PUBLIC TRACES

AMAZON 2023 · COMPLETED RESEARCH

Inspect the public-trace footprint.

Explore Gijs Overgoor and coauthors’ completed analysis of the Amazon Reviews 2023 archive, collected by McAuley Lab at UC San Diego. Every setting below opens saved research results.

Source: McAuley Lab’s Amazon Reviews 2023 dataset on Hugging Face ↗

Original research · recorded results

SELECTED RESEARCH SCOPE

Verified-purchase five-star reviews

January 1–September 14, 2023 · full-release archive audit

Cohort entry count 10Minimum burst 5 reviews

The complete procedure was run on verified purchases. All shares below use the verified five-star denominator.

L0 · REVIEW FOOTPRINT37.9%

Cohort-authored reviews

5,203,677 reviews

Any five-star review by a cohort member.
L1 · REVIEW FOOTPRINT18.5%

Reviews in cohort bursts

2,536,408 reviews

Cohort-authored reviews with the selected minimum number of cohort reviews on the same product within 28 days before or after the review date.
L2 · REVIEW FOOTPRINT12.2%

Reviews by recurring burst authors

1,680,728 reviews

Reviews in cohort bursts whose author appears in qualifying bursts on at least two distinct products.
OVERALL DENOMINATOR13,742,903

Five-star reviews

Used for the overall, entry-count, and ordering results. Categories use their own denominators.

Jan 1–Sep 14, 2023

L0, L1, and L2 are nested footprints of public review structure. They are not additive and do not estimate confirmed fraud or an individual account’s fraud probability.

COHORT ENTRY COUNT

How the footprint changes.

7 completed settings

The review scope and minimum burst size stay as selected above. Each row shows a saved result for a different cohort entry count.

Review footprint at seven cohort entry counts, within the selected scope and burst size
Entry countL0 shareL1 shareL2 share
277.3%43.0%26.0%
562.1%33.2%20.8%
10Selected37.9%18.5%12.2%
145.1%1.1%1.0%
182.9%0.5%0.5%
202.0%0.3%0.3%
300.4%0.0%0.0%

L0: reviews authored by the cohort. L1: cohort reviews in bursts. L2: cohort-burst reviews by recurring authors. Bars use the same 0–100% scale.

ACROSS PRODUCT CATEGORIES

The footprint by category.

34 categories

Alphabetical categories from the completed analysis. Each row uses its own five-star review denominator, under verified-purchase five-star reviews.

Category footprints at cohort entry count 10 and minimum burst size 5
CategoryFive-star reviewsL0 shareL1 shareL2 share
All Beauty7,57333.2%2,514 reviews6.3%476 reviews4.2%315 reviews
Amazon Fashion17,41332.9%5,734 reviews6.4%1,108 reviews4.3%755 reviews
Appliances73,57832.4%23,820 reviews13.3%9,754 reviews9.4%6,909 reviews
Arts Crafts and Sewing278,84232.3%89,942 reviews6.5%17,989 reviews5.2%14,484 reviews
Automotive637,43928.8%183,281 reviews8.1%51,886 reviews5.6%35,823 reviews
Baby Products136,56036.3%49,507 reviews16.7%22,873 reviews13.2%18,003 reviews
Beauty and Personal Care893,95342.8%382,341 reviews24.5%219,286 reviews16.7%148,847 reviews
Books328,83526.9%88,409 reviews5.7%18,778 reviews4.1%13,374 reviews
CDs and Vinyl33,74826.1%8,819 reviews1.2%415 reviews1.0%342 reviews
Cell Phones and Accessories524,98942.0%220,502 reviews25.9%135,920 reviews15.6%82,002 reviews
Clothing Shoes and Jewelry1,981,72937.9%751,533 reviews19.4%383,526 reviews13.0%257,025 reviews
Digital Music2,61723.0%603 reviews0.3%7 reviews0.3%7 reviews
Electronics1,041,28744.0%458,644 reviews29.0%301,643 reviews16.9%176,105 reviews
Gift Cards4,52980.0%3,625 reviews71.3%3,228 reviews51.6%2,338 reviews
Grocery and Gourmet Food430,14639.7%170,772 reviews18.3%78,725 reviews13.8%59,344 reviews
Handmade Products30,77025.8%7,948 reviews4.1%1,261 reviews2.5%781 reviews
Health and Household850,02845.5%386,870 reviews29.0%246,717 reviews18.8%159,421 reviews
Health and Personal Care5,78934.4%1,992 reviews6.5%375 reviews5.1%293 reviews
Home and Kitchen2,174,52139.4%855,791 reviews19.9%431,638 reviews13.5%293,025 reviews
Industrial and Scientific156,42433.6%52,582 reviews10.9%17,120 reviews8.2%12,904 reviews
Kindle Store292,72333.8%98,984 reviews14.4%42,049 reviews10.7%31,223 reviews
Magazine Subscriptions13256.1%74 reviews0.0%0 reviews0.0%0 reviews
Movies and TV129,40037.2%48,120 reviews12.0%15,578 reviews7.8%10,057 reviews
Musical Instruments72,57428.6%20,759 reviews7.3%5,275 reviews5.3%3,812 reviews
Office Products374,13435.5%132,901 reviews12.5%46,861 reviews9.4%35,045 reviews
Patio Lawn and Garden502,05030.9%154,939 reviews11.2%56,266 reviews7.1%35,737 reviews
Pet Supplies522,50043.9%229,397 reviews27.8%145,477 reviews17.4%91,089 reviews
Software15,04530.5%4,584 reviews20.7%3,107 reviews9.8%1,482 reviews
Sports and Outdoors435,63333.1%144,257 reviews12.3%53,631 reviews7.6%33,060 reviews
Subscription Boxes22526.2%59 reviews0.4%1 reviews0.4%1 reviews
Tools and Home Improvement821,56535.9%294,714 reviews15.2%124,614 reviews10.1%83,079 reviews
Toys and Games471,75633.9%160,149 reviews9.4%44,209 reviews7.6%35,975 reviews
Unknown400,75133.3%133,444 reviews9.8%39,324 reviews6.3%25,187 reviews
Video Games93,64538.5%36,067 reviews18.5%17,291 reviews13.8%12,884 reviews

SEARCH ORDER

Variation across completed orderings.

3 recorded orderings

The research repeated the session ordering three times at cohort entry count 10 and minimum burst size 5. These are the observed ranges across those runs, not confidence intervals.

L0 share37.8603%–37.9004%
L1 share18.4561%–18.4636%
L2 share12.2298%–12.2375%
Inspect each completed ordering
OrderingL0 shareL1 shareL2 share
1 · seed 6100037.8645%18.4561%12.2298%
2 · seed 6200037.9004%18.4636%12.2375%
3 · seed 6300037.8603%18.4591%12.2309%

RESEARCH AND DATA

Explore the findings and the research behind them.

Source archive: McAuley Lab’s Amazon Reviews 2023 on Hugging Face ↗. The release contains reviews from multiple years. The reported audit uses reviews dated January 1–September 14, 2023 across all 34 release categories, including Unknown.

Analysis: Gijs Overgoor and coauthors applied the paper’s method to that archive to calculate the findings shown here. The earlier 2019–2020 campaign dataset is a separate public resource:

Public research dataset ↗

These are the paper’s existing archive results. This site does not accept online contributions or provide account-level historical membership checks. Use the investigator to collect and export a case locally.