A WORKED INVESTIGATION

What does repeated overlap look like?

Start with one reviewer’s history. Follow nearby five-star reviews of the same products. Use the connections you find to choose whose history to examine next.

Fictional people and products

ONE REVIEWER, FOLLOWED THROUGH THE RECORDS

6 reviewers connect back to
Rowan Hayes.

We examined other reviewers’ histories and found five-star reviews of the same products, posted close in time to this reviewer’s own five-star reviews. The pattern appears across 2 shared products in this generated case.

2products with supporting overlap

3–12 daysbetween the paired five-star reviews

The search examined 10 reviewer histories and found 71 reviewers through their neighborhoods. The records below explain this reviewer’s result. Names and product descriptions are fictional labels for the generated records.

How were these reviewers found? ↓Inspect the supporting products ↓

Start with this reviewer’s history 5 generated reviews across 4 products

These are all of this reviewer’s records in the generated panel. The 2 products with supporting overlap are a subset of this 4-product history.

All generated review records for Rowan Hayes
Product	Review date	Rating	Verified purchase
Glass storage jars	2019-03-26	5 stars	Marked
Cotton throw blanket	2019-04-03	5 stars	Marked
Adjustable desk lamp	2019-04-11	5 stars	Marked
Magnetic kitchen timer	2019-06-23	5 stars	Not marked
Cotton throw blanket	2019-06-24	5 stars	Marked

HOW THE SEARCH GETS HERE

Each discovery helps choose the next reviewer.

The search follows the connections it uncovers. A reviewer who keeps appearing around histories already examined becomes a stronger lead for the next step. This is an adaptive search: the evidence found so far guides where to look next.

Examine one history
This run starts with Riley Bennett, chosen at random from reviewers with five-star activity. Examining a history means looking through that reviewer’s available review records: products, dates, and ratings.
Find nearby reviewers
For each five-star review, look for other people who gave the same product five stars within 14 days before or after. Together, those reviewers form the history’s neighborhood, or ego.
Update the counts
Each eligible history adds +1 corroboration to every other reviewer in its neighborhood. Several matching products with that same history still add only one. Appearing around another examined history adds another.
Follow the next lead
Among reviewers not yet examined, choose the highest current count within the search cap of 5. Examine that history, find its neighbors, and update the counts again. This run repeats until 10 histories have been examined.

Why examine that person next? Their repeated appearances provide an evidence-based lead. Their history can reveal further connections and add support to reviewers already found.

A simple choice example, and how the cap works

For illustration, suppose two reviewers have not yet been examined: one appears in three neighborhoods found so far and the other in one. With this cap of 5, the search chooses the reviewer with three corroborations first, unless another candidate has a higher count within the cap.

The cap controls which lead to follow: reviewers with a current count above 5 are set aside for this choice. Their evidence still counts, and their reported corroboration can keep growing beyond 5. If no unexamined candidate has a positive count within the cap, the search chooses at random from all remaining unexamined histories, including those above the cap. Ties between candidates are resolved by the order they first appeared in the evidence.

“Stronger lead” means more corroborations under this selection rule. It is not a probability of coordination or fraud. The count used to choose a reviewer is the count at that step; the final count shown in a result may be higher. A history needs at least two five-star reviews to contribute an eligible ego. This is checked after examining it, so a history without an eligible ego still uses one step of the budget.

FOLLOW THE EVIDENCE

Same products. Nearby five-star reviews.

Open a product to compare the dates and records. A pair qualifies when both reviews give five stars and their dates are no more than 14 days apart, including exactly 14 days.

SHARED PRODUCTAdjustable desk lamp3 supporting histories9–12 days apartView all paired recordsOne example pair · both reviews give five starsCameron Reedhistory examined2019-03-3012 days apartRowan Hayesthis reviewer2019-04-11

Rowan Hayes gave five stars on 2019-04-11. Each row below pairs one of those reviews with another reviewer’s five-star review of this product. We examined each of those reviewers’ histories during the search.

All five-star review pairs supporting Rowan Hayes on Adjustable desk lamp
Reviewer whose history we examined	Their review date	Rowan Hayes’s date	Rowan Hayes’s timing	Verified purchase marks
Cameron Reed	2019-03-305 stars	2019-04-115 stars	12 days after	Cameron Reed: markedRowan Hayes: marked
Elliot Palmer	2019-04-015 stars	2019-04-115 stars	10 days after	Elliot Palmer: markedRowan Hayes: marked
Riley Ellis	2019-04-025 stars	2019-04-115 stars	9 days after	Riley Ellis: markedRowan Hayes: marked

“Before” and “after” describe Rowan Hayes’s review relative to the paired reviewer’s review. Purchase marks are generated fields, as are the dates and identifiers.

SHARED PRODUCTCotton throw blanket3 supporting histories3–7 days apartView all paired recordsOne example pair · both reviews give five starsRowan Hayesthis reviewer2019-04-034 days apartRiley Bennetthistory examined2019-04-07

Rowan Hayes gave five stars on 2019-04-03. Each row below pairs one of those reviews with another reviewer’s five-star review of this product. We examined each of those reviewers’ histories during the search.

All five-star review pairs supporting Rowan Hayes on Cotton throw blanket
Reviewer whose history we examined	Their review date	Rowan Hayes’s date	Rowan Hayes’s timing	Verified purchase marks
Avery Foster	2019-03-275 stars	2019-04-035 stars	7 days after	Avery Foster: markedRowan Hayes: marked
Sage Foster	2019-04-065 stars	2019-04-035 stars	3 days before	Sage Foster: not markedRowan Hayes: marked
Riley Bennett	2019-04-075 stars	2019-04-035 stars	4 days before	Riley Bennett: markedRowan Hayes: marked

“Before” and “after” describe Rowan Hayes’s review relative to the paired reviewer’s review. Purchase marks are generated fields, as are the dates and identifiers.

The same reviewer’s history can support more than one product. Count each supporting history once across all products; adding the product totals can count a history twice.

Account for all 6 corroborations Every supporting history and its shared products

An ego is the neighborhood around one reviewer whose history we examined: other reviewers giving five stars to the same products within ±14 days. Each history below has at least two five-star reviews and contributes one ego to this reviewer’s count. The date pairs above show why.

Distinct histories supporting Rowan Hayes’s corroboration count
Reviewer whose history we examined	Five-star reviews in its history	Shared products	Count added
Riley Bennett	3	Cotton throw blanket	+1 ego
Avery Foster	2	Cotton throw blanket	+1 ego
Sage Foster	2	Cotton throw blanket	+1 ego
Cameron Reed	4	Adjustable desk lamp	+1 ego
Elliot Palmer	9	Adjustable desk lamp	+1 ego
Riley Ellis	2	Adjustable desk lamp	+1 ego
Distinct supporting histories			6

READ THE RESULT WITH ITS LIMITS

What is observed, and what is missing?

Observed in this example

6 supporting histories, 2 shared products, and the paired five-star dates shown above. The method has all 610 generated records available; 10 histories were selected for the search.

No live source check

No marketplace page was captured. These fictional records come from Paper’s synthetic generator · seed 42. This example cannot show whether a real profile is truncated or a live product window is missing pages.

A pattern to investigate

The count records repeated overlap. It does not establish payment, coordination, or fraud. A live investigation needs source checks and an explicit record of inaccessible pages and incomplete coverage.

View the underlying identifiers Match the fictional names to the generated records

Names and product descriptions are display labels only. The method uses the original reviewer and product identifiers, which stay the same across settings. These labels do not identify real people or products.

Reviewers in this generated result
Fictional name	Generated ID
Alex Bennett	`demo-0`
Avery Bennett	`demo-1`
Jesse Bennett	`demo-10`
Kendall Foster	`demo-100`
Lee Foster	`demo-101`
Morgan Foster	`demo-102`
Parker Foster	`demo-103`
Quinn Foster	`demo-104`
Reese Foster	`demo-105`
Riley Foster	`demo-106`
Rowan Foster	`demo-107`
Sage Foster	`demo-108`
Taylor Foster	`demo-109`
Jordan Bennett	`demo-11`
Alex Hayes	`demo-110`
Avery Hayes	`demo-111`
Casey Hayes	`demo-112`
Cameron Hayes	`demo-113`
Dana Hayes	`demo-114`
Drew Hayes	`demo-115`
Elliot Hayes	`demo-116`
Emery Hayes	`demo-117`
Harper Hayes	`demo-118`
Jamie Hayes	`demo-119`
Kendall Bennett	`demo-12`
Jesse Hayes	`demo-120`
Jordan Hayes	`demo-121`
Kendall Hayes	`demo-122`
Lee Hayes	`demo-123`
Morgan Hayes	`demo-124`
Quinn Hayes	`demo-126`
Reese Hayes	`demo-127`
Riley Hayes	`demo-128`
Rowan Hayes	`demo-129`
Lee Bennett	`demo-13`
Sage Hayes	`demo-130`
Taylor Hayes	`demo-131`
Alex Lane	`demo-132`
Cameron Lane	`demo-135`
Dana Lane	`demo-136`
Drew Lane	`demo-137`
Elliot Lane	`demo-138`
Emery Lane	`demo-139`
Morgan Bennett	`demo-14`
Harper Lane	`demo-140`
Jamie Lane	`demo-141`
Jesse Lane	`demo-142`
Jordan Lane	`demo-143`
Kendall Lane	`demo-144`
Lee Lane	`demo-145`
Morgan Lane	`demo-146`
Parker Lane	`demo-147`
Quinn Lane	`demo-148`
Reese Lane	`demo-149`
Parker Bennett	`demo-15`
Riley Lane	`demo-150`
Rowan Lane	`demo-151`
Sage Lane	`demo-152`
Taylor Lane	`demo-153`
Alex Palmer	`demo-154`
Avery Palmer	`demo-155`
Casey Palmer	`demo-156`
Cameron Palmer	`demo-157`
Dana Palmer	`demo-158`
Drew Palmer	`demo-159`
Quinn Bennett	`demo-16`
Elliot Palmer	`demo-160`
Emery Palmer	`demo-161`
Harper Palmer	`demo-162`
Jamie Palmer	`demo-163`
Jesse Palmer	`demo-164`
Jordan Palmer	`demo-165`
Kendall Palmer	`demo-166`
Lee Palmer	`demo-167`
Morgan Palmer	`demo-168`
Parker Palmer	`demo-169`
Reese Bennett	`demo-17`
Quinn Palmer	`demo-170`
Reese Palmer	`demo-171`
Riley Palmer	`demo-172`
Rowan Palmer	`demo-173`
Sage Palmer	`demo-174`
Taylor Palmer	`demo-175`
Alex Reed	`demo-176`
Avery Reed	`demo-177`
Casey Reed	`demo-178`
Cameron Reed	`demo-179`
Riley Bennett	`demo-18`
Dana Reed	`demo-180`
Drew Reed	`demo-181`
Elliot Reed	`demo-182`
Emery Reed	`demo-183`
Harper Reed	`demo-184`
Jamie Reed	`demo-185`
Jesse Reed	`demo-186`
Jordan Reed	`demo-187`
Kendall Reed	`demo-188`
Lee Reed	`demo-189`
Rowan Bennett	`demo-19`
Morgan Reed	`demo-190`
Parker Reed	`demo-191`
Quinn Reed	`demo-192`
Reese Reed	`demo-193`
Riley Reed	`demo-194`
Rowan Reed	`demo-195`
Sage Reed	`demo-196`
Taylor Reed	`demo-197`
Alex Shaw	`demo-198`
Avery Shaw	`demo-199`
Casey Bennett	`demo-2`
Sage Bennett	`demo-20`
Casey Shaw	`demo-200`
Cameron Shaw	`demo-201`
Dana Shaw	`demo-202`
Drew Shaw	`demo-203`
Elliot Shaw	`demo-204`
Emery Shaw	`demo-205`
Harper Shaw	`demo-206`
Jamie Shaw	`demo-207`
Jesse Shaw	`demo-208`
Jordan Shaw	`demo-209`
Taylor Bennett	`demo-21`
Kendall Shaw	`demo-210`
Lee Shaw	`demo-211`
Morgan Shaw	`demo-212`
Reese Shaw	`demo-215`
Riley Shaw	`demo-216`
Rowan Shaw	`demo-217`
Sage Shaw	`demo-218`
Taylor Shaw	`demo-219`
Alex Brooks	`demo-22`
Avery Brooks	`demo-23`
Casey Brooks	`demo-24`
Cameron Brooks	`demo-25`
Dana Brooks	`demo-26`
Drew Brooks	`demo-27`
Elliot Brooks	`demo-28`
Cameron Bennett	`demo-3`
Harper Brooks	`demo-30`
Jamie Brooks	`demo-31`
Jesse Brooks	`demo-32`
Jordan Brooks	`demo-33`
Kendall Brooks	`demo-34`
Lee Brooks	`demo-35`
Morgan Brooks	`demo-36`
Parker Brooks	`demo-37`
Quinn Brooks	`demo-38`
Reese Brooks	`demo-39`
Riley Brooks	`demo-40`
Rowan Brooks	`demo-41`
Sage Brooks	`demo-42`
Taylor Brooks	`demo-43`
Alex Carter	`demo-44`
Avery Carter	`demo-45`
Casey Carter	`demo-46`
Cameron Carter	`demo-47`
Dana Carter	`demo-48`
Drew Carter	`demo-49`
Drew Bennett	`demo-5`
Emery Carter	`demo-51`
Harper Carter	`demo-52`
Jamie Carter	`demo-53`
Jordan Carter	`demo-55`
Kendall Carter	`demo-56`
Lee Carter	`demo-57`
Morgan Carter	`demo-58`
Parker Carter	`demo-59`
Elliot Bennett	`demo-6`
Reese Carter	`demo-61`
Riley Carter	`demo-62`
Rowan Carter	`demo-63`
Sage Carter	`demo-64`
Taylor Carter	`demo-65`
Alex Ellis	`demo-66`
Avery Ellis	`demo-67`
Casey Ellis	`demo-68`
Cameron Ellis	`demo-69`
Emery Bennett	`demo-7`
Dana Ellis	`demo-70`
Drew Ellis	`demo-71`
Elliot Ellis	`demo-72`
Emery Ellis	`demo-73`
Harper Ellis	`demo-74`
Jamie Ellis	`demo-75`
Jesse Ellis	`demo-76`
Jordan Ellis	`demo-77`
Lee Ellis	`demo-79`
Harper Bennett	`demo-8`
Morgan Ellis	`demo-80`
Parker Ellis	`demo-81`
Quinn Ellis	`demo-82`
Reese Ellis	`demo-83`
Riley Ellis	`demo-84`
Rowan Ellis	`demo-85`
Sage Ellis	`demo-86`
Taylor Ellis	`demo-87`
Alex Foster	`demo-88`
Avery Foster	`demo-89`
Jamie Bennett	`demo-9`
Casey Foster	`demo-90`
Cameron Foster	`demo-91`
Dana Foster	`demo-92`
Drew Foster	`demo-93`
Elliot Foster	`demo-94`
Emery Foster	`demo-95`
Harper Foster	`demo-96`
Jamie Foster	`demo-97`
Jesse Foster	`demo-98`
Jordan Foster	`demo-99`

Products shown in this result
Fictional product	Generated ID
Adjustable desk lamp	`P00000`
Insulated travel mug	`P00001`
Canvas tote bag	`P00002`
Bamboo cutting board	`P00003`
Cotton throw blanket	`P00004`
Glass storage jars	`P00005`
Compact phone stand	`P00007`
Stainless mixing bowl	`P00009`
Folding laptop stand	`P00011`
Reusable water bottle	`P00014`
Felt desk organizer	`P00017`
Metal bookends	`P00021`
Bamboo drawer divider	`P00023`
Pocket notebook	`P00025`
Fabric pencil case	`P00026`
Silicone spatula set	`P00027`
Clip-on reading light	`P00030`
Stainless tea infuser	`P00033`
Magnetic kitchen timer	`P00035`
Portable picnic blanket	`P00038`
Cotton oven mitts	`P00042`

FROM A RESULT TO A CONTRIBUTION

What would you share?

An investigation records what was observed and where it came from. See a small excerpt from this same example and how future reviewed contributions would work.

View an example contribution →

Change the search settings 10 histories · cap 5 · ±14 days

The budget limits histories examined. The cap guides which reviewer is examined next; it does not cap the count reported for that reviewer. The time window defines nearby activity.

Inspect the full result Counts, supporting egos, product bursts, and search history

Reviews in the example610Full generated panel; the search examines selected histories.

Histories examined10

Complete product windows8seed egos

Highest corroboration6distinct seed egos

COLLECTION PROGRESS

Each history adds context.

71 reviewers reached

CORROBORATION LEDGER

Inspect the reasons.

Up to 30 reviewers

A count records how many examined reviewers’ neighborhoods include this reviewer. Each distinct history adds at most one. The reported count uses eligible histories with complete product-window coverage. Expand a row to see the supporting reviewers, products, and weeks.

Rowan Hayes2 products6 egos

Generated record ID: demo-129

6 distinct product-window-complete seed egos; 6 observed seed egos.

Observed count, including incomplete coverage: 6

Reviewer whose history we examined (seed)	Product	Week (day index)
Riley Bennett	Cotton throw blanket	90
Avery Foster	Cotton throw blanket	90
Sage Foster	Cotton throw blanket	90
Cameron Reed	Adjustable desk lamp	97
Elliot Palmer	Adjustable desk lamp	97
Riley Ellis	Adjustable desk lamp	97

Riley Foster1 products4 egos

Generated record ID: demo-106

4 distinct product-window-complete seed egos; 4 observed seed egos.