The Inside Track: Turning a Decade of Supply Chain Data Into Decisions

Open a shipment-tracking app today, and the dot is usually there—which is no small thing if you remember how much freight used to run on check calls, stale ETAs, and someone's best guess. Real-time visibility has done its first job. It found the freight.

The harder question starts right after that: what should you do with what you can now see?

That's where my team comes in.

My name is Cathy Slesnick, and I’m the Director of Data Science at Tive. I joined as Tive's first data hire in December 2021, which means I had the rare privilege of building the data function from scratch. The team, the stack, the plumbing, the direction. All of it.

This is where the story really begins: not with simply seeing freight, but with understanding it. Years of supply chain data, network patterns, and plenty of behind-the-scenes work are what make it possible to turn real-time shipment visibility into decisions people can act on.

How I Got Here

To explain where we are now, it helps to start at the beginning—because when I arrived, almost none of this existed yet.

Walking Into a Candy Store

I didn't join Tive because I have a lifelong passion for logistics. I joined because Tive looked, to a data person like me, the way a candy store looks to a kid. Here was a company with sensors moving all over the world, generating an enormous stream of data about shipments, conditions, and routes, and almost none of it was being fully used.

For someone who gets a real thrill out of finding patterns in data (I have a PhD in astrophysics, so staring at enormous datasets and waiting for something to reveal itself is kind of my thing), that was irresistible.

Building the Plumbing 

Before you can do anything clever in data science, you have to do a lot of grunt work. When I started, we were collecting raw information but lacked the underlying foundation to actually use that data for meaningful insights. As such, the first few years went to the least exciting but most necessary work there is: cleaning, organizing, and building what we call a data transform layer.

In plain English, that means taking all the raw data coming off our sensors and systems, scrubbing it, joining it together, and turning it into reliable datasets someone can build on. It's the plumbing inside the house: you never see it, but nothing works without it, and it's the part most people underestimate.

Crawl, Walk, Run

In my mind, the arc is crawl, walk, run.

When I arrived, the vision for supply chain data products was barely formed. But once our current VP of Product, Kyra, got behind it, things started moving fast. The timing made sense: hardware grows only so quickly, while software and data products scale far faster, and the companies that win turn their data into momentum.

Now I lead a team of 13 people across three disciplines: data science, data analytics, and data engineering, plus analytics partners embedded across the company. The team is spread out, with data science in Boston, analytics between Boston and Kosovo, engineering distributed, and QA up in Canada.

With that team in place and the data clean, we now get to spend our time deriving insights and building products instead of wrestling with the basics.

What We're Seeing Now: People Want Context, Not Only Piles of Data 

Once the “crawl, walk, run” work is done and the data finally behaves, the conversation changes almost immediately. Teams stop asking for raw information and start asking for actionable information. As supply chain data matures, the focus moves away from visibility for its own sake and toward helping people decide what to do next.

Start With This One

Most freight teams, though, aren't wired that way. They open the day to alerts that arrived while they slept, trackers that paused somewhere inconvenient, temperature readings that moved just enough to annoy everyone, and a map full of dots that all look equally needy until one of them becomes expensive.

That's the first trend I keep seeing: the appetite for raw supply chain data has dropped, while the appetite for actionable insights has exploded. Teams want the system to read the room. They want it to say, "Start with this load, then look at that carrier, and don't waste your morning on the other eight alerts unless something changes."

"Normal" Has Become the Hard Part

A shipment only looks strange after you know what strange means, which sounds obvious until you try to build the system behind it.

A truck leaving its expected route might mean the driver knows a better local road. It might also mean the load just moved into a theft pattern. The difference depends on context: the lane, the commodity, the corridor, the stop history, the time of day, and the thousands of quiet trips that taught the system what normal usually looks like.

Smart Route Deviation Alerts were created to combat that exact problem. The system has to understand the roads shipments usually take before it can flag the road that deserves a second look.

The Network Starts Filling in the Blanks

Once the system understands normal movement, it can begin to infer the things customers never enter into a form. Ground truth data matters immensely for this reason. Of course, I would love a world where every customer carefully labels every shipment for us, but that world doesn't exist.

Take berries. A customer may never tell us the truck carries strawberries. However, the shipment can still move like berry freight, behave like berry freight, and resemble the alert patterns we see across other berry freight. From there, the system can recommend a smarter alert setup without asking the customer to build one from scratch.

Trust has to sit inside that kind of intelligence from the start, too. We cluster endpoints, leave private one-off stops out of route boundaries, and use the network pattern rather than the customer’s private footprint. Collective intelligence should make each shipment smarter without making anyone feel like the network has started peeking over the fence.

The Real Ask is Earlier Warning

After customers get used to priority, context, and cleaner alerts, they ask the question everyone has wanted to ask all along: can you tell me something is wrong before the problem gets expensive?

You see that request everywhere. Which ports hold freight for three days so often that teams should plan around it? Which carriers keep tripping the same alert? Which lanes create reefer issues, cargo theft risk, or recurring delays that look random from one shipment but obvious across the network?

The reefer alert we just rolled out fits that pattern. Customers used to wonder whether a driver turned the unit on when the load needed protection. Now they can see it, prove it, and hold the right party accountable when a shipment fails — the same record that flags the problem in the moment becomes the evidence that settles who owes what later, and the lesson that tells the next shipment where to look first. 

Supply chain data starts to earn its keep at that point. Yesterday’s freight teaches today’s system where to look first.

The Best Part of the Job: Seeing the Whole Picture

I'll end with the part of this job I love most. When people ask what keeps me in it, these five things are the honest answer:

• A view across everything: Because my team runs the data warehouse, essentially every kind of Tive data flows through us—from sensor readings to shipment records to who is ordering what. A team working with a single slice can't see how those pieces connect, and that vantage point is what I value most about the work.
• The moment a pattern appears: Remember those Magic Eye posters that looked like noise until your focus shifted and an image surfaced? Good data work has the same payoff: you sit with an overwhelming amount of information, wait, and a pattern that was there all along resolves. That moment of recognition never gets old.
• Running an anomaly to ground: Some of the most satisfying work is pure detective work. A teammate recently noticed we were being overcharged by a vendor, and kept pulling the thread until he traced it to the root cause. Following a strange signal all the way to its source is genuinely rewarding.
• Building it from scratch: As Tive's first data hire, I've had the rare chance to shape this function from nothing: hiring every person, choosing the tech stack, and setting the direction. Not many people get to build the thing they then get to run.
• Turning data into decisions: This is what makes the rest matter. A pattern that's invisible inside any single shipment becomes obvious across the whole network—which is what enables us to hand a customer a clear answer instead of a pile of data. Watching that reshape how someone runs their supply chain is the most meaningful part of all.

The View From Here

When I look back at the candy store I walked into in 2021, the change is hard to overstate. A flood of underused data became a clean foundation, and that foundation became a platform that doesn't just show you where your freight is: it tells you what to do about it. The next decade of this industry belongs to intelligence, not just visibility, and that future gets built one pattern at a time.

This is exactly where Tive lives. We're one of the few companies that pairs real-time tracking hardware and visibility software with more than a decade of network data. We turn that combination into intelligence you can act on: smarter routing through Smart Route Deviation alerts, earlier signals on theft and risk, foresight about where shipments will get stuck, and recommendations drawn from the behavior of an entire network.

We're not trying to be the only company doing this. We're trying to do it the best.

Get started with Tive today to learn more.