More defined questions are arguably better suited to how analytics are positioned in this process. For example, it’s true that predictive analytics will be correct more often than humans in determining the last day of usefulness of a machine part. It’s also true that analytics will be accurate more often than people — even experts — in detecting patterns in massive amounts of data.
But it’s also important to understand the limitations of this approach. Machines can do only so much. They don’t think or learn as you and I do. For example, machine learning can sort data faster than humans in any this-is-a-cat-this-is-not choice of well-defined options. But even then, machines are likely to mislabel a few cat pictures that a human child would instantly recognize to be distinctly feline.
By comparison, decision intelligence melds your mind to the machine’s “mind” so that the strengths of each one overcome the other’s weakness. Obviously, it’s not an actual melding of human and machine, but rather a blend of decision capabilities.
That’s a vital advance because it renders outputs of greater significance to problem solving and additional analysis of real world impacts. An unintended side effect of using the traditional model was the devaluation of the worth of human input outside of programming and other software development activities. We humans got into the habit of putting total faith in algorithm outputs — even when the outputs are in conflict with our thoughts and experiences. We do this despite it being well known that data is rarely perfect or complete. But that’s not the only problem challenging our blind faith in data analysis.
Unfortunately, as Cassie Kozyrkov, Google’s head of decision intelligence, often reminds us: “Strategies based on pure mathematical rationality are relatively naïve and tend to underperform.”
Anyone who has been chased by ads — for items they have already purchased and don’t care to buy again — across social media and the Internet can attest to this most annoying “relatively naïve” underperformance of some analytics.
Even so, general wisdom still has it that data — and not people — should take the lead in business decisions. It’s machine over gut instinct every time, goes the mantra, even when that isn’t done in practice. But why must this be an either-or question? The correct answer, of course, is that it doesn’t.
Enter decision intelligence, the practitioners of which “usually emphasize details of broad business decision systems; these include analytics, management of data and information resources, business rules, integration of decisions into operational systems and other functions,” according to Meta Brown, the author of Data Mining For Dummies.
The aha moment here is that analytics and data mining are parts of the decision-making process rather than the whole of it.
Decision intelligence is a huge umbrella under which all the activities necessary to produce decisions huddle and are put into practice to yield a preset, desired outcome.
Predictable and even preset questions are still good questions in many business pursuits. There’s nothing wrong with continuing to interrogate the data in this way for many common use cases. After all, you always want to know your sales for the day and how that number compares to last year’s sales. And, you likely always want to know which products are hot sellers this week, which employees were the most productive, and so forth. So go ahead and keep asking the data these questions.
But it’s time to see what else you can do with data and analytics and AI. It’s time to rethink how to go about the business of deriving decisions at scale as well. And that is exactly how some in the data sciences arrived at decision intelligence.
Reinventing Actionable Outcomes
Perhaps no buzzword is more touted in the data analytics industry than actionable outcome. To be fair, some outputs are actionable and some of those actionable items can even be fully automated — no humans needed outside of those who built the machines that are now doing all that work.
However, more often than not, actionable outcomes are insights that might enable an action. That’s quite a different concept than analytics that can deliver actual decisions, or a rated range of them, complete with expected impacts.
Decision intelligence aims to change outputs from insights to decisions, at any scale and by using varying blends of human and machine tactics. This is what Google’s Cassie Kozyrkov means when she so often describes the difference between traditional data science with machine learning, and decision intelligence as “the difference between those who make microwave ovens and the cooks who use them.” It’s the recipe and the outcome that matter, she says, because the chef has no need to build a microwave or even understand how it works.
The focus is shifting, in other words, from data explorations and building more technology to delivering a specific payload.
Living with the fact that we have answers and still don’t know what to do
Decision modeling is maturing to include more pointed pursuits, broader considerations in the decision-making processes, and more accountability for the results. It won’t surprise you to hear that this new phase is also labeled decision intelligence and that it’s measured by the value of its outputs, whether that’s in terms of impact or return on investment or both.
There is no more time, patience, or money for fishing in data lakes or panning for gold in data streams in the hope of discovering valuable knowledge. Decision intelligence insists on moving with purpose to achieve a predetermined end whose significance has been well defined.
Here’s a handy example: You may ask the data what the weather will be like tomorrow. But that isn’t the question. Nor will the answer “Partly cloudy with a high of 70 degrees” be of any significant use to you.
Think hard. What is it that you really want to know?
Perhaps it’s whether to plan a picnic tomorrow. In that case, you likely need an assessment of the weather, plus pollen counts, projected traffic at the park, and maybe even water sports availabilities and/or wait times for picking up prepacked picnic lunches at your favorite deli.
Perhaps you wanted the analytics to tell you that your best pick for a picnic tomorrow is “Happy Park on the north beachside with shelter from the wind but not the warmth of the sun, and plenty of tables, because it’s not a high traffic park. Also, your route has three delis, and two have less than a 10-minute wait for order pickups.”
Decision intelligence can be applied for a relatively-speaking best decision for a problem or question of any size, ranging from the highly personalized (like the picnic questions) to the truly huge (like a global pandemic).
I talk earlier in this chapter about how the COVID-19 epidemic revealed the limits of a data driven approach to problem solving. Some of the lapses in the initial response to the epidemic were certainly caused by the urgency of the threat and the novelty of both the virus and the vaccines. Yet several factors worked in favor of making sound public health decisions under pressure. For one, Israel struck a deal with Pfizer to share patient data on the efficacy and side effects of the Pfizer vaccine in real world use. Israel also has one of the world’s most efficient healthcare systems, complete with highly developed electronic healthcare records (EHRs) capable of collecting massive patient data in real time. The resulting database is well organized and filled with clean data — accurate