AI is changing how small online sellers decide what to make
Foto: MIT Tech Review
More than 40% of sellers on the Etsy platform declare that their e-commerce business is their primary source of income, and now their success is increasingly determined by predictive analytics algorithms. Small entrepreneurs are moving away from intuitive design in favor of precise trend analysis supported by artificial intelligence. Tools such as Everbee or Alura allow them to scan thousands of competitor listings in seconds, identifying niche keywords and products with the highest sales potential. For creative makers, this represents a fundamental shift in the business model: instead of creating unique handicrafts and hoping for market interest, they first analyze data from Google Trends and social media before starting production. AI shortens the research process from several days to just a few minutes, allowing for an immediate response to viral trends. The practical implications are clear—the barrier to entry in e-commerce is dropping drastically, but at the same time, pressure to optimize for search engine algorithms is increasing. In global online trade, an advantage is gained not only by artists but primarily by analysts capable of turning data from generative AI tools into a ready-to-ship product. The democratization of advanced analytics means that small, one-person brands can now compete with major retail chains in terms of offering relevance.
In the e-commerce world, where millions of small sellers fight for customer attention on platforms like Amazon or Shopify, traditional market research methods are becoming a thing of the past. For years, the decision-making process regarding what is worth producing relied on intuition, tedious browsing of spreadsheets, and analysis of historical sales trends. Today, this role is being taken over by artificial intelligence, which not only speeds up processes but is also able to spot market niches invisible to the human eye. The story of Mike McClary and his outdoor brand shows that AI is becoming a key advisor in the process of reanimating forgotten products and optimizing offerings for specific customer expectations.
McClary enjoyed success for years selling the Guardian LTE Flashlight – a robust, high-power black flashlight that became one of his best-selling products. Even though he withdrew it from the range around 2017, loyal customers never stopped sending inquiries about the possibility of purchasing it again. What once required manual sorting of hundreds of emails and subjective business risk assessment is today the domain of algorithms that can precisely indicate whether a given product has a chance for a second life in changed market realities.
Algorithmic sentiment analysis and the return of a legend
The use of artificial intelligence in retail allows small players to operate with a precision that was previously reserved exclusively for giants with massive budgets for focus group research. When Mike McClary faced the decision to resume production of the Guardian LTE Flashlight, he did not rely solely on nostalgia. Modern AI tools allow for the aggregation of data from multiple sources: from social media opinions and search engine queries to analyzing the content of customer emails. Natural Language Processing (NLP) algorithms can extract specific product features that users valued most from thousands of messages – in this case, brightness and durability.
Read also
Thanks to AI, sellers can create "ideal product" profiles based on negative reviews of the competition. These tools scan comment sections under rivals' products, identifying recurring complaints. If customers complain about battery life being too short in other brands' flashlights, AI suggests to McClary that the new version of the Guardian LTE Flashlight should have an improved cell. This is a reversal of the traditional design model: instead of creating a product first and then looking for a market, sellers use AI to map "holes" in the market that the product is meant to fill.
- Real-time trend analysis instead of quarterly reports.
- Automatic detection of gaps in the competition's offering.
- Personalization of technical specifications based on specific wishes reported by the community.
- Optimization of prototyping costs thanks to demand simulations.
Artificial intelligence as a designer and strategist
The role of AI does not end with the analysis of historical data. Generative tools are beginning to participate in the visual and technical design process itself. Small brands are using AI to generate product renders before they even hit the production line. This allows for A/B testing on advertising graphics to see which color version or button layout generates the most interest. In the case of the Guardian LTE Flashlight, the return to the classic black design could be backed by hard data indicating that this specific aesthetic is associated by customers with "heavy-duty" reliability.
Modern sellers operating on global platforms must face immense competition, where time-to-market determines survival. AI shortens this time by automating the creation of SEO-optimized product descriptions and generating keywords that convert best in a given category. For a small outdoor company, this means the ability to compete with corporations without having to hire an army of analysts and copywriters. This is the democratization of access to advanced analytics, which is changing the balance of power in electronic commerce.
Challenges and limitations of digital advisory
Despite the enormous possibilities, relying solely on algorithms carries risks. AI is based on past data, which can lead to the creation of "trend bubbles." If every seller uses the same tools for market analysis, the market could be flooded with identical products, differing only by their logo. Mike McClary, in deciding to bring back the Guardian LTE Flashlight, had to balance suggestions from data with the authenticity of his brand. Over-optimization for Amazon algorithms can strip a product of the unique character that originally attracted customers.
"Artificial intelligence will not replace an entrepreneur's intuition, but it will allow them to make decisions with a much lower risk of error, turning guesswork into a precise product strategy."
Another aspect is the issue of market saturation. AI may indicate that heavy-duty flashlights are in demand, but it does not always account for sudden changes in supply chains or logistical costs that could make production unprofitable. Therefore, experienced players treat AI as one of many instruments in their arsenal, rather than the sole oracle. The key remains understanding that this technology works best at identifying patterns that a human cannot perceive in the information noise, but the ultimate responsibility for the product and its quality rests with the human.
Democratization of production in the age of algorithms
We are witnessing a fascinating moment in the history of trade, where the barrier to entry into the manufacturing industry is dropping drastically. Thanks to AI, a small seller can manage a product's life cycle with the same precision as a global retail chain. The case of Mike McClary is proof that even products withdrawn from the market can return in style if their reanimation is backed by intelligent analysis of consumer needs. The Guardian LTE Flashlight is not just a flashlight – it is a symbol of a new era where data becomes a more important raw material than the aluminum used for the device's housing.
In the coming years, we will witness even deeper integration of AI with inventory management systems and "on-demand" production. The ability to predict when a given trend will fade will help avoid overproduction and waste, which is crucial for smaller brands with limited working capital. Artificial intelligence is changing the definition of a "small seller" – today, it's not the size of the warehouse that counts, but the ability to quickly process information and adapt to the dynamically changing expectations of a global audience. The future of commerce belongs to those who can combine the passion for creating physical objects with the computing power of the cloud.
