Why Do 90% of Data Projects Fail? Meet the "Last Mile" Problem

Wholesale and distribution organizations invest massive capital in ERP systems and advanced BI tools, yet most still struggle to translate mountains of information into a tangible bottom line. This post dives into comprehensive research by McKinsey, exposing the structural failure known as the "Last Mile Problem," and explains why shifting from passive dashboards to Actionable AI is the only way to join the 8% of companies that are truly beating the market.

The Paradox of the Information Age

We live in an era where organizations are "swimming" in data. If you manage a distribution company, you have an ERP (like SAP, NetSuite, Priority, or others) that knows everything: order history, inventory, price lists, and financial transactions. You likely also have a BI layer (like PowerBI or Qlik) presenting colorful charts in board meetings.

And yet, CEOs and Senior VPs of Sales report frustration. They see the potential, but they don't see the money.

In their foundational research, "Breaking Away: How leading companies scale analytics", as well as recent reviews on The State of AI in 2024, strategic consulting firm McKinsey points to a staggering fact: Advanced analytics can generate trillions of dollars in value across global industries. But—and this is a big "but"—most companies fail to capture even a fraction of this sum.

Why? Because they solved the problem of collecting data, and the problem of displaying data, but failed at the most critical problem: delivering the data to the decision-maker in the field. Welcome to the "Last Mile Problem."

What is "Last Mile Analytics"?

In the logistics world you inhabit, the concept of the "Last Mile" is well known: The most expensive and complex part of the supply chain is getting the package to the customer's door. In business analytics, the principle is identical.

The "First Mile" is data collection (your ERP). The "Middle Miles" are data analysis (statistical models, BI reports). But the "Last Mile" is the moment the insight meets the sales rep, the van-sale driver, or the account manager, exactly at the moment they need to make a decision.

McKinsey's research reveals that companies spend fortunes developing sophisticated algorithms but fail to embed them into daily workflows. The result: You have great reports, but field agents continue to work based on intuition ("gut feelings") and old habits.

"Insights are useless unless they are embedded into the daily workflow of the frontline employee. If your data doesn't drive a specific action right now, it’s just noise."

Why is PowerBI Not Enough for Field Reps?

A question that always arises in discussions with CIOs is: "We already have excellent dashboards in PowerBI, why do we need another system?"

The answer lies in the difference between Business Data Analysis and Driving Action. Imagine a field agent arriving at a retail client. They have 15 minutes for the meeting. Is it realistic to expect them to open a tablet, look at a pie chart analyzing sales over the last quarter, drill down into categories, cross-reference with inventory data, and independently conclude that they must offer this specific client a new product in the beverage category?

The answer is no. It doesn't happen. BI systems are passive. They require the user to be the analyst. In the distribution world, the agent deals with hundreds or thousands of SKUs. The cognitive load is immense. They need someone - or something - to whisper in their ear: "This customer stopped buying Product X two weeks ago. Offer them a 5% discount now, and they will return."

McKinsey Data: The Difference Between "Breakaway" Companies and the Rest

In their research, McKinsey identified an elite group of companies, comprising only about 8% of the market, termed "Breakaway Companies."

What is their secret? They don't necessarily collect more data. They simply invest their resources differently. While average companies invest most of their data budget in technology and infrastructure, leading companies invest over 50% of the budget in the "Last Mile" solution—embedding insights into decision-making.

McKinsey's conclusion is unequivocal: Instead of asking "What data do we have?", we must ask "What decision does the agent need to make right now?" and provide them with the answer.

The Perceptual Revolution: From Dashboard to Agentic AI

This is where Insighting's approach comes in. We are not another tool that displays data. We are an AI engine that connects to your ERP, "crunches" the data overnight, and in the morning serves sales reps a "Hit List" of actions. We flip the pyramid. Instead of the agent searching for insights, the insights find the agent.

Feature	The Old Way (ERP / BI / Manual)	The Insighting Way (AI-Driven)
Data Access	Passive: "Log in to the dashboard and search"	Active: "Here is an alert, this is what needs to be done"
Decision Making	Based on intuition or slow manual analysis	Based on Data-Driven Decision Making and prediction
Churn Detection	Retroactive: Discovered after the client leaves	Preventive: Identifying "Silent Churn" patterns in real-time
Upsell Growth	Generic: "Product of the Month" for everyone	Personalized: Specific recommendation per client (White Space)
The Last Mile	The Big Failure: Info doesn't reach the field	The Success: Info translates to action in a click

Proof in the Field – Connecting the "Brain" to the "Legs"

Implementing the "Last Mile" strategy is not theoretical. When you connect the analytic brain (Insighting) to the organization's legs (Sales Reps), the results are immediate.

Take, for example, an organization using a standard ERP. The system is full of historical data. Once Insighting connects, it identifies that a specific customer, who usually purchases $15,000 a month, dropped to $12,500. In standard BI, this looks like a negligible standard deviation. Insighting's AI, however, identifies that a specific category (e.g., cleaning products) has disappeared completely from their orders—a clear sign they started buying from a competitor.

The agent receives an alert: "Attention! The customer stopped buying cleaning products. Offer them Promo X now."

This is the closing of the loop McKinsey talks about. It is the shift from "being right" (with data) to "being profitable" (with actions). To reinforce the credibility of this approach, you can read the article we published in Leading Tech Publications.

Conclusion: Don't Get Left Behind with Dashboards

The McKinsey research is a wake-up call. Investing in technology without a solution for the "Last Mile" is an investment down the drain. To become a true Data-Driven organization, you must stop flooding employees with information and start equipping them with insights.

Is your organization stuck in the passive BI stage? Check out our BI vs. AI page to deeply understand the differences.

Frequently Asked Questions (FAQ)

1. What is the difference between BI and Actionable Insights?

Business Intelligence (BI) displays historical data visually and requires the user to analyze and draw conclusions. In contrast, Actionable Insights are active recommendations generated by AI, telling the user exactly what action to take to improve results, without the need for manual analysis.

2. Why do Big Data projects fail in B2B companies?

According to McKinsey research, failure often stems from the "Last Mile Problem." Companies invest in data collection but fail to embed insights into the daily workflows of field staff. The information exists, but it is not accessible in a way that enables real-time decision-making.

3. How do you implement "Last Mile Analytics" in sales organizations?

Implementation is achieved by shifting from static reports to AI systems that analyze data automatically and provide agents with "Hit Lists"—specific alerts on customers at risk of churn or opportunities for upsell, delivered directly to their mobile or tablet.

4. Does Insighting replace the existing ERP (like SAP or NetSuite)?

No. Insighting does not replace the ERP but sits on top of it as an intelligence layer. The system pulls raw data from the existing ERP, processes it using algorithms, and returns smart insights to sales reps.

5. What is the advantage of using AI for Churn Prediction?

Using AI allows for the detection of "Silent Churn" = a situation where a customer is still buying but has stopped purchasing specific categories. Unlike humans, who struggle to spot subtle changes across thousands of SKUs, AI identifies the pattern immediately and alerts you before the customer leaves completely.