Ce qu’il faut vérifier avant d’acheter un compte sur les réseaux sociaux : les quatre attributs que le marché valorise réellement

Three tabs, three listings, apparently the same product. The honest way to choose between them is not to guess which seller is nicer. It is to work out which of the differences between the listings are ones the market has already put a price on, and which are words anybody can type.

Measured across the whole live catalogue on this site, that question has a clean answer, and it is not the one most buying guides give.

A stated year is priced. The word "aged" is not

Listings here carry classifier tags, and each tag has a description that says what it is supposed to mean. Comparing the median price of the listings carrying a tag against the median of the catalogue gives a multiple, and reading the multiples in order tells you what buyers are short of.

Attribute on the listing	Listings carrying it	Price vs catalogue median
Registered 2017 to 2021	909	2.65x
Manually registered	1,354	2.03x
Registered 2022 to 2025	2,825	1.72x
Mailbox access included	5,442	1.70x
Age mentioned, no year given	2,687	1.26x
Two-factor details included	6,696	1.06x
Auto-registered	2,750	0.64x
Registered 2026	655	0.57x
"Fresh", no year given	2,698	0.35x

Read the three year rows in sequence. From this year's registrations to a five-to-nine-year-old account the price moves by a factor of about four and a half, in a straight line, with thousands of listings behind each step. Now read the two rows that describe age without giving a number. "Fresh" is the cheapest thing in the table and a bare mention of age sits barely above the middle. The market prices the number and ignores the adjective.

So the first thing to check is whether the listing states a registration year at all. If it says "aged" and nothing else, you are not buying age, you are buying a word that costs the seller nothing to type.

Provenance outranks almost everything except the year

The two rows worth putting side by side are manual registration at roughly twice the median and automatic registration at roughly two thirds of it, a spread of about three to one on more than four thousand listings between them. Nothing else in the table separates that cleanly on that much volume.

Bundled mailbox access sits just below manual registration. That is the practical one: where the mailbox comes with the account, recovery is something you can perform yourself, and where it does not, recovery is a conversation with whoever still holds it. Two-factor details barely move the price at all, which tells you they are close to standard rather than valuable.

The after-sales window is measured in hours, and on some listings it is zero

This is the check most buyers skip and the one that most often decides whether a bad order can be fixed. Every listing carries its own warranty, it is written in hours rather than days, and the clock starts when the order is delivered rather than when you pay.

Across the live catalogue the median is twelve hours. That is already much shorter than the thirty-day thinking most guides import from consumer marketplaces. The distribution matters more than the median:

• About nine in every hundred live listings carry a window shorter than ten minutes, which the platform treats as no after-sales window at all.
• Another fourteen in a hundred sit between ten minutes and an hour.
• A little under six in ten sit at exactly twelve hours.
• Around eight in a hundred run to a full day, and roughly two in a hundred to a week.

The zero-window group is worth one more sentence, because the shape of it is reassuring rather than alarming. It is not spread across the market. Six sellers account for all of it and one of them accounts for nearly all of that. It is a property of a particular shop, it is written on the listing, and you can read it before you pay.

Practical consequence: check your window, then plan the first hour rather than the first month. Test the login immediately, on the day, and do it before you do anything that would change the account. For what the clock actually gates, see opening a dispute.

Most listings have never sold here, and that is not a defect

Fewer than one live listing in five has ever carried a completed order on this site. That sounds worse than it is: a marketplace this size lists far more than it sells, and a new listing from an established seller starts at zero like every other.

What it means for you is that "this listing has buyers" is a much rarer and therefore much stronger signal than it looks, and that the signal is wildly uneven by category. On some shelves close to half the listings have sold; on others it is one in thirty. Before you treat a lack of visible history as a warning, check whether anything on that shelf has history at all.

One number you should not use for this: the sales count some listings show is imported from the upstream shop the stock came from, not from sales made here. It is present on only a few suppliers' listings, and where a listing has both that figure and real orders on this site, the two agree in two cases out of six hundred and thirty. Ignore it and look at whether the listing has reviews attached instead.

Read the written reviews, not the star average

Being direct about our own numbers, since the alternative is telling you to read a figure we know is not informative. When an order completes and is confirmed, this platform writes a silent five-star review against it automatically. That is why the site-wide average sits at 4.99. Of roughly eighty-three thousand visible reviews, about a hundred and sixty carry any text a buyer actually typed, and the average across only those is 2.49.

Neither number is the truth on its own. The first is a completion count wearing a star rating; the second is a complaint log, because a buyer who has to be motivated enough to override the automatic review is usually a buyer with something to say. Read them as a pair. The platform exposes both, a written count and a written-only average, alongside the headline figure for exactly this reason, and the written count is the one that tells you whether anybody has ever described this product in their own words.

The same logic applies to a seller: a large order count is real, a five-star average is close to automatic, and a written review is the only part a human wrote.

Check the engagement yourself, and distrust any threshold you are quoted

For a follower-bearing account, the audience is the part no listing field can describe. Count likes and comments across the last ten to fifteen posts, divide by follower count, and compare that against a few accounts of similar size in the same niche.

What you should not accept is a specific cutoff. Numbers like "above three per cent is healthy" circulate widely in this space and none of them has a source anybody can point to, including in our own data, which contains nothing that would let us calculate one. Use the ratio to compare listings against each other, not against an invented bar. A sudden change in tone or subject partway down a profile grid is a more reliable signal than any threshold, because it means the audience attached to the account was gathered for something other than what you plan to do.

What you can actually verify, in order

1. Does the listing state a registration year, or only an adjective? The year is what the market pays for.
2. Manual or automatic registration, and does the mailbox come with it? These are the two attributes that move the price most after the year.
3. What is the after-sales window, in hours, on this specific listing? It can be zero, and it starts at delivery.
4. Has anything on this shelf ever sold, and does this listing have written reviews rather than a star average?
5. How many sellers are on this shelf? On some you are comparing hundreds of them; on others one seller holds most of the stock and there is no comparison to make.

Roughly two minutes per row, all of it on the listing page, none of it dependent on a seller answering a message. For the platform-level version of the same question, whether the marketplace itself is worth trusting rather than the listing, see how to compare marketplaces. For a worked example of these checks on one shelf, see the Telegram supplier breakdown, and browse Instagram accounts or Twitter/X accounts to run them yourself. If Telegram is the only platform you are sourcing for, a specialist storefront such as tgaccounts.net publishes its own tenure and delivery notes worth cross-checking.