HN Debrief

The Amazon tax

  • Economics
  • Startups
  • Infrastructure
  • Regulation
  • Consumer Internet

Seth Godin’s post argues that Amazon’s search ads are not just annoying merchandising. They function like a platform toll. Sellers end up paying to appear in front of customers who were already looking for them, competitors can buy placement against branded searches, and the resulting ad costs get pushed back into product prices. He frames that as a hidden tax on both merchants and buyers, with the added problem that Amazon now has a direct incentive to keep organic search just bad enough to sustain the ad machine.

Treat big marketplace search as adversarial, not neutral. If your business depends on Amazon, Google, app stores, or similar platforms, budget for defensive ad spend and build discovery channels you control because the platform will eventually tax any purchase intent it can see.

Discussion mood

Strongly negative. Most people see Amazon search as deliberately degraded, ad-saturated, and hostile to exact intent, with broader frustration that the same rent-extraction pattern now shows up across search, maps, app stores, and other platforms.

Key insights

  1. 01

    Brand bidding is defensive, not incremental

    Buying ads on your own product name often looks high-converting because it captures customers who already intended to buy. That does not mean it creates new demand. Sellers still pay because competitors can bid on the same term and siphon traffic if they leave the slot open. This changes the economics from marketing into toll collection. Once that norm is established, every seller is forced into a defensive budget just to hold ground they had already earned organically.

    If you run marketplace ads, separate branded and non-branded spend and measure incrementality, not raw conversion. High return on ad spend on your own name can hide a protection racket rather than a growth channel.

      Attribution:
    • throwaway2037 #1
    • levocardia #1
    • cortesoft #1
    • jitbit #1
  2. 02

    Intent capture is breaking exact-match search

    People described systems failing at the one job users still expect from search, which is exact retrieval. A full doctor name and street address in Google Maps produced a sponsored competitor. Amazon searches for exact book titles or model numbers got replaced by unrelated products or by the new "Alexa for Shopping" layer insisting the item did not exist. That is a deeper failure than clutter. It means the platform is willing to override explicit user intent when monetization or recommendation layers think they know better.

    Do not assume exact strings will stay exact on commercial platforms. For critical workflows, link directly to canonical product pages, use raw addresses, and test the actual user path instead of trusting search boxes.

      Attribution:
    • ryandvm #1
    • qwerpy #1
    • QuantumGood #1
    • jeromechoo #1
  3. 03

    Amazon removed power-user search on purpose

    Several comments pointed out that Amazon used to support search operators like quoted phrases and term exclusion and then quietly stripped them away. That matters because it undercuts the idea that semantic search merely replaced older methods with something better. Amazon had tools that would let users narrow results precisely and chose to weaken them, which makes more room for vague matching, sponsored substitution, and profitable reranking.

    When a platform removes precision controls, read it as a business decision, not a product accident. Build your own catalog filters or external search paths if exact findability matters to your customers.

      Attribution:
    • mrandish #1
    • GuB-42 #1
    • marginalx #1
  4. 04

    The bigger problem may be trust collapse

    A lot of people thought ads were only one layer of the problem. Fake reviews, counterfeit goods, listing manipulation, and swarms of near-identical no-name products have made Amazon bad at trust, not just discovery. In that environment, better placement does not help users find better products. It helps them find whatever seller is best at gaming the system. That also weakens the argument that Amazon can simply optimize toward best-reviewed and least-returned items, because those signals are already polluted.

    If you sell quality goods, assume marketplace reputation signals are weaker than they look. Invest in off-platform trust signals like brand recognition, independent reviews, and direct channels instead of relying on marketplace ranking alone.

      Attribution:
    • RobotToaster #1
    • everybodyknows #1
    • rokkamokka #1
    • Brendinooo #1
  5. 05

    Users are already routing around sponsored results

    People shared simple tactics that materially improve the experience today. Sorting by Best Sellers instead of Featured removes many sponsored placements. uBlock Origin strips sponsored items automatically. Others search via Google or external review sites first, then click into Amazon only for fulfillment. The important point is behavioral. Users who care are no longer accepting the native shopping flow. They are building their own search stack on top of Amazon.

    Watch for users bypassing your platform’s intended path. If customers need ad blockers, external search, or third-party review sites to use your service safely, you have already damaged trust.

      Attribution:
    • 83457 #1
    • weberer #1
    • stickfigure #1
    • adamsmark #1
  6. 06

    Cold-start discovery is real but ads are a blunt fix

    The strongest defense of sponsored placement was that new products need some way to get initial visibility. Without that, old incumbents can lock in rankings forever. But commenters also noted this is a solved ranking problem. Marketplaces can reserve exploration slots, weigh recency, and test promising products without turning search into a pay-to-win channel. The existence of a real cold-start problem does not justify pages of sponsored clutter or branded keyword hijacking.

    If you run a marketplace, separate exploration from monetization. Use explicit new-product testing or recommendation slots instead of letting paid placement become the default answer to discovery.

      Attribution:
    • bko #1
    • Arainach #1
    • robotresearcher #1

Against the grain

  1. 01

    There is no single best product

    A credible minority pushed back on the article’s core simplification that Amazon should just surface the "best" item. In many categories, buyers value different things. A cheaper bare-bones option, a familiar brand, or a feature-rich premium model can all be the right answer for different people. That means some amount of merchandising or recommendation near the point of sale is not inherently anti-customer. The real failure is when paid placement masquerades as relevance rather than helping users navigate those tradeoffs.

    Be careful with product ranking models that assume one global winner. Good commerce UX needs clear segmentation by use case, budget, and constraints, not just one default bestseller.

      Attribution:
    • aeturnum #1
    • its_ethan #1
    • senordevnyc #1
  2. 02

    Consumers still have agency to ignore ads

    A few commenters argued the outrage goes too far because sponsored results are labeled and shoppers can sort, filter, compare, or simply keep scrolling. In that view, the deeper problem is that many buyers outsource judgment to the first thing shown, then blame the platform for exploiting that laziness. This does not excuse dark patterns, but it does challenge the idea that platforms alone determine outcomes.

    Even on hostile platforms, interface literacy matters. Teams building consumer products should assume many users will not scrutinize rankings, then decide whether they want to exploit that habit or protect users from it.

      Attribution:
    • whall6 #1
    • falcor84 #1
  3. 03

    This is normal advertising, not unique Amazon abuse

    Some commenters thought the post was rhetorically overheated because competitor bidding and paid placement have existed for decades in search and in physical retail. Brands buy shelf space, publishers pay for co-op placement, and Google has long allowed competitors to bid on branded terms. From that angle, Amazon is not inventing a new scam so much as applying an old one more aggressively in a digital channel where the manipulation is easier to scale.

    Do not mistake familiarity for harmlessness, but also do not frame this as an Amazon-only edge case. If your business depends on any demand aggregator, expect the same extraction logic to show up there too.

      Attribution:
    • noir_lord #1
    • iririririr #1
    • xyzelement #1

In plain english

site:amazon.com
A search operator used in Google and some other search engines to limit results to pages from Amazon’s website.
uBlock Origin
A browser extension that blocks ads, trackers, and unwanted page elements.

Reference links

Regulation and legal context

Tools and workarounds

Alternative search and shopping services

  • Geizhals price comparison
    Recommended as a strong price-comparison and filtering tool in Germany.
  • Kagi
    Mentioned as a paid search engine that avoids ads and feels closer to older search quality.
  • Amazon Haul storefront
    Cited as evidence of Amazon moving further toward low-end marketplace rent extraction.

Background reading and commentary