To get straight to the point: the model of “showing ads to humans to drive purchases” is structurally beginning to break down.

Meta’s personalized ads have stopped being effective in Europe. Amazon has shut out AI agents, while Walmart has welcomed them. When we line up these three pieces of news, one question arises.

“In a world where AI makes purchasing decisions, where should advertising budgets be allocated?”

For small and medium-sized e-commerce businesses, this is not just someone else’s problem. For companies with monthly advertising budgets of 500,000 to 2 million yen, misallocating their budget can lead to immediate losses. Let’s break this down step by step.

—

Why Meta’s Ads Have Stopped Being Effective

Due to the EU’s data regulations (strengthening of GDPR), Meta has been forced to offer “Less-Personalized Ads” to European users. Users can now refuse the use of their personal data, leading to a drop in targeting accuracy.

What has been the result?

• Meta’s advertising revenue in Europe is on a downward trend compared to the previous year.
• The CPA (Customer Acquisition Cost) for advertisers has risen. Some e-commerce operators have reported, “We have seen a 30% drop in acquisitions with the same budget.”
• Small and medium-sized e-commerce businesses are particularly affected. While large companies can continue to advertise for brand recognition, small businesses rely on how much they can acquire per transaction.

What’s important here is that this is not a temporary issue. EU regulations are likely to tighten further in the future. In Japan, amendments to the Personal Information Protection Law are progressing, and a similar trend is likely to follow.

In other words, the shelf life of the model that relies on “targeting through social media ads” is beginning to show signs of expiration.

—

Why Amazon Shut Out AI Agents

On the other hand, Amazon has taken a different approach. It has blocked external AI agents, such as Perplexity’s AI shopping agent and Meta’s Muse, from searching, comparing, and purchasing products on Amazon.

Why? The reason is simple.

For Amazon, AI agents are “entities that skip ads.”

Recall Amazon’s revenue structure. A significant portion of its revenue comes from advertising fees paid by sellers (such as Sponsored Products). In 2024, advertising revenue is expected to exceed $50 billion. If AI agents automatically select “the best products at the lowest prices,” they won’t interact with ads at all. The search results page displaying sponsored products will not even be viewed by AI.

What does this mean for small and medium-sized e-commerce businesses selling on Amazon?

• In the short term, the influx via AI agents is blocked, so traditional Amazon ads still function.
• In the medium to long term, if consumers develop a habit of shopping through AI agents, the number of people searching on Amazon itself may decrease.
• By shutting out AI agents, Amazon risks becoming “an inconvenient platform” for consumers.

Amazon has made the decision to sacrifice consumer convenience to protect its advertising revenue model. This structure is worth remembering.

—

The Rationality Behind Walmart Welcoming AI Agents

In contrast, Walmart has fully embraced AI agents. It has partnered with companies like OpenAI to create a system where AI agents can search for products on Walmart and complete purchases on behalf of users. However, they have implemented safety measures such as spending limits and approval steps before purchases.

The reason Walmart made this decision is clear.

Walmart earns money not from advertising revenue, but from product sales themselves.

Whether AI agents or humans come, Walmart profits as long as products are sold. In fact, if purchasing frequency increases through AI agents, sales will grow. Because it does not rely on an advertising model, Walmart can welcome AI agents as “new customers” rather than viewing them as “enemies.”

This holds important implications for small and medium-sized e-commerce businesses.

—

Three Things Small and Medium-sized E-commerce Businesses Should Consider Now

Now we get to the main topic. Considering the movements of Meta, Amazon, and Walmart, what should small and medium-sized e-commerce businesses with monthly advertising budgets of 500,000 to 3 million yen do?

1. Create Channels Other Than “Showing Ads to Human Eyes”

In a world where AI agents make purchasing decisions, ad banners and listing ads are predicated on the assumption that “humans will see them.” AI agents do not view banners, and emotional copy does not resonate.

So what works? Structured data.

• Product specifications, prices, inventory information, and review summaries need to be organized in a machine-readable format.
• Information that allows AI agents to determine “this product is optimal” should be available on your website or feed.

This is not technically difficult. Just aligning with Google’s structured data (schema.org) and organizing your product feed can make a significant difference. The cost can be managed at a few tens of thousands of yen per month.

2. Reduce Dependence on Meta Ads and Diversify

If the CPA for Meta ads is rising, you should try different channels with the same budget. Specifically:

• Google P-MAX: AI automatically optimizes delivery channels. It is easier for small and medium-sized businesses to keep CPA low.
• TikTok Shop (if applicable): More content commerce than advertising. Production costs are decreasing.
• Your own LINE or email: Promoting repurchases to existing customers. The CPA can be one-fifth to one-tenth of new customer acquisition.
• Listing on platforms that support AI agents: Walmart Marketplace (in the U.S.) or Shopify’s AI integration features.

The important thing is not to allocate more than 50% of your advertising budget to one channel. Small businesses should diversify to mitigate the risk of losing sales due to sudden changes in platform rules.

3. Design Products That Are “Chosen by AI”

This is the most fundamental point.

When AI agents select products, what criteria do they use? Based on what we currently know:

• Balance of price and quality (review scores, return rates, etc.)
• Completeness of information (specifications, materials, sizes, usage scenarios)
• Stability of inventory (products that frequently go out of stock are harder for AI to recommend)
• Delivery speed and reliability

Here lies an opportunity for small and medium-sized e-commerce businesses. Large companies often have too many products, making it difficult to keep information for each product up to date. For small businesses, even if they handle 100 to 500 products, they can thoroughly structure information for all their products. It is the nimble small businesses that can create products that are “easy for AI agents to choose.”

—

What Happens When the “Destination of Advertising Budgets” Changes

Let’s take a step back and think about this.

Over the past decade, advertising budgets for e-commerce businesses have concentrated on three platforms: Meta (Facebook/Instagram), Google, and Amazon Ads. However, all three are currently in the midst of structural changes.

• Meta: Declining targeting accuracy is worsening advertising efficiency.
• Amazon: By excluding AI agents, it risks future declines in traffic.
• Google: Reports are beginning to emerge that AI Overviews (AI summaries of search results) are lowering click-through rates.

When the destination of advertising budgets changes, it means that the assumption that “money will sell” is collapsing.

So what remains? Product strength and quality of information.

AI agents do not recommend products because they are paid to do so (at least not for now). They purely select the products that best meet user demands. In other words, there is a possibility of shifting from a war of attrition based on advertising budgets to a competition based on the quality of product information.

For small and medium-sized e-commerce businesses, this is not necessarily bad news. Companies that could not compete with large corporations on advertising budgets may find a new arena to compete in through the structuring of product information and adapting to structured data.

—

So, What Should We Do?

Just three things.

① Structure the product data on your website. Adapt to schema.org and organize your product feed. Even if outsourced, it can be done for 100,000 to 300,000 yen. If done yourself, it can be free. This will be the minimum requirement to “be found by AI.”

② Gradually reduce the budget for Meta ads and diversify into 2 to 3 channels. Instead of cutting it all at once, shift 10-20% of the budget to other channels each month and compare CPAs. Make decisions based on the numbers.

③ Try creating one product page that is “chosen by AI agents.” Create a product page that includes specifications, comparison information, review summaries, FAQs—essentially all the information that AI might use to make decisions—and observe any changes in traffic or purchases.

All of these can be started next week. No significant investment is required.

In the era of AI shopping, continuing to spend advertising budgets solely on “human eyes” is a risk. As the destination of advertising shifts from “humans” to “AI,” small and medium-sized e-commerce businesses should act first. While large corporations take three months to make organizational decisions, small businesses can experiment next week. That is their greatest weapon.