Because CaliberMind leverages machine learning, our ears perk up when research, surveys, and articles come out on the topic. When we read “Why machine learning strategies fail” by Ben Dickson on VentureBeat, we weren’t surprised by the title. What shocked us was that 4% of survey respondents told Rackspace Technologies that there are “no barriers” to actionable insights.
Who are these people?
Companies routinely underestimate the work needed to get actionable insights out of their data. As Ben pointed out, skilled data scientists are really expensive, and there’s a significant deficit in data maintenance personnel who have the understanding and time needed to connect disparate data sources and cleanse unhygienic data.
Businesses that outsource their advanced machine learning analytics and don’t have full buy-in from leadership don’t dedicate the time and energy necessary to pass along much-needed business context to produce something valuable (Ben also pointed this out). Data only reduces overhead and improves revenue if you know what it’s telling you and can adjust accordingly.
It’s worth pointing out that all of these issues come into play with advanced marketing analytics. So when someone says they have zero barriers to achieving actionable insights, we had to wonder who would or could make that claim.
To paraphrase Law & Order:
The following story is fictional and does not depict any actual analysts or organizations. Particularly ours.
The Jedi
It’s rare to find a talented individual who would rather not be in the spotlight after successfully wrangling hundreds of thousands of rows of data into something that could save the business a lot of money. Few people have the willpower to refrain from a happy dance or company-wide email blast.
The Jedi are among this unique crowd.
They know management wouldn’t understand (or would be horrified if they knew) the calisthenics necessary to whip data into shape. Instead, management has learned to trust the Jedi because they consistently get stuff done. There’s a vague sense that what they do is difficult, but beyond that, people are spared the gory (boring) details.
Because of these elusive coworkers’ secretive nature, they are often drawn to magic (the performance art and sometimes The Gathering) at a young age. They have a talent for noticing patterns, logic, and sleight of hand.
If you have one of these brave data warriors in your organization—the kind of warrior who consistently delivers sound insights without complaint—be thankful they protect you from the dark side of disparate data sources and manual data entry error rates. And may we also suggest gifting them a Japanese Whisky flight as a thank you.
The Visionary
The Visionary is often misunderstood by those around them. Sometimes they present as flaky and lacking common sense, but the biggest hurdle is the distance between their intelligence and ours. I remember a young man in my computer science courses who liked to tell jokes in binary (do you have any idea how long that takes???). Sometimes Visionaries like to do something for the sake of saying they did it.
(On a related note, visionaries in the life sciences are how we will get zombies or dinosaur clones.)
Despite their unconventional approach to work and life in general, somehow, they get things done. The logic they use is beautiful and unrepeatable, meaning you’re in deep doo-doo if they ever leave the company.
Because visionaries frequently are misunderstood, you better find a way to keep them happy and compensate for the cold shoulder they get in the lunchroom because of their 30-minute lessons on home-crafted IPAs to anyone who doesn’t scamper out of their path in time. Might we suggest this mug as a small token of your thanks?
The Rookie
When a data scientist or analyst is fresh out of college, they have this sparkly look to them. No, not sparkly like Twilight vampires. We’re thinking the fresh, sparkling glow of someone who hasn’t yet spent four days going down an unexpected rabbit hole of “clean” data that is very much the opposite of clean. The experience is much like biting into an “apple” and finding out it’s really an onion. There are a lot of layers, and they don’t necessarily improve as you dive deeper.
Everything is “what you see is what you get” the first few months to years of your career, but then we all learn to anticipate that those “apples” aren’t what they seem and set expectations with clients accordingly.
We get it. As data and systems analysts, we’ve all been there.
We’re not saying that Rookies mature into more jaded individuals. We are saying that they learn to pad their delivery timeline like they’re setting up a bounce house for a horde of sugar-fueled toddlers. This is precisely why so many of us learn to negotiate timelines with upper management after they react in shock that a “simple” request takes three weeks.
Until then, the Rookie takes requirements and underlying data at face value. Which is adorably naive and completely understandable! We all need to learn some things the hard way, and setting expectations is certainly one of those things. Until then, those of us who are a little more seasoned will look on in fond recollection of simpler times.
The Perpetual Optimist
A few good-natured Rookies don’t mature into “seasoned” professionals. Instead, they retain their optimism that this time things will be different. The data will be clean, and everything will go as expected.
And you know what? That’s okay. We need the optimists to balance out the “realists” (aka pessimists). These professionals understand what a face-value timeline looks like, and will help rein in an analyst who’s been particularly traumatized by a landmine of duplicates and a plethora of null values.
For organizations thinking about diving into advanced analytics, it’s far from impossible to get actionable insights. However, understand that getting the numbers right means an investment in the right talent, technology, and processes. Even if you outsource your analytics to a kickass company like CaliberMind, you’ll need to dedicate time and resources to passing on valuable business context and deciding which normalization and data hygiene best practices to mirror.
