About recommendations and ranking

1. Default ranking

When browsing and searching you can choose a sort order (e.g. newest, price ascending/descending, size). The default order is not personalized and is the same for all visitors — it is based on the chosen sort criterion and relevance to your filters (location, price, property type, etc.). Ranking that does not use your history is always available.

2. “Recommended for you”

For signed-in buyers we show a “Recommended for you” section on the home page. Recommendations are computed from your activity on the portal over the last 90 days; the main parameters and their weights are:

From these signals we derive the most frequent property types, cities and municipalities, the average price range and size, and suggest listings that match. If we do not have enough signals, we show popular or most-viewed listings. Recommendations do not use sensitive personal data and do not constitute automated decision-making within the meaning of Article 22 GDPR.

3. How to turn off personalization

You control the personalization mode in the Personalization section of your account:

• Recommendations (default) — display of the “Recommended for you” section.
• Active personalization — automatic filter suggestions while searching (requires explicit consent).
• No personalization — we delete the preference profile immediately; we treat you as an anonymous visitor.

Details about the legal bases are in the Privacy policy.

4. Paid and sponsored listings

The portal does not currently show paid or sponsored listings. If we introduce featured or paid placements in the future, they will be clearly labelled (e.g. “Sponsored”) and visually separated from organic results and recommendations; we will explain how the payment affects the placement.

5. Artificial intelligence (AI) features

Besides recommendations, we also use AI for: smart (semantic) search (it understands a descriptive query and ranks it by meaning within your filters), automatic listing tags (e.g. “bright”, “renovated”), suggestions to improve the description, and help with moderation of listings. Texts are processed by a provider with servers in the EU; we do not send buyers’ contact details.

Moderation is advisory — AI assesses a listing’s potential risk, but the final decision (publish, reject) is always made by a human. This is not automated decision-making within the meaning of Article 22 GDPR. We log moderation assessments so we can explain the reasons.

6. Your rights and access to data

You can access the data AI derives about you (preference profile, a possible “taste” vector, logged decisions about your listings) via the data export in the My account section (JSON export). You can turn off personalization at any time (point 3), which immediately deletes the derived profile. For an explanation of an individual decision, contact us via the contact form.

7. Property valuation

Any property valuation is indicative and is not a substitute for an official appraisal by a court-sworn appraiser. It is based on public statistical data (see Privacy policy — statistical data).

8. Ranking and paid placement

The default order of search results is chronological — newest first. You can change it at any time (price, size, price per m²) or choose Recommended, which uses your profile (section 2). Payment does not affect ranking in the search results list.

There are two places where featuring does affect placement, and we label both:

• Featured properties on the homepage — featured listings are ordered ahead of the rest. Listings featured as part of a paid package are marked “Paid promotion”; editorially chosen ones are marked “Featured”.
• The map — when a viewport contains more than 200 listings, featured ones are included first. They are labelled in the popup that opens when you click a marker.

The agency directory is likewise ordered with featured agencies first; those carry a “Featured” badge.

Google (AdSense) ad slots are separate from portal content and always labelled “Oglas”.

The match score in the buying process

In My buying process, each tracked property can show a match score against your preferences (e.g. "Match 82%"). It is computed automatically and is not AI text generation — it compares data.

• Your saved preferences: property type, sale/rent, town or municipality, price range and size.
• Similarity to your activity: how close the property is to the ones you viewed, saved or enquired about.

The score appears only when personalization is on. Every calculation is logged so we can explain it and it is included in your data export. Open the buying process and tap "Why?" next to a score for the breakdown.