I have already followed digital retail operations that treated all consumer actions in the same way: same contestation, same thesis, same rite. The volume was large, the strategy was standardized and the result was predictable, but not in a good way. The loss rate varied wildly from one type of demand to another, and no one had stopped to ask why.
Brazilian e-commerce closed 2025 with around 439 million orders. For 2026, the projection is close to 461 million transactions. With this volume, every tenth of a percentage point of failure represents hundreds of thousands of points of friction with the consumer. Part of it becomes a complaint. Part becomes process. And the difference between the two situations increasingly depends on how the company structures its operation before the conflict reaches the Judiciary.
The problem is not the volume itself. Many companies still treat these actions as if they were all the same. They are not.
Four demand profiles, four different strategies
Consumer litigation in digital retail is organized into at least four distinct groups, each with a specific average cost, loss rate and defense strategy.
The first is that of logistics: missed deadline, wrong product, lost in the carrier. It is the one with the largest volume and, in general, the one with the least legal complexity. The company either has or does not have proof of delivery. The SLA was or was not met. With adequate traceability, defense tends to be more consistent and risk more predictable. Without it, it becomes a lottery.
The second is that of product: hidden defect, manufacturing defect, item that does not correspond to the offer. The CDC establishes joint and several liability among suppliers in the chain, which often includes the retailer even when the origin of the problem is in manufacturing. The evidential complexity and potential cost are greater, and the decision to defend or agree needs to take this into account.
The third is that of relationship: cancellation denied, refund not processed, difficulty in returning, non-compliance with the right of withdrawal. Here, the outcome depends less on declared internal policy and more on proof that the consumer received adequate information and was able to exercise their rights without hindrance. This cluster is the one that most exposes companies with poorly documented service.
The fourth is marketplace: the product was offered by a seller within the platform. The risk depends on the role effectively played by the marketplace in offering, payment, logistics and the relationship with the consumer. The greater the integration into the operation and the trust generated by the purchase, the more difficult it is to maintain that there was mere intermediation.
Treating these four groups with the same standardized response is the equivalent of using the same medicine for four different diseases. It works sometimes. Most of the time, no.
The cost that does not appear in the report
When the legal department presents a litigation report to the board, the number that appears is the number of active cases and, when done well, an estimate of the provision. What rarely appears is the total cost by type of demand, with the distinction between conviction cost, fees, operational management and average settlement value per cluster.
This distinction matters because the decision to defend or agree changes radically from one group to the next. A logistics demand with consistent proof of delivery tends to offer better defense conditions and more controllable financial exposure. Depending on the value and circumstances, it may make sense to defend. A relationship demand in a district with an unfavorable history, without documentation of the return policy, has an expected condemnation cost higher than the settlement cost. Insisting on defense may be financially irrational, but that’s what happens when the strategy is standardized.
In practice, the cost per process can vary significantly depending on the type of demand, the district, the strategy adopted and the quality of the documentation available. A well-informed logistics action can require very different effort and exposure than a relationship demand with no history of service. The difference is not just in the legal thesis. It starts with the operation.
The window that the STJ is opening
There is an ongoing discussion at the STJ, within the scope of Theme 1,396, which will define whether the consumer needs to demonstrate a prior attempt at extrajudicial resolution to be interested in acting in consumer actions of a benefit nature. The public hearing took place in May 2026, with broad participation from exhibitors, and the judgment on the merits does not yet have a set date.
For digital retail, this debate has direct implications for logistics and relationship clusters. These are exactly the types of demands in which a documented service channel, with a traceable response and a deadline met, could resolve the conflict before filing a lawsuit.
Reading the Consumidor.gov.br indicators requires care. According to the platform’s methodology, complaints completed without consumer evaluation are included in the solution index. This does not take away the usefulness of the channel, but it recommends caution before transforming the percentage disclosed into automatic evidence of effective resolution.
Companies that already structure these channels will be better positioned whatever the result of the STJ. Not because the law will require it. But because mathematics already shows that resolving beforehand is cheaper than defending later.
What the operation needs to deliver to legal
Efficient management of digital retail litigation begins outside of legal matters. It starts with the operation.
Delivery traceability, with evidence of acceptance or refusal. History of customer service before the process, with protocol and results recorded. Return policy documented and communicated at the time of purchase. Clear contracts with sellers, which organize information, documentation, logistics and service duties, in addition to providing for the right of return and internal distribution of responsibilities, without restricting consumer rights.
When the legal department receives a case without any of these elements, the result becomes a lottery. When you receive it from everyone, it becomes management.
The difference between the two situations is not just legal. It is operational. And it is precisely in this transition between operations and legal that much of the digital retail litigation begins to be won or lost.
