These apply across every region at once: the tiles below and the brand lists
read the same set, so the numbers agree with the rows. The discovery window
filters the cases inside a brand, and a brand with no shop in the window drops
out of the lists entirely.
To narrow one region instead, open it below: the drill-down has its own
filters for name, account owner, revenue and size, applied inside that region
only.
Fake shop detection
Attribution used to arrive only by spreadsheet, so the lead pool grew as
fast as somebody typed. The detector sweeps every identified storefront
BrowseCat has, roughly 116k of them, against a brand vocabulary built from
the brands already in LeadEye and from the World Brand Library, and
proposes which brand each shop is trading on.
A proposal is not an attribution. It reaches the lead lists only when you
confirm it, and confirming never overwrites a brand somebody already put on
the domain. Rejecting is permanent: the detector will not offer that brand
for that domain again, which is the point of keeping the row.
The exception is a brand that already works with us. There is no lead to
qualify, so those go straight in and appear here as applied.
Maintenance
Detection runs on its own every night at 01:30, and the liveness pass an hour after it. Starting one here does the same sweep over the whole corpus: it takes several minutes, the counts above move as it goes, and brands that already work with us are applied without review. There is no need to run it to see last night's proposals, which are already in the table.
Reports for the marketing offices
A periodic snapshot of what the detector found, split the way the offices
work: shops against brands that already work with us, separately from shops
against brands that do not. The first is service delivery, the second is
new business.
A report is frozen when it is generated. Rejecting one of its detections
afterwards does not change it, because the number an office was given has
to stay the number an office can check. Regenerating a period is the one
thing that replaces it.
Nothing is emailed yet: the host has no mail credentials. Reports are built
every Monday by cron and wait here to be opened or downloaded. The HTML is
already the shape a mail body would take.
Live tools
Brand Lead Finder
Type part of a brand name to list every storefront attributed to it. Each row is a
lead: the infringing domain, the brand's home country and region, how many products
were extracted, and when it was discovered. Substring match, case-insensitive,
against scanned_domains.brand. Every matched brand gets a Salesforce
block above the table, where every annotation of that brand can be set: owner,
Salesforce link, the region it really belongs to, its parent company, whether it
is already a client, and how big it is. Saving needs you to be signed in. The copy
button puts every infringing URL of the brand on the clipboard, one per line,
across all pages. Only the discovery window from the filter bar applies here, so
a search is never emptied by a filter set for the region views.
Report a fake shop
A shop somebody sent you, or found in an ad, that our crawl has not seen. Put
it in here and it is a lead for the whole team straight away, in the lists and
the region tiles like any other case.
All three are needed: the brand being ripped off, the shop's address, and the
country the brand owner sits in, because that is what decides the
sales region. The discovery date is filled in for you, today's date, since
that is the day anybody here first saw the shop.