Moderation rules

Every advertisement is checked automatically before it is published. The check compares the text, price, category and the seller’s account against the posting rules: prohibited items, signs of fraud, contact details in the title, duplicates, and so on. A clean advertisement is published as soon as the check is complete. If the rules only raise doubt, the recognition model described below decides at once — if it finds nothing, the advertisement is published. An advertisement that appears to breach the law or the rules, and one about which the model confirms the doubt, is held: it is visible only to its author until a member of staff has looked at it, usually within 24 hours. The author is notified at once.

Photos are checked as well. The text and photos of an advertisement are reviewed by an automated image and text recognition model: a fast model looks at every advertisement, and if it sees something that may be prohibited — weapons, drugs, medicines, alcohol, tobacco, nudity, counterfeits, or photos that do not match the description — a more capable model looks again. While the check is running, usually a few seconds, the advertisement is not visible to anyone else. If the fast model cannot complete the check, the more capable model decides. If the model confirms the doubt, the advertisement is held for a person as described above. If no model can complete the check and the rules found nothing, the advertisement is published at once.

The automated check never takes the final decision. A person reviews every held advertisement and either publishes it or refuses with a statement of reasons; a refusal may be appealed as described below. We keep the outcome of each check, including the rules that were triggered, together with the decision for 3 years.

The subject of the advertisement is prohibited from circulation, or its sale requires a permit which the seller does not hold.

The advertisement is misleading: the price does not correspond to the item, the photographs belong to another person, the description does not relate to the item offered.

The advertisement is placed in the wrong category, duplicates an earlier one, or contains contact details instead of a description.

The content infringes the rights of third parties, contains insults, discrimination or calls to break the law.

The author has given false information about themselves or their business, poses as another person, or has stated a business they are not entitled to represent.

The advertisement breaches the posting rules: it repeats one already published, has been republished after removal, contains brand names or words unrelated to the item to influence search, or has a bait price.

The author breaches the rules on communication and trust: sends spam, fraudulent links or requests, harasses other users, or creates fictitious deals or reviews.

Anyone may report content they consider illegal or in breach of the rules through the “Report” function on an advertisement, a profile, a message or a review, or by e-mail to info@nordo.ee. A notice should state why the content is illegal, where it is (link), the name and e-mail of the person reporting it, and a confirmation that the notice is made in good faith. We confirm receipt, decide without undue delay, and inform the reporter of the decision and how to challenge it — in the notifications of the service or, if the report was made without an account, by e-mail.

When people with different accounts report the same advertisement, it is hidden until a member of staff has looked at it; only accounts older than one day are counted, so that a competitor cannot hide someone else’s advertisement with newly created accounts. When a message is reported, an automated model reads that conversation and gives our staff a short assessment; the decision is always taken by a person. Photos from advertisements removed for breaching the rules are remembered by their fingerprint: if the same photo is uploaded again, the advertisement is held for review.

Notices from trusted flaggers designated under the DSA are handled with priority.

When we remove or restrict content, suspend or terminate an account, or restrict a user’s ability to earn money on the service, the affected user receives a clear statement of reasons: what was done, on what facts, whether the decision was based on law or on our terms, whether automated means were used, and how to appeal.

The author is notified of a withdrawal in the service’s notifications, with the ground stated. The advertisement stays in the author’s list of advertisements with the status “Removed”.

If the decision is considered erroneous, it may be appealed within six months of the notification: on the website under Profile → Notifications → Moderation decisions, or by writing to support. The appeal is reviewed by a person — where possible, one who did not take the original decision.

Restriction of access to an account is applied in cases of systematic breaches or conduct bearing the signs of fraud. The ground and duration of the restriction are communicated to the user.

A user affected by a decision, and a person whose report was rejected, may lodge a complaint within six months of the decision. Complaints are free of charge, are reviewed by a person and not solely by automated means, and are answered without undue delay. If the complaint is justified, the decision is reversed.

If you disagree with the outcome of a complaint, you may refer the matter to a certified out-of-court dispute settlement body under Article 21 of the DSA in your country. We engage in good faith with such bodies. This does not affect your right to go to court.

Users who frequently provide manifestly illegal content and users who frequently submit manifestly unfounded notices or complaints are warned and may then be suspended for a reasonable period, taking into account the number and gravity of cases and the intention behind them.

Mass messaging is limited automatically. A message with signs of fraud that is sent to three or more conversations within an hour is not delivered. Such sending, the same text sent to many conversations, and a new account that opens many conversations within an hour are reported to our staff, who decide whether to warn or restrict the account; the automatic limit itself never restricts an account.

The posting rules apply to a provider card and to the messages of a job request in the same way as to an advertisement. Before a card is published, its text is checked automatically by the same rules as an advertisement (prohibited goods and services, contact details and links, signs of fraud), and photos that were already refused in an advertisement are recognised. A card that raises doubt is held: it does not appear in the catalogue until a member of staff has looked at it, usually within 24 hours, and its holder is notified at once. The automated check never takes the final decision. We keep the outcome of each check and the decision on it for 3 years.

A provider card can be reported directly from its page, by anyone, including without an account; a complaint about what was written in a request is made about the message. The procedure is the same as for advertisements: a decision with a statement of reasons, the possibility to appeal within 6 months, and the same protection against abuse of the complaint procedure. A complaint that is upheld removes the card from the catalogue.

Where a card or a request breaches the rules, we may hide the card, remove the kinds of work or the photos that breach them, warn the holder or restrict the account, as described above. The holder is told the reason.

We publish reports on our content moderation with the frequency and content required of us under the Digital Services Act: the number of notices received and orders from authorities, the actions taken, the use of automated tools and the outcomes of complaints.