Editor’s Note: This is the second in a five-part series.

“Data attribution turned pricing from estimates into a strategy.”

During the inflation period from 2020 to 2022, one of the largest distributors in North America faced a tidal wave of cost increases. While their legacy pricing process could measure risk, it lacked the ability to generate structured recommendations and present them in a format the business could confidently act on.

That changed with the configuration of a purpose-built pricing analytics solution.¹

This solution allowed the distributor to reengineer the entire contract pricing process and was purpose-built to meet the distributor’s specific business conditions.

The solution framework was anchored on five key pillars:

1. Break down pricing logic, customer contracts, and supplier cost changes into digitized attributes that can be used as both dependent and independent variables²
2. Create a parameterized rules engine to create formulas from these variables
3. Combine these rules, variables, transactional history and other data inputs
4. Automate the process to transform the combined data to generate comparative scenario-based output³
5. Format the output so that both senior leaders and sales teams can quickly understand and act with confidence

The results were transformative

• Modeled opportunity increased by 200%
  The number of recommendations presented to the sales team doubled, revealing far more margin opportunity than ever before.
• Proposal acceptance increased by 50%
  Recommendations were delivered with full context, including relevant customer contract terms. Sales reps could answer 90% of pricing-related questions instantly, building trust and confidence to adopt and approve recommendations.
• Review effort decreased by 80%
  While customer/SKU level pricing proposals were available, along with all supporting details, proposals were also framed at the contract and customer levels, providing the ability to more quickly (and confidently) make aggregated decisions.
• 100% of inflation and tariff costs were recovered
  For two consecutive years, the pricing framework’s recommendations fully offset the impact of rising costs. Previously, the distributor had only reached 40% at best in their first attempt to recover margin related to an inflation event.
• Total margin recovered: $1 billion
  We estimated the value of this solution to be $500 million over 2-3 years. We sent an email to the client confirming our understanding. They politely corrected our estimate:

The Secret to Success? If I had to choose one factor that made the greatest difference in our success, it would be data attribution.⁴

This was not a simple task.⁵  The pricing team attributed more than 40,000 contracts to discrete, structured data points. Examples of these attributes include:

• Contract type (tiered, volume-based, fixed-term)
• Pass-through definitions
• Escalation clauses
• Category and/or product-specific exceptions
• Price formulas / capping limitations

In total, the team identified over 100 unique customer contract attributes across the contract portfolio. Digitizing the attributes was difficult. Converting those digitized attributes in a way that could be used for modeling initially seemed insurmountable.

The breakthrough came when we created a strawman modeling process, an iterative approach that allowed stakeholders to review, easily validate and quickly refine how attributes were defined and used.⁶
(If you only read one footnote, read the one associated with this paragraph.)

The process began with a small sample set of customers, each with varying contract constraints, paired with a small selection of SKUs carrying different inflation parameters. A few pricing rules were then applied to this input, initiating a transformation process that generated a variety of summarized and detailed outputs.⁷

The detailed output included not only the final price recommendation for each customer and SKU, but also a step-by-step breakdown of how the recommendation was derived. This included a clear and easy-to-understand view of data lineage.⁸ In other words, business leaders had full visibility into how each contract constraint, attribute, and formula contributed to the transformation sequence. This transparency gave the team members managing the purpose-built analytics project complete confidence in how pricing logic would be applied and eventually scaled into the broader pricing model.⁹

Each attribute became part of a broader pricing formula, powering simulations across hundreds of millions of customer-SKU combinations — transforming weeks of work into hours. As a result, the company was able to deliver concise, precise, and explainable pricing recommendations.

An important note: The system did not require perfect data. It succeeded because of structured thinking, repeatable logic, and most importantly, the team that brought it to life. One of the concerns from a couple of the team members was this: “We can’t do everything.”  And they were right. There were about 5% of the contracts that we did not attribute (representing about 10% of the business). We chose progress over perfection. We didn’t chase perfection. We built for impact and managed exceptions the old way — on purpose.

If we had to summarize the reason for our success, it would come down to two concepts: data precision and understanding of technology’s capabilities.

Data Precision built trust.

Trust gave confidence.

Confidence enabled action.

Action optimized results.

The precision of data attribution allowed for complete transparency¹⁰ and strengthened the confidence of both the pricing and sales teams. It transforms what might feel like a random surcharge into a well-reasoned, defensible change. It also allowed for consistent external communication as every recommendation was traceable to its source, giving the sales team the clarity they needed to trust and confidence to act on the recommendations.¹¹

Understanding Technology’s Capabilities allowed the team to imagine new possibilities.

Technology may be the instrument, but it takes people to recognize the capabilities of that instrument. In other words, an instrument doesn’t make music on its own. It takes skilled, imaginative hands to turn potential into an enjoyable performance. And music does not come without many “screeches” (we had two daughters who played the violin!).

The same is true for technology.

We were fortunate to work with individuals who understood that to make advancements, there would be “screeches.”  They took the time to dig into the details. They were not afraid to experiment. They learned and recognized the capability of technology. They trusted the team, grasped the capabilities of the tools, and orchestrated outcomes that no technology could achieve on its own.

There are additional insights from this case study that we will reference throughout this paper.

Data Attribution — The Foundation of a Scalable Model

At the heart of any scalable pricing model is a commonly misunderstood but essential foundation: discrete data attribution.

Discrete Data Attribution
Data attributes are the building blocks that enable repeatable, logic-driven financial models. They allow us to break down complex elements such as pricing, customer agreements, contracts, and product details into measurable variables that can be structured and calculated.

These attributes are more than text descriptors. They are operational variables that drive formulas and algorithms.

In our case study, more than 100 unique attributes were extracted from over 40,000 customer contracts. These attributes became the basis for a modeling framework capable of handling high-volume, high-velocity change. When an inflation event occurred, the system executed hundreds of billions of atomic operations to generate precise output.

If you are a business leader still unsure about the value of data attribution, consider the previous example carefully. It took me years to fully understand its impact. What we accomplished in that case study could never have been done with traditional tools or manual effort. Data attribution was not just about managing data. It was the foundation that made scale possible.

This level of processing was possible because the data was structured in a way the system could interpret and use. Discrete, well-defined attributes allowed the model to move beyond manual analysis.¹² They enabled us to turn complexity into clarity and logic into action.

But data attribution goes beyond structuring information. Once attributes are defined and organized, they become the engine for simulation and strategic planning. A well-designed model can evaluate the best case, the worst case and most likely scenarios across hundreds of thousands of products, tens of thousands of customers, and thousands of contracts. These simulations can often be completed in a matter of hours.

These same attributes also support predictive and prescriptive pricing logic. By combining inputs such as SKU-level cost, contract terms, customer profitability, and transaction history, and applying customized rules, the model evolves from a forecasting tool into a strategic prescriptive analytics engine. It becomes a mechanism for balancing margin targets and risk tolerance, while giving leaders the ability to make faster, more informed decisions.

Data Ontology — Providing Context and Meaning

Data attribution alone is not sufficient. Data attributes must be organized and communicated within a broader framework. This is where data ontology plays a central role.

Data ontology¹³ helps consumers of pricing output understand the relationship between what is presented (the recommendation) and other related variables. For example, if a recommendation is made to adjust a price, the system will show these types of metrics:

• The % change of that item and $ impact of that item to the customer (Ex: item increase of 7% and extended price impact to the customer of $50/year)
• The total items affected, average % change of all items, and overall $ impact on all items to the customer (Ex: 6 items affected, average increase of all items by 5.5% and extended price impact to the customer of $350/year)
• For each customer, the overall impact of proposed changes as a % of customer total (Ex: 3% of the items affected, 10% of the total business affected)

Data ontology connects the dots. It turns recommendations into decisions by showing how each input affects the outcome.

Data Lineage — Trust through Traceability

Data lineage shows the full path from raw input to final recommendation, creating traceable trust. This traceability is critical to gaining confidence in high-stakes scenarios like tariffs. Communicating the source of a pricing decision becomes essential to acceptance.

Too often, tariff communications from suppliers are vague, such as stating that increases range from 4 to 23%. This kind of ambiguity weakens pricing automation. Strong data lineage replaces vague statements with specific, verifiable details. For example: “Item A increased by 6 percent under Tariff Act XYZ, dated MM/DD/YYYY,” supported with documentation.

Consumers of price adjustment recommendations (the sales rep) should be able to trace the proposed price to quickly understand the source data and rules behind the recommendation. Without this, confidence erodes.

Data lineage supports both internal alignment and external credibility by providing audit-ready transparency to understand the data and logic used to create the recommendation. Access to this level of detail builds trust. Trust is essential for turning analytics into action and driving adoption of the system’s output.

 

The Distributors’ Analytics Imperative Series

• Part 1: When Survival Demands an Analytics Revolution
• Part 2: Attribution Breakthrough – Building on Bedrock
• Stay tuned for parts 3, 4, and 5!

 

Footnotes

1. We will refer to this solution as: Precision Pricing Optimization. ↩︎
2. The word “analytics” comes from the Greek term Ἀναλυτικὰ, meaning “to loosen” or “break apart.” This definition offers a more accurate way to think about analytics. Today, many associate “analytics” with output: a dashboard, a spreadsheet, a report. But in practice, effective analytics begins with decomposition — breaking complex problems into detailed (attributed) components. This is especially valuable when it comes to pricing events. Whether it’s attributing contract language or interpreting tariff effects into discrete data elements, meaningful analysis starts by structuring data that can be used in formulas, modeled, and acted upon. ↩︎
3. One of the most powerful aspects of this framework was to allow the business to see comparative scenarios. This allows leaders to understand and balance the three critical aspects of any pricing event: Return (benefit to the company), risk (the magnitude of the impact to the customer) and administrative burden (the level of effort of how the recommendations — the prescriptive output — gets actioned). ↩︎
4. Fortunately, with the advent of generative AI, this task — one that took months (80% of the timeline to build this solution was dedicated to this task) — could now be accomplished in days. ↩︎
5. Fortunately, with the advent of generative AI, this task — one that took months (80% of the timeline to build this solution was dedicated to this task) — could now be accomplished in days. ↩︎
6. And looking back, if we had not done this, the skepticism discussed in the next section would have claimed victory, and this project would have been scuttled. ↩︎
7. An important factor to our success is that we started small and evolved. ↩︎
8. This concept is not well known, but critical for adoption. As the term implies, when someone views information (summarized data), often there is the question: “Where did this data come from?” Data lineage provides the consumer of the information an easy way to see and understand how this information was constructed. This is critically important at all levels of any successful analytics effort. Data lineage provides not only visibility but understanding. That understanding instills confidence. Confidence strengthens courage to act. ↩︎
9. This transparency, with the technique data lineage, gave the distributor confidence. Not only did this help with moving the model along but would be another critical success factor once the recommendations of the model fell into the hands of the sales team. ↩︎
10. Transparency is so important to achieve adoption success with a pricing event. To contrast this, let’s look at another distributor who attempted to pass through transportation costs using a single invoice line labeled “Surcharge.” The result? Confusion, mistrust, and customer pushback. Even though the cost increase was contractually valid, without detail, credibility was lost. Tariff-related communication must do more than justify a price — it must provide evidence, and ideally an alternate path to minimize the customer impact. ↩︎
11. While some distributors seek to simplify this process by applying a flat markup across all tariff-affected SKUs, they fail to understand that they introduce new forms of risk. This methodology, while much easier to implement, often leads to overcharging some customers and undercharging others. More critically, it erodes trust — internally with the sales team and externally with customers. ↩︎
12. To clarify: The process was a series of disparate scripts and hard-coded queries that required manual oversight. ↩︎
13. The definition of data ontology can be summarized into one word: Context. Context provides meaning. Meaning provides understanding. Understanding provides confidence to act. ↩︎