TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

The Standard BI Playbook Wasn't Built for the Physical Economy

Generic analytics advice gets a lot wrong when applied to industrial distribution and marine operations.

• By Hasan Zafar
• September 16, 2026

An unpopular opinion from someone who builds analytics for industrial operations every day: most companies in distribution, manufacturing, and marine transportation don't have an analytics problem. They have an analytics democratization problem.

The analytics exists. It's just trapped in one team's spreadsheet, one analyst's legacy BI report, or a homegrown tracker nobody else knows about. An organization becomes data-driven only when analytics moves from the individual to the organizational level. And almost everything in the mainstream BI playbook (built for SaaS companies with clean cloud data and digitally native users) quietly assumes that problem doesn't exist.

I work at the intersection of industrial distribution and marine transportation—sectors still running significant operations on manual reports unchanged in decades, yet moving an enormous share of the physical economy: roughly 465 million tons of commodities worth over $158 billion move on U.S. inland waterways every year, feeding sectors representing about a third of U.S. GDP, according to the ASCE 2025 Infrastructure Report Card. Here's what the generic advice gets wrong.

1. The Playbook Assumes Your Data Is in One Place. Yours Is in Twelve.

Modern BI content starts at “connect your data warehouse.” In industrial environments, there is no warehouse to connect. There's a massive legacy ERP, powerful but user-hostile. There's a homegrown system, grown massive over the years, that only a handful of people truly understand. There are apps capturing pieces of the operation, and spreadsheets—so many spreadsheets—each providing a slightly different version of the truth.

High-value reporting cannot be built on fragmented data, so the first real analytics investment here isn't a visualization tool; it's the unglamorous work of centralizing onto a modern platform. Skip that step and you're automating the fragmentation.

2. The Playbook Assumes People Want the Report. Everyone's Too Busy Making Their Own.

In a decades-old manual reporting culture, every team is heads-down producing its own numbers just to hit deadlines: reconciliation, validation, compiling, hours every cycle, human error baked into every handoff. The risk isn't hypothetical: research on operational spreadsheets led by Raymond Panko at the University of Hawaii finds field audits uncover errors in roughly 88% of spreadsheets examined. An organization running its reporting on spreadsheets is, statistically, running it on errors.

Worse, teams don't know what others have built—I've watched groups spend weeks constructing something that existed one department over. Even good tools go unused when socializing analytics is treated as an afterthought instead of half the job.

A real example: a sales team I worked with, renting out machinery, couldn't reliably determine (even with a report) whether inventory was actually available to rent, wasting hours getting to the right info every week. The data existed; a single source of truth didn't.

3. The Playbook Assumes Analysts. Industrial Teams Have a Different Kind of Expert.

The generic advice assumes a bench of data-fluent people in every function. In industrial companies, analytical skill is unevenly distributed; some teams have it, most don't, and those that do work solely for their own team's success. Meanwhile the teams needing analytics most—operations, maintenance, dispatch—are deep domain experts with earned skepticism about a new dashboard changing twenty years of habit. That resistance isn't a character flaw; it's what happens when every previous “system rollout” made their lives harder. Adoption here is a trust problem, not a training problem.

4. The Playbook Optimizes for Analysis. Operations Needs Action in Minutes.

SaaS analytics culture thinks in quarterly funnels. Industrial operations think in today: which barge is waiting, which order is short, which claim is stuck. Without live or near-real-time visibility, teams cannot act—and with a modern data platform, any refresh frequency is on the table. Real-time where minutes matter, daily automation everywhere else.

Another pattern: an operational team running at a hundred miles an hour just to file and process warranty claims, with zero visibility into amounts claimed, amounts received, or aging payments equals an unquantifiable audit risk. All motion, no instrumentation. Integrating claims data directly from the source changed the question from “did we file everything?” to “where is our money and why is it slow?”

5. The Playbook Measures Dashboards Shipped. The Real Metric Is Hours Returned.

The most honest KPI for industrial analytics is weeks of work turned into hours. Field notes from my projects:

• A sales commission process that was entirely manual—weeks per month compiling sales data, email back and forth, chasing approvals, corrections, and finally payments. Automation gets accurate data to managers immediately and approved numbers to accounting far faster. Both the accounting and sales teams are able to do so much more.
• Daily operational reports manually assembled and emailed for years are now generated and distributed automatically, with the narrative summary written by AI. Same report; human hours near zero.
• Engineers spending hours hunting across systems for whether a part exists, is in stock, and who makes it are replaced by a search experience answering in seconds.

None of this is exotic, which is precisely the point. In digitally mature industries, these would be table stakes; in industrial sectors they're transformations, meaning the ROI on basics is enormous and largely unclaimed. Deloitte's Digital Maturity Index research on global manufacturers finds higher digital maturity translates directly into higher EBIT and revenue. The laggards aren't just less modern, they're leaving money on the table.

What to Actually Do Differently on Monday

Invest in an organizational-level data strategy, not team-level tooling. A central platform combining fragmented sources is the prerequisite for everything else; every dollar spent on visualization before centralization is spent polishing fragments. Most organizations fail here: McKinsey's global survey on digital transformations found only 16% succeed at both improving performance and sustaining the gains—in traditional industries, even fewer.

Treat socialization as half the project. A solution nobody knows about has the same ROI as one that doesn't exist. Catalog what exists, demo cross-functionally, celebrate reuse as loudly as new builds.

Audit prioritization for influence bias. What gets built is often decided by the loudest voice; some of your highest-ROI solutions sit in the backlog because their proposer lacked pull. Score requests on business impact, openly.

Break team boundaries deliberately. Teams leveraging each other's skills and work is the cheapest capacity you'll ever add.

Automate one painful, visible manual process end-to-end. Not the flashiest use case, the most felt one. Nothing converts skeptics like a hated weekly task disappearing.

Design for the frequency the decision needs. What action does each report drive, and how fast must it happen? Frequency is a design decision now, not a technical constraint.

Measure hours returned, not dashboards delivered. “Time given back to the business” is the number executives feel.

The physical economy—the barges, warehouses, engines, and parts everything else depends on—runs on a fraction of the analytical capability digital-native industries take for granted. That is the biggest open opportunity in analytics today. Companies that centralize their data, democratize their analytics, and socialize what they build won't just save hours, they'll change what their people spend their working lives doing.