A mass consumer customer had a problem that, viewed from the inventory report, made no sense: a full warehouse, and at the same time, constant shortages at the point of sale. Product left over on one side of the chain and missing on the other at the same time. The...
The initial diagnosis pointed to logistics: poorly planned routes, inconsistent delivery times, and excessive safety stock. None of these diagnoses were false. None of them was the root cause either.
The root cause was simpler, and more uncomfortable: no one in the organization saw, in the same place and at the same time, how much product was entering the channel (Sell In) and how much was actually being sold at the point of sale (Sell Out). One team saw Sell In. Another saw Sell Out, weeks later, in a different report, with a slightly different product definition. By the time someone crossed both numbers manually, the decision they had had to make two weeks earlier no longer served.
The problem wasn't the inventory. It was the fragmentation of the data.
Sell In and Sell Out are not just two metrics of retail. They are two different business questions that, without joint visibility, are answered separately and almost always incorrectly. Sell In answers «how much are we pushing into the channel?». Sell Out answers «how much is really moving?». When those two questions live in systems that don’t talk to each other, the organization ends up optimizing the one that can be seen, not the one that matters.
In this case, Sell In lived in the distribution system. Sell Out lived in reports from the own retail outlets, loaded with delays and without a standardized definition of what counted as «one unit sold» versus «one unit returned» or «one unit on promotion.» It wasn’t a problem of a lack of data. It was a problem that the data that existed could not communicate with each other with confidence.
Therefore, before touching any business process, the work was about integration and data quality: an ETL (Extract, Transform, Load) process to bring Sell In and Sell Out into a single model, at the same cadence, with the same business rules. And, underneath that, an exercise in MDM (Master Data Management) — a single master data of product and customer, because without that, the «same unit» could appear with three different names depending on the system that reported it, and any cross-sell-out would not be reliable.
This is the part that almost no one discusses when talking about commercial visibility: you can’t have a reliable dashboard of Sell In/Sell Out if the product and the customer are not unified at the source. A dashboard built on data without governance does not reduce commercial uncertainty. It only gives a prettier interface to the same confusion.
What changed when the data was unified
With Sell In and Sell Out integrated over a reliable master data, the organization was able to see, for the first time in the same place, where the product was being pushed into the channel faster than it was actually being sold, and where the opposite was happening. That is not a technical finding. It is a different business decision: at what point in the sales cycle to replenish first, at which point to stop shipping, and with what lead time.
The business decision that changed wasn’t a campaign or a discount. It was the criteria the team used to decide, week by week, who to serve first. Before, that criterion was based on what was seen from the Sell In side — what was coming out of the warehouse — because that was the most available data. Later, it was based on real Sell Out, which is, in the end, the only metric that matters to the business.
The objections that always appear before starting
The first is that the data already exists: the distribution system already reports Sell In, the point of sale already reports Sell Out. That’s true, almost always. What almost never holds true is that both reports share a product, customer, and period definition that allows them to be cross-referenced with confidence. Confirming that —or correcting it— is the first step, not an optional step that can be skipped to get to the dashboard faster.
The second is time: the commercial team is already operating, already making decisions week by week; stopping that to «fix the data» feels like unnecessary friction in the face of the urgency of day-to-day operations. In this case, the data integration and quality work didn’t stop the commercial operation—it ran in parallel, and the team continued to make decisions with what they had while the unified model was being built. What changed wasn’t the speed of the decision. It was the confidence that the decision was being made with the right data.
And there is a third objection that comes later, not before: once integrated, who owns the master data? In this case, the answer was not «from systems» nor «from the commercial department» separately, but from a single owner with authority to decide how each product and each customer is named, classified, and updated. Without that owner, the master data degrades by default: someone adds a new product with their own nomenclature because they need to report that same week, and months later the same problem returns even though the technical integration still works well on paper. Technology solves the integration once. Governance is what prevents the fragmentation from returning.
What this means for any commercial team
If your organization reports inventory and sales separately, and no one has unified definitions of product and customer across both, you don’t have channel visibility. You have two reports that coexist without talking to each other. The question worth asking, before investing in more dashboards or more reports, is the same one that solved this case: Do the things that enter the channel and the things that are actually sold live in the same data model, with the same rules, at the same time?
If the answer is no, the problem is not logistical or commercial. It is data-related—in terms of integration and quality—and that is where we need to start.
This pattern is not exclusive to mass consumption. It repeats itself in any organization where two areas measure the same things with different definitions and no one has noticed because everyone trusts their own reports: in telecommunications, between what the commercial system invoices and what the billing system invoices; in manufacturing, between what the production report and the quality report on the same lot report; in financial services, between what the transactional system records and what the accounting system recognizes as a single customer. In all cases, the symptom is seen as an operational problem and the initial diagnosis almost always points to the process. The root cause, more often than recognized, is that two parts of the same organization are seeing the same reality with two different data definitions, without anyone reconciling them.
The sequence we recommend before investing in more commercial visibility
Before buying another dashboard or hiring another reporting platform, it is worth following a different order than that followed by most sales teams by habit.
First, fragmentation diagnosis: which areas measure the same with different systems, and with what definition each one. Second, master data: before any integration, a same product and customer catalog, because without that no data cross-check is reliable, no matter how many connectors the platform has. Third, integration (ETL): bringing the relevant data to a single model, with the same cadence and the same business rules. Fourth, the decision: build the view or dashboard only after the three previous steps are resolved, not before.
Investing this order —starting with the dashboard and leaving the quality of the data for later— is the reason why many commercial visibility initiatives end up producing beautiful reports that no one uses to make decisions, because no one fully trusts what they show.
Next step
If your sales team is making decisions with separate Sell In and Sell Out processes, or if you simply don’t trust the data you see to represent what’s really happening in the channel, that’s exactly the kind of diagnosis we do in the Data Intelligence practice at Scanda (Data Quality, Single Customer, MDM) before recommending any technology. Write to us.
Español 
English