How AI Helps Small Online Sellers Decide What to Make

MIT Technology Review has described how small online sellers are using artificial intelligence to decide which products to make next, a question that has for years been one of the most expensive and risky for small brands. The article starts with the story of Mike McClary, who for years sold the Guardian LTE Flashlight through his small outdoor brand. It was a rugged model in a black housing, designed for brightness and durability, and over time it became one of the most popular items in his lineup.

The Story of One Flashlight

According to the publication, McClary stopped selling the Guardian LTE Flashlight around 2017, but that did not end interest in the product: customers kept emailing him to ask where they could now buy the flashlight. The situation is a classic example of a problem familiar to almost any small manufacturer: demand for a specific product often becomes obvious only in hindsight, after it has been discontinued, rather than at the moment of deciding what to put on sale.

Key facts from the story:

• The product is the Guardian LTE Flashlight, a heavy, rugged model in a black housing.
• It was sold through Mike McClary's small outdoor brand.
• It became one of the most popular products in the brand's lineup.
• Sales were discontinued around 2017.
• Customers kept writing with questions about buying it even after it was taken off sale.

How AI Helps Decide What to Make

For small online sellers, the main constraints have traditionally been money and time: full market research, analysis of competitor reviews, prototype testing and audience surveys were a luxury available mainly to large companies with dedicated marketing and product teams. AI-based tools change that balance of power. They can analyze large volumes of customer emails, reviews, social media discussions and search queries, and surface demand patterns that used to take an analyst months of manual work. Such signals, from repeated customer questions about a discontinued product to trends in adjacent categories, become the basis for better-grounded decisions about which version of a product to launch next.

Why This Matters for Small Businesses

For the owner of a small brand like the one in the flashlight story, every decision to launch a new product is a real financial risk: buying a batch, manufacturing, warehousing, marketing. A wrong demand forecast can mean money frozen in unsold stock, while a missed signal of real customer interest means lost revenue that goes to competitors instead. By lowering the cost of quality market analysis, AI levels the playing field between retail giants with staffs of analysts and owners of small niche brands, who have historically made such decisions more by intuition than by data.

The Guardian LTE story is also telling because customer emails have always been an accessible source of information about demand. The problem was not a lack of data but that a small team physically lacked the time and resources to analyze it systematically and turn it into concrete product decisions. This routine but labor-intensive work, reading and categorizing thousands of emails, reviews and online mentions and picking out recurring requests from the noise, is what modern AI tools can take over. The business owner keeps the final decision but no longer has to review every message by hand to spot a pattern that might previously have gone unnoticed until it was too late.

Another layer of opportunity opens up where demand analysis meets generative design tools. Instead of merely noting that customers want a discontinued product back, a small brand can test hypotheses about product variations, such as color, material and configuration, faster, before investing in a trial production run. For companies at the level of Mike McClary and his outdoor brand, this fundamentally changes the speed of iteration between an idea and a launch decision, narrowing a gap that used to be accessible only to large players with full market research departments.