top of page

The 3 Pillars of a Data-Driven Organization

We talk to logistics providers every day, and they all understand there's an advantage to being a data-driven organization. But while many 3PLs and freight brokers have taken steps in the right direction, they quite often hit a wall after investing in business intelligence tools and data warehouses.

What can a 3PL do to become data-driven? Unfortunately, there's more to it than simply using analytics or standalone AI features. Truly data-driven organizations must demonstrate three essential qualities: adaptive analytics capability, the experimental mindset, and strategic alignment.


Adaptive Analytics Capability

It's not enough to rely on the analytics provided by standard operational or transactional software (like a transportation management systems). Truly data-driven organizations possess the capability to continually adapt analytics, key performance indicators (KPIs), and metrics to address their own unique and evolving needs. While tools like Excel provide a good starting point, a more advanced analytics solutions – like Axiom-One – will elevate your analytics capabilities to a new level.

It's also important to remember that before your unleash any analytics algorithms on your data there's initial technical groundwork that's required, including the establishment of data pipelines for cleanup and transformation. This up-front prep work ensures the data is reporting-ready and enables swift and efficient decision-making.


The Experimental Mindset

Data-driven organizations cultivate a scientific, critical, and impartial mindset, steering clear of the natural human bias to embrace only the data or results that align with our preconceived notions. An experimental mindset leads to structuring processes to capture and evaluate data impartially, regardless of any alignment with existing beliefs. Approaches like A-B testing can objectively assess the effectiveness of different strategies. Rapid capture and reporting of insights are emphasized, and individuals who apply experimental techniques are recognized and rewarded (regardless of the message told by the final results).


Strategic Alignment 

Bottom-up analytics and AI approaches, pursued opportunistically instead of strategically, can still give you sporadic victories. But the path to repeated wins with analytics and AI, and the hallmark of a truly data-driven organization, is a strategic and top-down focus where analytics are closely aligned to the high-level objectives of the business. To identify which strategic initiatives are best supported by analytics, considering both data availability and accessibility. When first starting out, prioritize a limited number of strategic initiatives that are well supported by easily-accessible data, and then grow out from there. Craft metrics that are tailored to measure the health of your selected strategic initiatives, and consider how those metrics should be communicated at different levels of the organization. Customize visualizations for different audiences to ensure that insights are both effectively communicated and best suited to enable informed decision-making.


A Data-Driven Process Example

In order to provide a practical example of the use of a data-driven process, let's consider a topic significant to freight brokers – the management of negative and low-margin shipments.


Step 1: Align analytics to a strategic objective

Imagine the challenging economic landscape where brokers are under heightened pressure and seeing a surge in low-margin or negative-margin shipments. This sets the stage for strategic concern, and it's easy to picture senior management launching an initiative that asks the organization to find innovative ways to lessen the impact of these low-margin shipments (and thereby enhance quarterly results). 


Step 2: Identify broad correlations, trends & opportunities

In a data-driven organization, the first step focuses on utilizing data to gauge the magnitude of the issue and measure the overall ability to affect change. Several questions are asked at the outset, and adaptive analytics work can provide the answer:

  • Is there relevant and accessible data for analytics purposes?

  • How big is the current observed change in negative/low-margin shipments?

  • Are there any observable trends in volumes and margins? What kinds of forecasts can you make about future periods (watch out for seasonality), and where does it seem likely that the numbers would top/bottom out?

At this stage, the data is a compass that guides the organization to set realistic goals for volumes and margins as it undergoes transformative change. If it seems realistic that the organization can only improve the situation by $X dollars, then the cost of the analytics initiative needs to be significantly smaller than X or else the entire project itself may not be worth the effort.

Step 3: Experimental design

Assuming that the initial analysis indicates that it makes sense to proceed, the next step involves defining what changes the business will make (e.g., change pricing) and defining one or more specific ways to improve the situation. It's important to pay close attention to how metrics will be captured and tracked in the next phase, because the impact of the changes needs to be measured. At this point, a more in-depth discovery analysis and exploration of metric correlations is helpful for suggesting (and qualifying) which specific strategies to try:

  • Do low-margin shipments correlate significantly to specific lanes, carriers, or customers?

  • Are any special escalating costs, perhaps due to new or unforeseen accessorials, contributing to the issue?

  • Are customers just paying less for services?

The historical data holds valuable insights, and it may even hold past instances where similar strategies were implemented, along with their outcomes.


Step 4: Test Your Strategies

You are now ready to implement the strategies you selected in the last step. (Huzzah! Maybe reward the team for all their data-driven work so far?) It's not like you will have endless time to run different experiments, but it's still important to undertake a small sample run at the beginning to make sure you don't have any errors in your new processes. When everything looks good, then move forward to full-scale strategy testing.

Communication takes center-stage here as the organization tries to adapt to new business strategies. If new processes are required, such as tagging shipments with special references to help later analytics tracking, then the business needs constant feedback on how well users are implementing those new required processes. Analytics shown to the data-entry employees is likely going to be delivered differently than anything designed for managers, so understand the audience's need and help them do the best job possible. Remember to recognize and reward those making the effort to change.

At some point, you will have enough testing data so that management can decide which strategy to adopt (e.g., a specific price increase strategy). If nothing else, time will dictate that a decision has to be made and the organization has to act, and the analytics need to be ready when that time comes. You likely don't have the luxury to take additional weeks just for results analysis.

Step 5: Design the Forward Process

Now that a new process has been decided upon, moving forward you should use analytics and AI to ensure that your new process is consistently implemented. The most basic application of analytics here would capture and report on metrics that indicate how successfully the new process is being followed. A more advanced implementation may try to automate the decision process, if possible, so that better decisions are made (more quickly and with less errors).

Also, consider deploying predictive analytics to forecast the essential metrics that provide continued insights into the business's trajectory on this issue. A strategy that works well at time A may no longer work at time B.



It's common to see organizations struggle to define processes that will enable them to act in a data-driven manner. If you commit to the three pillars of a data-driven approach (analytics competency, an experimental mindset, and strategic alignment) then you will become truly data-driven and you will optimize your most important business decisions using high-quality and highly-relevant analytics initiatives.




20 views0 comments


bottom of page