Logistics Disruptors: Using big data to spot shipping anomalies

Intelligent Audit analyzes data gleaned from its customers in shipping and third-party logistics. It searches for anomalies, and when it spots interesting information—exorbitant cost outliers, glacial delivery times—it strives to transform numbers into actionable insights. Based in New Jersey, with regional offices across the globe, the company says it audits more than a billion shipments a year.

Hannah Testani’s father started Intelligent Audit more than two decades ago. Testani, 34, had always wanted a career on Wall Street, but after the 2008 financial crisis complicated that plan, her father offered her an opportunity to join the family firm. “At first, I hated it,” says Testani. “This was before logistics was sexy. No one knew what supply chain was because consumers didn’t feel personally affected by it. Now, everyone knows supply chain because they have been personally affected by some supply chain shortage or crisis somewhere.”

In this installment of Logistics Disruptors, Testani speaks with McKinsey partner Sandy Gosling about how to improve logistics by sifting reams of data, implementing machine learning in business, and throwing away everyone’s ballpoint pens.

The following is an edited transcript of their conversation.

McKinsey: Your company’s name says what it does. Can you walk us through an example of the kind of audit you do when you begin working with a customer?

Hannah Testani: We recently onboarded a customer that ships more than one hundred million shipments a year. They were using database software that couldn’t support even the most basic analytics. So they didn’t have the ability to properly look at their own data in a way that would let them understand their transportation spend and identify opportunities to optimize. We ingested their raw data and cleansed it to make it easier to interpret. Once we did that, we quickly found opportunities for them to reduce their transportation spend, reduce the time in transit for their shipments, and reduce their carbon footprint.

For instance, we saw that they were using boxes that were bigger than necessary for their shipments, which meant they were wasting a lot of money paying for unused space. We also saw that they were using air shipments in situations where ground would actually have been faster and cheaper—because if you pay for a two-day air shipment service, the carriers will take two days to deliver even when it’s one day by ground. And we also spotted some consolidation opportunities: they had multiple shipments leaving the same delivery center, on the same day, using the same service, going to the same customer, which means they were paying multiple times for something they could’ve paid for once, because three one-pound shipments cost triple the amount of one three-pound shipment.

illustration autonomous truck driving off into the future

Logistics Disruptors

Meet the companies shaping how goods will move tomorrow.

We believe that every shipper should have access to clear, actionable transportation data that enables them to ship smarter—meaning faster, cheaper, and with less delivery exceptions. We use our proprietary technology to turn their complex, global transportation data into actionable intelligence that shows our customers exactly what they can do to optimize their transportation spend. If they spent a billion dollars last year on shipping, we’re showing them how to make that $900 million, with quicker execution and better customer experiences.

We also care about the transportation carrier’s experience. We feel like it’s our responsibility to get them paid on time for their services while holding them accountable. Within two hours of receipt of any electronic invoice, we complete the freight audit and work with the carriers preemptively, before the invoice is paid, to achieve a resolution on any disputed amounts. When we identify any discrepancy, we work directly with the carriers to identify the root cause as quickly as possible and collaborate with them to avoid similar discrepancies in the future.

McKinsey: There continue to be many unknowns in logistics—disruptions, conflicts, labor shortages—to navigate these days. What are some new kinds of challenges your customers are facing?

Hannah Testani: One central issue is that our customers are being forced to diversify across their supply chains, between their manufacturing and their transportation carriers. If a customer was previously sourcing their manufacturing solely from China, they’re now also manufacturing in Taiwan, Vietnam, or Latin America because they feel they have to. The easy button for many of our customers used to be to source their transportation with as few transportation carriers as possible. But after the recent disruptions, they have diversified their carrier mixes. Nodes have multiplied. Supply chains have become more complex, and so it can be much harder to keep your finger on their pulse.

If you can’t process your data and understand what it means, then you can’t understand the nature of your supply chain spend.

The upshot is that, because of how complicated things have become, the amount of noise has grown exponentially. If you look at other big industries with this amount of data—like finance and banking—they are light years ahead of where we are in transportation. So many companies in our industry still rely on paper bills of lading to run their supply chains. To supplement the paper documents, they might use an Excel file, which can’t support that much information. What this means is there’s a lot of garbage data out there. If you can’t process your data and understand what it means, then you can’t understand the nature of your supply chain spend. It’s impossible to save money in an area you can’t properly measure.

McKinsey: How do you approach problem solving?

Hannah Testani: We have always leveraged technology to find the most efficient and repeatable solutions for our customers. About two years ago, we onboarded a new data science team. The leader that we chose came from the healthcare space—using big data to analyze MRIs, find patterns, and detect fatty liver disease before a more invasive biopsy would.

We’ve tried to use a similar approach. We use deep learning models, which take our customers’ data and analyze it through both long-term and short-term lenses, at various levels of granularity—from a macro level down to a division or a business unit. Using these models, we can find anomalous patterns in our customers’ data.

For instance, one of our customers is a big retailer that had just started shipping a new SKU from their stores. Our anomaly detection sent us an alert because it noticed a shipping cost that was 30 times more expensive than this customer’s average cost per shipment. So we investigated. What we found was that the new SKU happened to be a large, bulky item, and it was incurring an additional handling fee that brought the total cost to $140 for a single shipment—an amount that was far higher than the margin on the item. Within days of the alert, we were able to bring that information to the customer, which made it decide to make that SKU only available for in-store pickup.

With another customer, our anomaly detection tool noticed disproportional costs for shipments coming out of a Texas location. It turned out they’d had a team member come from Europe to that delivery center in Texas. In Europe, they use a comma the way we use a decimal point. But the software in Texas wasn’t treating the comma as a decimal point, so instead of a five-pound package it was interpreting the weight as 500 pounds.

Our anomaly detection software has been able to find many issues in our customers' transportation data, almost all of them a result of either a technical glitch or a human error upstream.

McKinsey: If you had a magic wand, what would you change to make logistics better for the world?

Hannah Testani: I would tell everyone that they could no longer write anything down using a pen. The amount of manual processes, the use of paper, and just the general lack of data to make good decisions are all huge issues. Most people in logistics are aware of that, and the first step in solving a problem is admitting that you have one. So I am confident that we are, as an industry, moving in the right direction.

But it is slower than I would have liked. There needs to be much more investment in technology. Ultimately, this industry should reach a place where it uses technology to allow customers to self-serve. I equate it to banking: it used to be that you needed to go to a bank branch to do anything, and now you can do almost anything on your phone, and it’s easy and intuitive. We’d like to empower our customers in the same way.

McKinsey: You’ve been in this industry for 15 years. What has been the most surprising lesson for you along the way?

Hannah Testani: The biggest lesson I’ve learned in this industry is that it’s very relationship based. It is a much smaller industry than I expected. And if you’re able to find the right connections and deliver good products to them, your customers will be your best referrals. Because they tell their peers who to use.

I’ve also learned that as a woman, you’re a unicorn. There just aren’t many of us in this industry. And that makes it hard for someone new to ask, “What’s the path?” I was at a conference last year of about 300 executives, and there were 15 women. The lack of diversity in the industry is shocking, but that gives me an opportunity to show others out there that they can do it—because I’ve done it, and I’ve enjoyed it and have been successful at it. So I hope others will be able to enjoy it and be successful at it, too.

Explore a career with us