Project Starfish: What It Is and What It Means for Brands

Amazon’s Project Starfish has been making headlines. If you haven’t heard about it yet, it’s the AI‑driven initiative that surfaces products from external websites within the Amazon Shopping app.

Most of the coverage — including a recent Forbes piece — frames it as being an example of the double standards that exist in Big Tech. i.e. while Amazon blocks other companies from scraping its own catalogue, it’s extracting product data from other retailers’ sites to populate discovery features inside its app. (Forbes)

To be honest, while it’s important to understand the framing so that you can preempt its roll-out and/or changes to how it’s implemented in the future, I’m not particularly interested in debating the politics. Sure the dynamics can be frustrating, but we can’t control them and dwelling on them won’t help our businesses. What we can control is how we respond to, and mitigate, the impact that the symptoms of these dynamics have on our eCommerce operations. That’s the focus of this article.

I want to walk you through what Project Starfish actually is, and what it could mean for brands. Its impact isn’t restricted geographically, so I’ve made sure that this discussion is relevant to US, UK and EU businesses.

By the end of this piece, you’ll:

• understand the facts

• be aware of the potential impact on your business, and

• have a practical plan for protecting revenue, margins, and customer relationships.

What is Amazon’s Project Starfish?

At its core, Project Starfish refers to a set of systems and experiments that Amazon is running to pull (scrape) product data from publicly accessible brand and retailer websites. It then structures that data for use inside Amazon’s shopping experience.

Industry reporting clearly states that the outward expressions of Amazon’s internal data aggregation strategy are consumer‑facing elements like Shop Direct and Buy for Me. Here’s how these features currently behave:

Shop Direct

This feature shows products from other brands in Amazon search results and links customers to the brand’s own website for checkout. Consider it a referral link, if you like. Clicks take users out of Amazon and onto the retailer’s site (About Amazon).

With Shop Direct, while Amazon isn’t capturing the sale, it’s capturing the data. The signals it needs to then understand what customers are looking for, how they’re searching, and which brands they’re most drawn to. All juicy information to then inform Amazon’s decisions around UX, M&A activity, product catalogue expansion etc.

Buy for Me

Buy for Me lets Amazon handle the checkout on behalf of the shopper even when the item isn’t sold on Amazon. The aim here is for Amazon to keep the experience within its app — also completing the transaction for the buyer from the external retailer’s site (About Amazon).

The thing here is that — having guided discoverability and then made the transaction for the brand (yes, even without their say-so) — Amazon takes a cut. Although we don’t yet know how much.

The tension in the marketplace is that brands are not asked to opt in to either of these experiences. Instead, many have found their products appearing automatically, and only later discovered they must request removal if they don’t want to participate (EcommerceBytes).

To rein this in a bit, it’s worth noting that this is a beta experiment in the US. It may broaden in scope, but it isn’t a global product yet (About Amazon). It’s worth keeping an eye on it though, as some brands have reported incorrect product information or listings they never meant to show (Business Insider).

What does Project Starfish mean for brands?

When you strip away the headlines and look at the practical reality, the implications of Project Starfish are about visibility, control, and financial flows.

Potential Benefits

Doom and gloom as it might all seem so far, there is a potential upside — of sorts:

• Expanded reach: Your products may appear to new audiences without having to invest in Amazon listing creation or optimisation. Even brands without specialist Amazon resources can be discovered by shoppers via Amazon search.

• Data structuring: AI and automation may organise product information in ways that make it discoverable within Amazon’s ecosystem. Even if this doesn’t move a lot of units immediately, it can inform how you think about category terms or attribute optimisation across channels.

• Market insight: You can observe how external visibility compares with performance on your own channels, which may uncover customer interest you weren’t previously tracking. I say “external” because we don’t yet know whether some/none of the internal analytics will be shared with brands.

This is all said as an eternal optimist. I’m wide-eyed to the fact that there is no public evidence that AI‑generated or aggregated content consistently improves conversion or discoverability yet. Consider that these benefits are plausible, but theoretical at this stage.

Potential Challenges

This section is somewhat easier to swallow, mainly because they’re yet another manifestation of the ongoing issues that Amazon creates for brands. As ever, it’s always best to be forewarned about the sharper edges. Then you can factor them into your strategy:

• Revenue diversion: If customers use Amazon’s Buy for Me option, orders will be directed through Amazon’s checkout experience rather than your own site. The exact fee and commission structures aren’t yet public, and early reporting suggests that in some test scenarios Amazon doesn’t collect commissions, but this remains unclear (SiliconANGLE). You’ll note that I’ve assumed they will — if not immediately, then certainly at some point. It’s Amazon, after all.

• Control over brand experience: Some brands have reported listings with inaccurate names or pricing pulled from their sites, which can undermine trust and cause operational headaches (Business Insider). This takes us back to Amazon circa. 2010, when the brand risk of listing on Amazon often outweighed the potential sales uplift.

• Operational burden: Checking whether your products are showing up, managing opt‑outs, and tracking any resulting orders requires time and resources that many brands didn’t plan for. Ultimately, if you’ve already chosen your sales channels, your focus is being taken away from actually running your business, your way.

• Customer relationship erosion: If Amazon completes the purchase, you may not receive full customer contact information, making it harder to build lifetime value or repeat business. It’s a fair concern, and it keeps coming up in industry commentary (EcommerceBytes).

What are the financial considerations? How can brands plan for Project Starfish?

For this section, I want you to think like a business owner. Detach from the headlines and the subjective narrative that comes with anything Big Tech. By focusing on measurable real-world scenarios, you can remove a lot of ambiguity and brace yourself for the potential impact of Project Starfish.

This is your four-step to-do list:

1. Audit listings now: Search your brand and key products on Amazon from different devices and locations to see if anything shows up under Shop Direct or Buy for Me. This gives you your baseline.

2. Track orders tied to Amazon referrals: If purchases occur through Buy for Me, capture those orders and track them through your financial reporting. Even a handful of units should be visible in your revenue data if they’re real orders.

3. Compare margins realistically: Rather than guessing Amazon’s cut, calculate your actual net revenue from any Buy for Me orders and compare that to what you would have kept selling directly. This gives you real numbers to act on.

4. Set SKU‑level rules: Decide whether to opt out in full. Or… whether to opt out for specific products (particularly high‑margin items) while accepting visibility for others (e.g. overstock or low‑margin products).

Before you call me out for the fact it’s in beta testing and not yet market-wide, this is as much about being super quick off the mark if/when it does affect your brand. So, if you don’t see any activity for your brand, at least get into the habit of regularly checking. Then, the day you show up, you can run through the next steps.

This approach lets you integrate Starfish‑related activity into your cashflow and forecast models so it’s not a surprise when it becomes business as usual.

What are the global implications? UK, EU, and beyond

A lot of eCommerce developments are limited to US brands — not least Shopify’s push into conversational commerce via OpenAI, Gemini and CoPilot. But in the case of Project Starfish, it’s not just US brands who should care.

If your product pages are publicly accessible — whether from a UK domain, EU host, Shopify, Magento, or another platform — your items could be surfaced via Amazon’s experimental features.

For international brands, this means:

• Products can be visible on Amazon without you signing up for Amazon’s marketplace.

• Orders initiated via Amazon (if Buy for Me applies) could divert margin away from DTC channels if fees apply.

• Regional regulations like GDPR or the EU’s Digital Markets Act could shape how opt‑outs and data use compliance are handled.

Even if Amazon’s features aren’t fully rolled out globally yet, being prepared and informed is the practical decision.

Bottom Line

We can’t label Project Starfish as being inherently good or bad. First, we don’t have enough information about it — or evidence of its impact at scale. Second, it might actually be good for some brands, even if it’s not great for others.

But if you walk away from this with a clear actionable takeaway, it would be to audit where your Amazon listings exist. Then, if you show up, start monitor resulting sales, and model the actual financial impact — then make informed, SKU‑specific decisions rather than reacting to headlines.

If you want help working out how to prepare for tomorrow while strengthening your eCommerce business for today — get in touch. I work as a fractional eCommerce and commercial advisor, helping brands build profitable, resilient commerce systems during growth and change.

You can find me on LinkedIn — or reach out directly.