Amazon says its shopping recommendations can draw on purchases, searches, reviews, Lists and Alexa for Shopping conversations. That tells you what Amazon publicly describes about its own service—not whether its recommendations are better than another retailer’s. To compare fairly, look at the same kind of recommendation, on the same shopping task, and check what each retailer actually discloses about inputs and customer controls.
What Amazon says informs shopping recommendations
Amazon describes personalization as part of the shopping experience on its homepage and throughout the shopping journey. It says shopping activity informs recommendation types and also describes personalized deals, emails and product descriptions as part of that experience. The recommendation you see may therefore depend on where it appears and what you have been doing on the site or app.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
Amazon eGift Card - Amazon Logo | $50.00 | Buy on Amazon |
| 2 |
|
Amazon eGift Card - Happy Birthday | $50.00 | Buy on Amazon |
| 3 |
|
Amazon eGift Card - Birthday Wishes | $50.00 | Buy on Amazon |
| 4 |
|
Amazon Physical Gift Card in a Gift Box - Better than Gold - Black | $50.00 | Buy on Amazon |
| 5 |
|
Amazon Physical Gift Card in a Mini Envelope - Candlelight Celebration | $50.00 | Buy on Amazon |
As an Amazon Associate I earn from qualifying purchases.
Amazon’s About You announcement identifies several kinds of activity that contribute to shopping personalization:
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →- Purchase history: items you have bought.
- Searches: products or categories you have looked for.
- Product reviews: feedback you have written.
- Lists: products you have saved to lists.
- Alexa for Shopping conversations: shopping-related interactions with Alexa.
These are Amazon’s descriptions of its practices, not an independently audited or exhaustive inventory of every signal, model or rule involved. They also do not establish that every signal influences every recommendation surface. A homepage row, a suggestion on a product page and a shopping response after a conversation may answer different questions.
#1 Best Overall
- Amazon.com Gift Cards never expire and carry no fees.
- Multiple gift card designs and denominations to choose from.
- Redeemable towards millions of items store-wide at Amazon.com or certain affiliated websites.
- Available for immediate delivery. Gift cards sent by email can be scheduled up to a year in advance.
- No returns and no refunds on Gift Cards.
Amazon’s company description says that, based on a customer’s shopping activity, it reviews preferences to create recommendation types on the homepage and throughout the shopping journey. It also says product descriptions may be personalized to be more relevant. That is Amazon’s own characterization of its service; it should not be mistaken for a technical account of every recommendation or an independent assessment of how well a suggestion matches a shopper’s needs.
The company has also described recommendations as extending beyond a single row of products. Personalized deals and emails, for example, are part of the broader experience Amazon describes. A shopper evaluating one homepage shelf should not assume that observation tells them how recommendations work in email, on product pages or in other parts of the journey. Keep the surface clear when discussing what appeared and where.
Recommendation labels do not mean the same thing
A personalized homepage suggestion is not necessarily equivalent to a product shown because it is related to the item currently being viewed. Nor is either one necessarily equivalent to a bestseller list or a response tailored to a conversational request. When comparing retailers, first identify the kind of result in front of you. Otherwise, you may be comparing a personal suggestion on one site with a popularity list or a context-based product match on another.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchAmazon has also described Help Me Decide as an AI feature that can make recommendations after a shopper has browsed several similar products. A company executive characterized it as a way to offer product recommendations tailored to a shopper’s needs. That is a description of the feature’s intent, not independent evidence that it improves decision quality. Availability and presentation can vary by marketplace, app version and shopping surface.
Daniel Lloyd, identified by Amazon as its vice president of personalization, described Help Me Decide this way: “Help Me Decide saves you time by using AI to provide product recommendations tailored to your needs after you’ve been browsing several similar items, giving you confidence in your purchase decision.” The statement explains how Amazon presents the feature’s purpose. It does not establish that shoppers actually save time, feel more confident or make better decisions in an independent test.
In practice, a shopper should note the label and context rather than treating all product suggestions as one category. “Based on your browsing,” “related to this item,” “popular” and a conversational answer can signal different recommendation formats. The label is useful evidence of how a retailer frames a result, but it is not a complete explanation of the system behind it.
What About You and privacy controls can—and cannot—tell you
Amazon describes About You as a place where customers can review and edit details used to personalize shopping recommendations. It can help a shopper understand or adjust preference details, but it should not be treated as a complete view of all information associated with the account or every input that may affect a suggestion. The feature and available controls may differ by marketplace, app version and surface.
Amazon’s privacy notice also discusses recommendation-related information, including recently viewed products that form a basis for recommendations. It describes settings for certain data uses and personalized advertising. Those are not interchangeable controls: changing an advertising setting should not be assumed to disable product recommendations, and a control over browsing or preference information should be understood according to the scope Amazon assigns to it.
Rank #2
- Amazon.com Gift Cards never expire and carry no fees.
- Multiple gift card designs and denominations to choose from.
- Redeemable towards millions of items store-wide at Amazon.com or certain affiliated websites.
- Available for immediate delivery. Gift cards sent by email can be scheduled up to a year in advance.
- No returns and no refunds on Gift Cards.
That distinction matters when judging what a control accomplishes. A setting whose stated purpose is personalized advertising should be described as an advertising setting, not a general switch for every shopping suggestion. Similarly, a page that lets a customer edit preference details is not proof that all recommendation inputs are visible there. Read the setting’s description and scope rather than inferring a broader effect from its name.
A careful way to inspect your preferences
- Find About You: Look for Amazon’s About You feature in the marketplace and app or website you use. Its availability and interface may vary.
- Review the preference details shown: Note what Amazon presents as editable information used for shopping personalization. Do not assume the page exposes every account signal.
- Make only a change you understand: If you edit a preference, record what changed and which recommendation surface you want to evaluate afterward.
- Check the privacy notice for the relevant setting: Distinguish controls for personalized advertising from controls described in relation to product recommendations or other data uses.
- Revisit the same shopping surface: Compare the result after the change, allowing for the possibility that a recommendation may not update immediately or may be affected by other activity.
This is a way to observe a change, not proof of exactly how a recommendation system works. Other activity, account state, product availability and the page being viewed can affect what appears.
For a useful record, write down the marketplace or country, whether you were signed in, the app or website surface, the date, and the preference or setting you changed. That makes a later comparison easier to interpret and avoids treating results from different countries, versions or account states as if they were identical. It still cannot isolate every influence on the result, but it makes your own observation more transparent.
Why AWS Amazon Personalize is not proof of Amazon.com’s implementation
AWS offers Amazon Personalize, a developer service that retailers and other organizations can use to build recommendation systems. Its technical documentation describes systems that can use interaction data and item metadata. Depending on the use case, recommendations can respond to recent recorded events, and filters can exclude items or constrain which results are returned.
That documentation is useful for understanding how a configurable recommender can work. It does not establish that Amazon.com’s own consumer shopping recommendations run on Amazon Personalize, or that Amazon.com uses the same models, signals, filters or implementation details described for the service.
The documentation also explains why a recommendation should not be read as a neutral product-quality judgment. For certain recipes, a score is a relative estimate of the likelihood that a user will interact with an item. Exploration can introduce items that might otherwise be less likely to appear, and filters can alter the returned list. A suggested item is therefore not a guarantee that it is the best product, the most suitable choice or the highest-quality option.
Those technical concepts can help readers understand what recommendation systems may be designed to do, but they should not be transferred to Amazon.com without evidence. For instance, documentation that a developer service supports recent events says that such a service can be configured to react to them; it does not verify that a particular retail page uses that capability. The same caution applies to filters and scores.
A ranking or score is also not the same as a consumer-facing explanation. The service documentation describes relative estimates for particular use cases, not a universal quality grade that tells shoppers whether one product is objectively superior. Exploration may intentionally bring a less-obvious item into a list, while filters may remove otherwise relevant choices. The order and composition of a list can reflect system design and constraints as well as predicted interaction.
Rank #3
- Amazon.com Gift Cards never expire and carry no fees.
- Multiple gift card designs and denominations to choose from.
- Redeemable towards millions of items store-wide at Amazon.com or certain affiliated websites.
- Available for immediate delivery. Gift cards sent by email can be scheduled up to a year in advance.
- No returns and no refunds on Gift Cards.
How to compare Amazon with another online retailer fairly
A meaningful comparison needs evidence about the same dimensions for each retailer, the same country and a clearly identified shopping surface. Without that, an apparent difference may come from the type of recommendation being displayed, the account’s activity or the product category—not from one system being more personalized or effective.
Also record when and where you made the comparison. Features, labels and settings can vary by marketplace, app version and product surface, and retailers can change their experiences over time. A statement about one country’s app at one point should not automatically be generalized to every version of a retailer’s site.
1. Compare disclosed signals, not assumptions
Check whether each retailer says recommendations use purchases, searches, viewed items, lists, reviews or other activity. Keep disclosed facts separate from inference. A retailer’s failure to describe a particular input is not proof that the system does not use it; likewise, a general statement about personalization does not establish the complete set of inputs.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →For each retailer, you can make a short evidence note: the retailer’s own description, the kind of recommendation it refers to, and what remains unspecified. For Amazon, the described inputs include purchase history, Alexa for Shopping conversations, product reviews, Lists and searches. Do not turn that list into a claim that every listed activity affects each recommendation, or into a list of what another retailer must use.
2. Match the recommendation type
Record whether the suggestion is tied to an account, related to the item currently open, drawn from a bestseller or popularity list, or generated in response to a conversational shopping request. These formats serve different purposes. Compare like with like—for example, an account-based homepage row against the equivalent surface elsewhere, rather than against a list labeled as popular.
Write down the displayed label and placement as well as the product. A recommendation on a product detail page may be responding to the current item, while a homepage shelf may be presented in the context of the account’s broader shopping activity. If the retailer does not say which basis applies, mark it as unknown rather than assuming personalization from the mere presence of a suggestion.
3. Check controls and their scope
Look for controls to inspect or edit preference details, remove items from browsing history, or change recommendation settings. Note exactly what each control claims to change. An advertising opt-out is not automatically a recommendation opt-out, and a history control may have a narrower effect than its label suggests.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Controls are comparable only when their stated purposes are comparable. One retailer might expose editable preference information, while another might offer a history control or a setting for a specific data use. Describe those differences directly, and avoid reducing them to a single privacy score unless there is a clearly defined, evidence-based method for doing so.
Rank #4
- Gift Card is redeemable towards millions of items storewide at Amazon.com
- Gift Card has no fees and no expiration date
- Gift Card is nested inside a specialty gift box
- Free One-Day Shipping (where available)
- Scan and redeem any Gift Card with a mobile or tablet device via the Amazon App
4. Test freshness and context
See whether the retailer says recent activity can change suggestions, and whether the claim applies to the surface you are testing. A system may incorporate recent events or show exploratory items, but those possibilities should not be attributed to a specific retailer unless its own documentation supports the claim.
When carrying out a personal comparison, keep the shopping task and the relevant account state as steady as you can. If you browse several products on one site but not another, the different histories may be part of the explanation for different results. If you change an available preference or history control, record that change and return to the same kind of page rather than comparing unrelated sections.
5. Separate recommendation quality from personalization
A more tailored-looking result is not necessarily more accurate or useful. A credible performance comparison would need a defined outcome—such as whether shoppers find a relevant item—and a comparable testing method. Offline metrics can assess recommendation lists against recorded data, while online evaluation can measure behavior in a live experience; randomized A/B testing is one method Amazon recommends for evaluating solutions built with its developer service. None of those service-level evaluation materials is a published head-to-head result for Amazon’s retail site and other retailers.
Amazon Personalize documentation includes measures such as coverage and mean reciprocal rank. Coverage concerns how broadly a recommender includes available items in its recommendations; mean reciprocal rank is a ranking measure that gives weight to how high a relevant item appears. These terms can help explain the kinds of questions developers evaluate, but they are not results showing that Amazon.com performs better than another retailer. Nor does any one metric settle whether suggestions are useful to a particular shopper.
Amazon’s service guidance recommends evaluating with offline and online measures, including A/B testing. In general terms, offline evaluation uses recorded data to assess a solution, while online evaluation examines outcomes in a live experience. A fair comparison across retailers would still need comparable definitions, conditions and outcome measures. The technical guidance is for evaluating solutions built with Amazon Personalize; it does not publish a head-to-head result for Amazon’s retail site versus Walmart, Target, eBay or another named retailer.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.A practical comparison test for game shopping
For a video-game shopper, use one specific task across the retailers you want to compare—for example, finding a game in a genre you already play, or finding a suitable game for a particular platform. The point is not to treat one browsing session as a scientific ranking; it is to make observations consistently and see which evidence is available.
- Choose a task and product context: Keep the genre, platform and shopping goal consistent. Do not compare one retailer’s suggestions for a new release with another’s suggestions for a broad category.
- Use the same account state where possible: Record whether you are signed in and whether you have relevant purchase, search, viewing or list activity. A fresh account and an established account are not equivalent test conditions.
- Visit the same type of surface: Compare homepage recommendations with homepage recommendations, or product-page suggestions with the corresponding product-page format. Note the label and placement of each row.
- Record what the retailer says: Capture the stated basis for the suggestions and any relevant preference, history or recommendation control. Mark anything not disclosed as unknown, rather than inferring it from the results.
- Change one documented input or control: If a retailer provides an editable preference or history control, change one thing and return to the same task and surface. Avoid changing several settings at once if you want to understand which change may be associated with a difference.
- Compare relevance, not just novelty: Note whether suggestions fit the task and whether they repeat products already considered. A changed list shows that the displayed output changed; it does not by itself establish why, or prove that the new list is better.
For a stronger personal comparison, repeat the observation rather than relying on a single visit, and keep the account state, shopping goal and surface as similar as practical. Note whether a suggestion appears to follow recent activity only when the retailer’s own description supports that interpretation. A shopper’s small test can be informative about their experience, but it is not a controlled evaluation of the retailer’s system as a whole.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsFor a robust claim about which retailer performs better, observations would need to be repeated under comparable conditions and evaluated with a defined measure. The available evidence does not establish a sourced ranking of Amazon and other retailers by recommendation accuracy, personalization depth or privacy protection.
Best Value
- Gift Card is redeemable towards millions of items storewide at Amazon.com
- Gift Card has no fees and no expiration date
- Gift Card is affixed inside a mini envelope
- Free One-Day Shipping (where available)
- Scan and redeem any Gift Card with a mobile or tablet device via the Amazon App
What the comparison can establish
Amazon publicly identifies several shopping activities that inform personalization and describes About You as a place to review and edit preference details. Its privacy notice distinguishes recommendation-related information from settings for personalized advertising and other data uses. Those disclosures give shoppers useful questions to ask and controls to inspect.
They do not support a general verdict that Amazon is more personalized, more accurate or more privacy-invasive than other online retailers. A fair comparison requires retailer-specific disclosures and controls for the same country and surface, plus comparable evidence of recommendation quality. Where that evidence is absent, the accurate conclusion is that the systems cannot yet be ranked on those grounds.
The practical takeaway is to treat each recommendation as one retailer’s suggestion in a particular context, not a neutral verdict on product quality. Ask what kind of recommendation it is, what the retailer says may inform it, and what the relevant controls actually cover. Then separate what you observed from what the retailer disclosed—and from anything that remains unknown.
Free tools Windows power users keep installed
One-click scans. No signup required.
Frequently Asked Questions
How does Amazon decide what to recommend?
Amazon says shopping personalization can draw on purchase history, Alexa for Shopping conversations, product reviews, Lists and searches. It also describes shopping activity as informing recommendations across the homepage and shopping journey. Amazon does not present that as an independently audited, exhaustive list of every signal or a guarantee that each signal affects every recommendation surface.
Can I change what Amazon recommends to me?
Amazon describes About You as a place to review and edit preference details used to personalize shopping recommendations. Availability and controls may vary by marketplace, app version and surface, and About You should not be treated as a complete view of all account information or recommendation inputs.
Does turning off personalized advertising turn off product recommendations?
Not necessarily. Amazon’s privacy notice describes settings for certain data uses and personalized advertising separately from recommendation-related information. Check the stated scope of a setting; an advertising control should not be assumed to disable product recommendations.
Does Amazon Personalize power Amazon.com’s recommendations?
The AWS documentation describes Amazon Personalize as a developer service for building recommendation systems. It does not establish that Amazon.com’s own retail recommendations use that service or the same implementation.
Recommended Free Tools
Are Amazon’s recommendations better than those from other retailers?
The available evidence does not establish a comparable ranking of Amazon and other retailers for accuracy, personalization depth or privacy. A fair quality comparison would require retailer-specific evidence, comparable conditions and a defined outcome measure; a personalized label alone is not evidence of better recommendations.
The Bottom Line
Amazon discloses several activities that can inform shopping personalization and offers About You for reviewing and editing preference details, but those disclosures do not show that its recommendations outperform other retailers. Compare the same recommendation type and shopping task, check each retailer’s controls and stated scope, and treat undisclosed inputs and untested quality claims as unknown.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
