In the world of big data, some things are easy to objectively measure, such as how much money a person spends, or how much time it takes to accomplish something. Other things are notoriously difficult to quantify, such as moods, subjective opinions, and beliefs.
As our technology enables us to make and store more measurements, and as demand for big data analysis continues to grow, we’re going to have the option to quantify these unquantifiable metrics. There are some major advantages to this approach, but are they worth the potential costs?
The Plus Side of Quantification
These are some of the biggest advantages this kind of data quantification can offer:
- Tools for decision making. When you’re making a major decision on behalf of a company, staking thousands to millions of dollars on your conclusions, you can’t cite your instincts or beliefs as hard evidence in favor of your chosen position. It’s better to have something evidence-backed and provable on your side. For example, you can calculate a priority score for projects in your portfolio management (PPM) strategies, or even rely on average user ratings to gauge satisfaction.
- Avoiding blind speculation. Numerical values also hold us to some degree of objectivity; without a figure to point to, we allow our speculations and assumptions (and therefore, our biases) to run rampant. Cognitive biases are powerful, but they hold much less power over real numbers than they do over unquantifiable impressions and beliefs.
- Inching toward a true value. Our measurement tools may not be perfect, but they’re the best tools we have for the moment. Right now, we may not be able to express a feeling with a number, but we can still get closer than we ever have before. Gradually, as new tools and techniques emerge, we may be able to inch closer and closer to a true objective figure, even if it’s not a comprehensively straightforward numerical one.
The Downsides
So what are the potential downsides?
- Method of measurement. No matter what method you’re using, there’s been some human influence involved in its creation. Humans came up with the questions. Humans came up with the scales. Humans invented the tools. That means all the tools and resources we have are naturally biased, at least in some ways. Even though the final results will appear to be objective and numerical, they’re still going to be based in systems we created for ourselves, and that makes them impossible to be totally unbiased.
- Overconfidence. When you start seeing everything in terms of numbers, and comparing those numbers against each other, it’s easy to become overconfident. Your decisions can all be plugged into basic formulas, and you can find an excuse to favor any decision, so long as the numbers are in your favor. That overconfidence can lead to questionable decisions, and ones that don’t fully examine all the variables in play.
- Reducing the irreducible. Quantifying abstract or subjective experiences is an attempt to reduce what is inherently irreducible. If you ask a question about how much time your employees waste per day, you’ll have to define parameters that are somewhat undefinable; for example, what qualifies as time ? Does it include personal time or break time? How do you define waste? Is it any passing second that isn’t used optimally, or does it only refer to chunks of time? Does inefficiency qualify as time waste, or must the time be spent doing nothing? What about time the employee is required to spend, but doesn’t improve your company’s bottom line? No single numerical measurement can answer all these questions.
- Ignoring subjective qualities. Finally, quantifying data points forces you to ignore the subjective qualities you’re looking for. Let’s say someone rates your app two out of five stars, but there’s one feature they absolutely loved. That piece of information, and the enthusiasm that went along with it, is completely lost in what is ”overall ”a negative score. On some level, you need subjective information just as much as objective information to fully understand something.
It’s perfectly acceptable, and even advantageous to quantify points of data previously thought to be unquantifiable. However, those quantified data points shouldn’t be our only basis for understanding a broad concept, and shouldn’t be the crux of our decisions. Instead, we need to use quantification as one of many tools we use to understand our environments, and pair it with subjective understanding and reasoning to form the best possible conclusions.
