Aggregator crash reshapes Amazon business valuations, and smart sellers are adapting
Marketplace data shows the average revenue of online stores sold has fallen dramatically, from $61,000 in 2020 down to only $27,500 by 2023, a drop of 55 percent. Traditional Amazon aggregators are struggling as rising interest rates push investors away from risk and toward stable, predictable yields, meaning fewer buyers and lower valuations for Amazon-centric businesses.
Software aggregators, by contrast, are thriving on subscription revenue models and lower operational risk. Industry reports describe an unquestionable buyer's market, with sellers facing tougher negotiations.
Advisors recommend several moves to maximize value in this climate: diversify beyond Amazon into Shopify, Walmart, or niche marketplaces; build recurring revenue through subscriptions or consumables; use software to cut overhead; keep financials clean and professionally managed; and invest in defensible branding. For buyers, the downturn is a chance to acquire solid businesses at attractive multiples.
In AI commerce news, Walmart's new shopping assistant is being watched closely as the retailer chases Amazon. Analysts say the chatbot is a step, but not the step, and warn sellers against five dismissive reactions: writing it off as just another chatbot, delaying optimization, assuming Walmart does not matter, waiting for the system to mature, or ignoring the results it surfaces.
The recommended actions are to study how AI is changing search behavior, optimize now for product-related Q&A since Walmart otherwise defaults to customer reviews, treat every AI rollout as an SEO opportunity, move early while competition is thin, and feed Walmart's model strong product metadata, Q&A, and titles.
A free star-rating management tool is circulating among sellers, offering trend analysis on ratings, projections of how many weeks until the next half-star, calculations of how many low reviews must be removed or how many five-star reviews are needed to move a rating, and guidance on which variations to split or merge.
Prompt engineering is getting fresh attention after a summary of Google's prompting course spread through the seller community. Noting that the average AI prompt is just seven words, the guide lays out a five-step framework: state the task clearly, add context about brand and audience, supply reference examples, evaluate the output, and iterate.
When output disappoints, the guide recommends revisiting the framework with more detail, splitting long prompts into steps, rephrasing the request with an analogous task, and introducing constraints on style, length, or tone. Multimodal prompting lets sellers upload product photos or brand assets and ask for matching captions or posts.
Advanced techniques include prompt chaining, breaking a big goal like a six-week campaign into linked tasks; chain-of-thought prompting, asking the AI to explain its reasoning; and tree-of-thought prompting, exploring several angles at once with pros and cons. The course also covers building specialized AI agents with defined expert personas, context, and stop phrases.
In other headlines, Amazon has cut 14,000 managerial positions, Google is reported to hold a semi-secret advantage in AI agents, and analysts are examining how Walmart built the biggest competitive threat Amazon has faced. Third-party sellers now account for 62 percent of all units sold on Amazon, and the UK marketplace posted the largest percentage increase last year, growing 12.7 percent.
