AR Data Sciences
EDITOR’S NOTE: This month, in lieu of a regular Guest Column, we bring you a handful of short-ish insights and observations that for various reasons didn’t make it into our featured Q & A’s over the past year. We try to keep our published interviews in the 2000-2500-word range, and yet some of the raw transcripts of the actual Q & A can, with digressions—many of them fascinating—run as long as 15,000 words. That necessarily leaves a whole lot of good, provocative stuff on the “cutting-room floor,” as evidenced by the following four examples.
COVID SKEWED OUR DATA
Let’s put 2020 in a closet and lock the door
Amy L. Hester, Ph.D., RN, BC
Chairwoman and Chief Executive Officer
HD Nursing, LLC
COVID REALLY AFFECTED the quality of healthcare in general. When you think about quality assurance, making sure that your quality stays stable, somebody has to keep the lights on all the time. There has to be somebody who’s always watching those quality indicators, whether it’s about falls, infections, pressure injuries, retained foreign bodies in surgery—whatever quality metric you’re looking at.
But when COVID came along, everybody turned their attention to that and so they turned off the lights on quality. And a lot of our quality outcomes started going in the wrong direction, because everybody was distracted. The hospitals that kept a team focused on quality assurance weathered the storm better, but organizations that didn’t have the team, the bandwidth, the resources, and so were all-hands-on-deck to fight COVID, those are the ones who’re now playing catch-up. By the end of 2020 they were like, “Oh, what happened? What’s going on, we’ve got to fix this!” So now they’re working like crazy to get their quality up to where they were so that they can be in that quality assurance cycle instead of quality improvement. How a hospital performs in various quality metrics affects its overall rate of reimbursement and penalties from CMS—the Centers for Medicare and Medicaid Services. It’s called Value-Based Purchasing. If a hospital’s quality rating is high, CMS reimburses it at a higher rate.
So now we’re in conferences talking about how we use big medical data, and whether we need to just put 2020 data in a closet somewhere and not even look at it. Because it’s all confounded data, it’s all affected. Every bit of 2020 data we have is affected in some way. People didn’t go to the ER’s for normal things; they waited until they got really, really sick with their chronic conditions because they didn’t want to go to the “COVID” hospital. And people weren’t having surgeries—the elective surgeries were completely cut off. We did the right thing by doing that, but it has really skewed numbers, and events, and outcomes. So we have a lot of data, but it is affected data, and I think we’re going to have to account for that in any kind of modeling that we do.
YOUR MICROWAVE ISN’T MAGIC
Updating the definition of a literate person
JAN P. SPRINGER, Ph.D,
Director, Emerging Analytics Center,
UA Little Rock
DO YOU KNOW how your microwave works? Probably not. We have many such gadgets in our lives, but the entry level for understanding them has become very high today. You can’t just run one program, as in the UNIX way, like 50 years ago. No, today you have to install a software development environment, and you need a license to use it, as if you were going to make money with that. And all you want to do is to know how your gadget works.
What Arkansas is doing with the computer science components in the middle and high school is a counter-action to this. It is a really good counter-action, even if we maybe do not agree with all of what the teachers are doing. It's better than not doing it at all.
The way I see it is, we have these “literacies” that we need to master if we’re to be educated people. We have to be literate in reading and writing, and then there is coding, or I would say doing things in an algorithmic way. And we need to have a certain literacy in math.
But math at a good high school level usually only gives you a good introduction—or intuition, if it's really good—to numbers, especially going into high numbers, and tendencies, and nondeterministic processes like stochastic statistics, because that's very hard. And there are basically things you can use to analyze that, too. But it doesn't give you a good insight into how your microwave works.
The classical thing people in Germany get introduced to in algorithm courses is an automated teller machine—an ATM. What does the ATM do? How does it go from one state to the other, how does it determine that you get money or you don't get money?
So to me, there should be a fourth component of literacy: reading, writing, math, and then something to do with algorithms. If we don't do that, people will not survive in this world anymore, because they think everything is magic. No, it's not. It's just highly complex.
LIFE IN A BLIND SPOT
Why aren’t there incubators for
existing tech companies?
Matt Olson, President,
Matmon Internet, Inc.
I'M PROBABLY JADED when it comes to Little Rock's efforts to help tech companies. Because they seem to want only the ones that are startups, that are starting from nothing. I've always been in a blind spot, it seems. I was always waiting for someone to knock on my door and go, “You're a tech company. You've been doing really well, and we've got something for you that's really going to take you to the next level.” But no, nothing, never. It was always more like, “Oh, you did it. Okay, well, how can we help more other companies be like yours?” It's never been, “How can we take the ones we have and help them compete at a different level.” So I really never found a place over at the tech park. I never found a place at any government funded anything ever.
During COVID, when I got this PPP, I couldn't believe that we got this much money. Getting two and a half times my payroll from the government—that’s huge, you know. I was like, Wow, they really did something for me. But I’m proud of what Little Rock is doing. I like the incubator things—I think all that is great, and it's needed, and I think it's been well received. I'm proud of Little Rock's venture center, and tech park thing, and what you guys at ACDS are doing. I'm proud of our state for all of that. I just haven't benefitted from it yet.
What I wish is that Little Rock could find ways to focus on efforts to help companies that are successful and need assistance going to the next level, instead of focusing primarily on startups. I bet efforts like that would have a nice success rate and a bigger return on investment for the city. Take what's already working and add to it, so to speak.
THE PARADIGM PARADOX
Don’t be blinded by “the way it’s always been”
Chairman and CEO, Inuvo
THE CORNERSTONE OF great marketing and advertising has always been the same, and it's the information that you have available to you to be able to make a decision. This has not changed since the dawn of marketing. Think back to even 50, 60, 70 years ago—there was no data, so to speak. But there was data. Your local banker knew everybody in his community. As a result of knowing them and what they did and how much money they had or didn't have, he was able to know who he should lend money to. So information, in whatever way it's available, is the difference between getting a customer or not getting a customer.
Back in our Acxiom days, this was our area of specialty. We understood information, consumer information—where it is, how you get it, how you bring it together. But using that information is a very cumbersome, human process involving a lot of trial and error. You kind of figure out who your clients are, then you buy data on them, you overlay it, take a snapshot of it, and say, okay, I guess my customer is someone who makes a hundred grand a year and loves fashion and maybe has a family. That's my audience—right?
Today, with advancements in technology and artificial intelligence, we see that process as limiting. It's limited not in the sense that it's bad. It's just a technology and process constrained by its paradigm—the way it is, the way it’s always been, at least for the last few decades since Acxiom invented that consumer data industry. But the challenge with paradigms is that they become normal when they shouldn't be normal.
For example, why are train tracks only a certain distance apart? Nobody ever asks that question. They just assume it's okay because that's the way it’s always been. But the fact is, they’re spaced that way because you go back into Roman times and the wheels on a chariot were spaced exactly the same distance as the train tracks are today, because that’s the only way you could get two horses together inside the tracks. So that became the paradigm. But think about the consequences of that today. Now they build tunnels, and you can only get merchandise on a train that’s a certain width. You can’t put bigger things on it. And all because you’re trying to fit two asses in between the chariot wheels.
At Inuvo, we’re asking the questions: Why is marketing still done this old way? Is there a way to create a different solution that gives us a better sense of what the real intent of the consumer is? And then, can we find a technology or technologies that can solve that problem with the added benefit of not invading people’s privacy? And the good news is, we did.