Etsy and eBay in one browser: what glues two shops

Different logins and different tabs do not make different computers. Etsy and eBay read one set of hardware, one font list and one way out to the network, and they read it before you press sign in.

One value in common proves nothing: machines like that run into the millions. What proves it is the set, the one that repeats across both of your shopfronts and almost nowhere else.

Screen, fonts, memory and graphics card answer both platforms the same way, because one machine is doing the answering. Changing the login leaves those answers untouched.

A separate profile gives the second shop its own set of answers, and that set is derived from the profile itself. Here is how it came out across 900 desktop profiles of the current generation.

Setting the second shop up as a copy of the first is convenient: same notes, same proxy, same scenario. The fingerprint is the one thing a copy does not inherit, and here is why.

Two profiles carrying one fingerprint are tied to each other more tightly than by anything else: to a platform that is one machine with two logins. So a copy comes with a machine of its own, and the values you picked by hand do not travel with it.

The E-commerce scenario opens a profile on Amazon and attaches two sets: the fingerprint and the protection against automation detection. Its counter lists are the general ones, link cleaning is on, and both shops get the same treatment.

What tells two shops apart is not the scenario settings, which a platform cannot see, but the address, the clock, the languages and the fingerprint. One scenario covers seven platforms, and the number of profiles equals the number of your shopfronts.

You copy a link to your own item out of an email, out of an ad account or out of somebody's post, and a tail of tags rides along at the end. A tag tells the platform who sent the visit and from which account.

The same tail arriving at Etsy and at eBay ties the shopfronts together with no fingerprint involved. So a profile strips those tags from the link before the visit: 40 of them today.

1. 1

   Where the tail comes from

   Ads, newsletters and other people's posts append tags to the end of an address.

2. 2

   What it gives away

   One tail on two shopfronts tells a platform that the person is one.

3. 3

   What a profile does with it

   A profile strips the tags before the visit and leaves the useful parts of the address.

The E-commerce scenario names seven platforms, and Etsy and eBay stand there beside Amazon, Walmart, Shopify, WooCommerce and BigCommerce. The per-day numbers are shared, because the platforms are counting the same things.

Below are the daily ceilings of that scenario. The app writes them out on the profile card, under Safe daily ceiling, as a pair, today’s and the full one, and counts only page moves itself.

During the ramp an action row shows today's figure and the full one behind a slash: two account switches on day seven are one on day one. From day seven a single figure remains.

Action rows carry no colour: the browser cannot count new listings or buyer replies, so there is nothing to colour them by. A second line under the figure gives the threshold, «warning at N», with no colour either: it is a guide, not a count. Colour lives only in the Safe pace bar of pages opened: amber from eight tenths of the day's figure and red at the figure itself.

The day's figure on the ramp never drops to zero. An action the scenario allows once a day shows one on day one as well, not zero.

Of the six numbers above the tightest is the account switch: two a day. The number is a guide: the app shows it with the threshold «warning at 1», but it does not count switches and warns you of nothing by itself.

A new shop gets its first week as a ramp: on day one Safe pace gives it 24 pages out of 120, and the number grows day by day. A shopfront opened yesterday then does not behave like a warehouse that has run for a year.

You close the shop and delete the profile, and the name of that profile easily stays in the launch log, with its rows still in the action count for the day. The profile list looks clean, and the traces of the old shopfront lie on the disk.

So deleting a profile takes the name out of the log, its rows out of the count and its pace along with them. Whatever could not be erased the app names to you outright, rather than showing a green tick over a directory it never cleaned.

The first step puts the profile in the trash. Before that the program closes its window, so the shop's browser does not live on by itself with no row in the list.

The second step is the final deletion from the trash, and only that one takes the traces away: extension state, schedules, notifications, the launch journal and the day's action count. Five stores forget the profile in one move.

If the profile's session lay in the shared folder between machines and could not be removed, the program says so outright. There will be no green tick over a directory that is still there.

The check takes an evening and needs no purchase at all. Open both shopfronts in two profiles and put side by side what each one learns about you.

Compare the whole set rather than one value. If the set matches, then to a platform this is one machine, however different the logins look.

1. 1

   Compare what the machine answers

   Open the fingerprint check in both profiles and put the answers side by side.

2. 2

   Compare the way out to the network

   Address, timezone and language should not match between the two shopfronts.

3. 3

   Look at the tail of a link

   Paste a working link and make sure the tags never reached the platform.