Photo By: Jack Lee
American consumers are still spending. That doesn’t necessarily mean they’re comfortable.
U.S. retail sales rose 1.2% in August, their strongest increase since March, while core retail sales climbed 1.4%. Yet the same economic picture includes persistent inflation, rising import costs and weakening sentiment among lower-income households. Economists cited by Reuters warned that continued spending could become increasingly difficult to sustain without further gains in household wealth.
For consumer brands, those signals create an uncomfortable problem.
Companies cannot wait for the economy to settle on a coherent narrative before deciding what to charge for a product, whether to run a promotion or where to put the next marketing dollar. Strong headline spending can coexist with consumers who are increasingly sensitive to price.
The economy can remain ambiguous. The pricing decision cannot.
The Problem With the “Average Consumer”
Aggregate economic data is useful for understanding the direction of the economy. It is less useful for answering a company’s next commercial question.
A national increase in retail sales does not tell a beverage company what will happen if it raises prices by 5%. Falling consumer sentiment does not reveal whether a discount will create incremental demand or simply reduce the price paid by customers who would have purchased anyway.
The further executives move from observing the economy to changing something inside their own business, the more specific the question becomes.
Suppose a competitor raises prices and a brand subsequently sees its own sales increase. That could indicate customers are switching between products. It could also reflect a temporary stockout, a regional demand surge or another factor occurring at the same time.
The observation matters. What caused it matters more. For consumer brands operating on tight margins, misreading that distinction can turn a seemingly rational response into an expensive mistake.
Every Commercial Response Creates Another Question
Pricing illustrates the problem particularly well.
If input costs rise, a company can raise prices to protect margin. But higher prices may reduce volume.
It can respond with promotions to protect that volume, but then executives need to know whether the discount generated new purchases or subsidized existing demand.
A company can increase advertising to support the new price, but higher sales following the campaign do not automatically establish that advertising caused the increase.
Commercial decisions rarely happen in isolation. Pricing affects demand. Promotions affect margin. Marketing influences acquisition. Competitors respond. Inventory constraints can change what customers buy regardless of what the brand intended.
That makes historical playbooks harder to apply when the conditions surrounding them change.
James Sun saw versions of that problem during roughly 15 years working with consumer brands including L’Oréal, Sephora, Mattel, Balmain and Amorepacific. Now CEO of Kapnova, Sun argues that companies need to distinguish between signals that identify a potential opportunity and evidence strong enough to justify putting capital behind it.
The distinction becomes more important when consumer behavior itself is uneven.
AI Can Find More Signals. Brands Still Have to Decide.
Retail and consumer companies are already deploying AI across many of these commercial functions.
Boston Consulting Group and The Consumer Goods Forum found AI being applied across innovation, pricing, assortment, forecasting, replenishment and customer engagement. Yet adoption remains uneven: roughly 75% of CPG respondents in their 2026 survey remained in pilot or exploration mode, while only 18% were scaling significant impact.
The challenge isn’t simply getting more information.
Modern AI systems can monitor competitor prices, reviews, search behavior, social sentiment and internal performance data continuously. That dramatically expands the number of potential opportunities a commercial team can identify.
But a signal still leaves an executive with the difficult part. What should we do about it?
That question can require different kinds of analysis depending on the decision. Forecasting can establish a baseline trajectory. Econometric models can help estimate price sensitivity. Causal inference can examine incrementality. Optimization can evaluate how capital should be allocated under constraints.
The important shift is from using one analytical method for every problem to matching the method to the decision being made.
From Understanding the Consumer to Evaluating the Decision
Kapnova is building around that distinction. The company describes itself as an agentic revenue and profit optimization system for consumer brands. Its AI agents scan internal and external information for potential revenue and profit opportunities, while quantitative methods are used to evaluate the decisions those opportunities create.
That might mean determining whether a promotion is likely to create incremental demand, evaluating the effect of a pricing change on volume and margin, or comparing different allocations of marketing capital.
The goal isn’t to produce another description of what consumers are doing. Consumer companies already have dashboards, forecasts and market research for that.
The harder problem is estimating what happens when the company intervenes.
That distinction could become particularly important in an economy producing contradictory signals. Consumer spending may remain resilient while affordability deteriorates for certain households. Some categories may sustain pricing power while others become more promotion-sensitive. Aggregate demand can remain healthy even as behavior fragments underneath it.
There may be no single consumer story for brands to follow. That means the advantage may increasingly belong to companies capable of evaluating each decision against the customers, products and market conditions actually affected by it.
Because executives don’t get to wait until the economy makes perfect sense. The next pricing decision comes first.
