< All Blogs

Why My Wife is Right and Google is Wrong

Business decisions are a little like marital ones. I’m not talking about the big, strategic decisions we make like “Should we acquire a small competitor?” or “Should we have a child?” I’m talking about the small but important decisions we make every day. The ones that are nonetheless consequential and all consuming in the moment, with a lot on the line to get it right. Like when you’re cutting it close on getting somewhere urgently. “How do we get from downtown to uptown, on time to pick up the kids after school?” Same as “How do we get from downtown to uptown, on time to make our sales call?”

By
Gary Edwards, PhD
,
on
March 1, 2016

A Primer On Data Science For Everyday Life

Business decisions are a little like marital ones. I’m not talking about the big, strategic decisions we make like “Should we acquire a small competitor?” or “Should we have a child?” I’m talking about the small but important decisions we make every day. The ones that are nonetheless consequential and all consuming in the moment, with a lot on the line to get it right. Like when you’re cutting it close on getting somewhere urgently. “How do we get from downtown to uptown, on time to pick up the kids after school?” Same as “How do we get from downtown to uptown, on time to make our sales call?”

Of course, you know the way home to your kid’s school and you know how to get to your prospects building. You live in the city. So for all those who simply could care less and take the route they always take and let the chips fall where they may, the story ends here. You make your decisions based on road knowledge and you are happy that this by and large works for you despite some fails. But if you’re curious and really want to know how to make the right decision on what route to take to get to your desired destination on time, you will need more than just road knowledge — you will need to combine your domain knowledge with data science driven technology.

Consider the following:Perhaps you like your technology and your smart phone. You pull up Google Maps even in your own hometown because it provides up to date, real time traffic conditions. That’s when the argument ensues with your partner sitting right next to you. Who knows better, your partner who drives to pick up the kids all the time from different directions or your device that is loaded with the latest version of Google Maps, equipped with the best data insights? Do you rely on domain expertise or technology? A lot is on the line. If your partner and Google disagree, do you want to own the outcome here if you choose technology over your partner’s best advice? No matter what, we can’t be late this time!

Use Case” For People, Process and Technology 

Let’s begin looking for an answer by taking a step back. Get a new perspective by thinking of how businesses make decisions and try applying it to the urgent, personal matter at hand. How would an executive in a large, successful business enterprise make this decision? As the technology people have taught us to say, the situation at hand is a particular “use case” for the people, process and technology involved. It’s definitely a moment when executives have to find the right balance of relying on Domain Expertise versus Data Analytics technologies to guide their decision-making.

Somewhere in the back of your mind the cheesy line “You are the CEO of you” comes back and you realize it’s your call.

So how do we decide?

What happens “in the moment” when we need to make a decision? We could be guided, as we are in business, with a Scorecard. Let’s imagine the visualization of its content. How many times have you been late for the kids in the past 6 months? How many complaints has that engendered? What do the trend lines show, improvement month to month, or how about from the same time last year? Number of days since the last late incident? How about some verbatim comments from the kids and some sentiment analysis on how it made them feel? We can quantify that too by using Natural Language Processing (NLP), converting the words into numbers and visualizing them in a Wordle for impact. Lots of big red words are bad. My scorecard seems to have a lot of them. As a researcher I could say “that’s interesting” but I’ve learned that CEOs and spouses never find it so. Just because something captures interest, doesn’t make it interesting. A fail is just a fail.

Let’s move on. From a load of Descriptive statistics, there is irrefutable evidence I am bad at getting to places on time. Nonetheless, as a technology for decision-making in the moment I offer that Scorecards are not particularly helpful. We are wasting time digesting and debating summaries of my past performance versus driving toward our goal. Perhaps it could be reviewed later, during an annual marital Performance Review.

How about a Dashboard? Businesses use them for decision-making but sometimes they get confused with Scorecards. We use a Dashboard when we need real-time feedback including alerts on critical information. The data must be digest-able at just a glance: do not drive with the parking brake on, do not run low on oil, watch your gas level. Pay attention to your speed en route — police do not issue tickets based on your mean, median or modal speed!

Streaming critical real time data will be helpful in getting you to your destination, helping you avoid major mistakes that would cost you considerable time against your urgent travel needs. In fancier cars, you may even be provided Predictive Analytics, letting you know how much further you can drive based on the way you have been consuming gas up to now, and so on. Note to the CEO here: it does little good to yell at your Dashboard “How do we get there? We’re in a hurry!” Those are the wrong insights to expect from your Dashboard. No you have not wasted your investment and yes the data is actionable. People who drive the car find it quite helpful. Point in case, with just a glance at our Dashboard (perhaps a second or two of effort) we know if we are “good to go” to pick up the kids. But unfortunately, we still have not moved.

Map it out

So now we need a map. Very obvious you say — but find me one at your workplace for all the urgent decisions at hand. Why are people pouring over Scorecards and Dashboards to answer the urgent questions of “How do we get there?” Simple questions can be difficult to answer. But confused application of technology and tools won’t help you solve the problem. Stop creating more Scorecards and Dashboards with more bells and whistles. In fact, do an audit of how many “point solution” software applications you have running, each with their Scorecards and sometimes Dashboards, and consider consolidating.

Since running businesses involves a lot of very urgent “How do we get there?” moments, the next “big thing” for technology in my view ought to be better Maps. Better Journey Maps and Mind Maps. They need to be dynamic, give us options and allow for heated and constructive debate with the Domain Experts.

An electronic map like Google provides real-time feedback including Prescriptive Analytics, where it still offers you different options for your route home. It also has behind the scenes Decision-Making Algorithms and Machine Learning which simply means that the more people use the technology over time, the smarter the analytics gets, with more data to inform the route options. Problem solved, I naively and confidently say. Go with technology to get to the kids on time because we have up to the second information on current traffic patterns. Nope. I still have not convinced my spouse, the Domain Expert. Please, no misunderstandings here — I am not “mansplaining” the value of technology and analytics. She is equally proficient.

Decision time!

We had a brief debate about my Scorecard, which as an aside was not at all motivating nor empowering like it’s supposed to be. I have taken a glance at the Dashboard and thankfully this is one is on my side! Not one of the days I decided to see if the car can actually still drive after the gas gauge shows empty.

And here is what my wife and I agree upon: The up to the minute map is helpful in knowing which route not to take because it will show unanticipated current traffic jams on the route she prefers as fastest. If there is an accident, which thankfully does not happen often, we take a different route. We also agree that if we did not know the city at all, we would be bound to follow Google Map’s advice because the option of relying on our ancient abilities to track our whereabouts by the position of the sun and stars is not all that accurate.

So what is the problem? She disagrees with Google Maps advice and our disagreement comes down to this. The real time data as helpful as it is, can occasionally be misleading. We leave downtown and the traffic patterns make clear which is the fastest route to take. I’m driving and I make the call — Google wins. I smugly look at our estimated time of arrival in 24 minutes versus the estimated 29 minutes that Google predicted for my wife’s preferred route. However, as we progress, I realize that Google does not seem to know that all the schools along its preferred path let out around 3:30pm. Not until we are within what looks like 10 minutes of our destination does the Google machine nefariously change its projected arrival time and stop us virtually in our tracks. Seriously?! It only tells us once the roads are clogged — not 20 minutes before when we could have chosen a better route. Our Domain Expert knew this. She had been adamant. She will be proven correct.

Unfortunately, for me, the other Scorecard showing argument “wins” is now also on a precipitous and statistically significant downward slope. I am not enjoying all these Scorecards and think a bunch of this data should be buried. But this is not about me. Let’s think about the kids. They had to wait for 10 extra minutes and had to even contemplate wandering home on their own, as I did when I was a child and none of this technology existed. The good news is we have been in touch by cell phone and according to Google Scholar, this does not tightly fit the definition of abandonment. Thank you modern life.

Big Data versus Domain Expert

So Google has helped but not saved me from another strike on my win/loss chart. Having “Big Data” provide instant up to date information to inform decisions was not enough to trump our Domain Expert. Between when we start a journey and when we finish it, we learn that the data can change. Domain Experts sometimes know better. It is why they get paid the proverbial big bucks and ought to make the urgent, important decisions.

Bottom line is Scorecards are great for telling us about the past, dashboards are great at telling us about the current, and maps (of various kinds) ought to be great at guiding our future decisions. However, the reality is even the most analytical and data driven companies are missing the opportunity to do better in mapping decisions by utilizing machine learning and artificial intelligence.

In an era of Big Data, much of what my Domain Expert knows, technology could know better. Previous history is a good start, especially for known repeated events like when schools get out during weekdays. Further, in an era of the “Internet of Things” we could also predict the impact of “one off” events, like when a ball game lets out. Contemplate additional fun with the concept, a prediction of whether a game will be won at the bottom or the top of the 9th. If you have driven by a ball stadium when it lets out, you know exactly what I mean. Let’s Money Ball these lively marital moments with what mathematicians can know better.

If we really want to displace our Domain Expert and improve on our odds of arriving on time and without argument, we need to consider Google’s brave foray into Artificial Intelligence and consider applying them further to this domain. Humans make human judgments. When I’m in a hurry I like to move fast, not sit through clogged traffic, even if it takes just a titch longer. I prefer beautiful boulevards versus unsightly roads, all other things considered. I choose safe over sketchy. Just as well as we can program options such as toll roads or not, why not program more options such as the ones above? Why not build artificial intelligence to reflect changing preferences and human decision-making?

Let’s return to the practical matter at hand and answer the original question.

Where did my decision-making go wrong?

To answer this, I offer my combined analytical and consulting advice garnered from decades of experience in business and in marriage, in process and in relationships, in technology for business and technology for personal use.

Hand your phone with the Google map loaded up to your partner and ask “what do you think?” This is the best data-driven, evidence-based, optimized business decision-making. Put the right technology in the hands of the Domain Expert in charge and let them solve it. Old thinking was “man versus machine”. New thinking is woman with smart-phone.

About the Author: Armed with a PhD in Social Work and a penchant for research and statistics, I set out on a course to better understand the world. I’ve spent the past 20 years helping enterprises better understand their customers, employees, and broader public sentiments, leading to better decision making.

back to blogs