and alter, however not grow to be a prisoner to what pattern information appears to be signaling. Effective decision-making can prevail regardless of velocity and uncertainty when there is room for experimentation as an emergent course of in relation to open and unwritten futures. The worth of knowledge is to tell evolutionary decision-making, not imprison. This dynamic course of continues by actioning choices whereas benefiting from experiential suggestions and adjusting future choices primarily based on earlier outcomes. This multilayered strategy acknowledges the assorted attainable futures forward. Quantifiable and unquantifiable; goal and subjective; measurable and unmeasurable drivers of change all contribute to imagining the colourful kaleidoscope of attainable eventualities and assist construct the capability to be future-savvy.

In our UN-VICE world (UNknown, Volatile, Intersecting, Complex, Exponential), count on discontinuity, instability, shocks, and randomness. These dynamic unpredictable programs are tough to mannequin as situations always change and new elements emerge. Irreducible complexity is not conducive to being prescriptive, however we are able to nonetheless form the longer term with out a dataset on the longer term – we merely want creativeness.

The insights derived from information will be invaluable as a suggestions loop to decision-making, however ought to by no means be confused with being a proxy for the longer term, a predictor of the longer term, nor the longer term itself.

Note:

Roger Spitz is the lead creator of The Definitive Guide to Thriving on Disruption (Disruptive Futures Institute, 2022), from which this text is tailored.

The submit How Reliable is Big Data in a Constantly Changing, Unpredictable World? appeared first on Datafloq.