This page discloses, in accordance with Article 5 of Regulation (EU) 2019/1150 ("P2B Regulation") as well as Article 27 of Regulation (EU) 2022/2065 ("Digital Services Act"), the main parameters by which listings on the McGesund platform are ordered in search results and list views ("Ranking"), together with their relative weighting.
We update this page on an ongoing basis. Material changes to the ranking procedure will be announced here and in the customer account at least 15 days before they take effect.
1. What is "Ranking"?
By ranking we mean the relative visibility of a listing compared with other listings in the same directory category, in the city context and in the search results of the platform's own search.
A higher ranking score leads to a more prominent placement — for example in the upper positions of a category list, in a teaser area of the city page, or in a map preview in the detail view of a related listing.
2. Main Parameters and Their Weighting
The order of listings is determined on the basis of the main parameters set out below. The weighting is a rough guide; depending on the search context (category, city, geographical search, detail recommendation), the effective weighting may shift.
| # | Parameter | Weighting | Explanation |
|---|---|---|---|
| 1 | Substantive relevance | very high | Correspondence of the search query with the company name, category, specialty and location of the listing. |
| 2 | Review trust score | high | Number, currency and authenticity markers of the existing reviews — in particular "Display-verified", quantum-resistant signature, anchoring in the Bitcoin blockchain. The number of reviews is factored in with a logarithmically flattening effect (see Section 4). |
| 3 | Currency and completeness | medium | Well-maintained address, category and contact data, supplementary descriptive and media content, regular updating. |
| 4 | Plan level | medium | Higher Plan levels may receive placements in teaser areas and category views that would not be attainable through purely substantive relevance. The plausibility of search results is not distorted: purely paid visibility is marked as such (see Section 5). |
| 5 | Geographical proximity | situationally high | For search queries with an identifiable location component (city, postal code, "near me"), the platform weights the distance between the presumed location of the searching person and the listing location. |
| 6 | Review average | medium | Aggregated heart rating across all four categories. Combined with the trust score (#2) into a combined reputation metric. |
| 7 | Response behaviour of the listing owner | low | Speed and proportion of reviews answered — as an indicator of active maintenance of the listing. |
| 8 | Click and dwell signals | low (aggregated) | Aggregated click and dwell-time signals of anonymous platform users on comparable listings. Never personal; session identifiers are discarded after a short lifetime. |
3. Order in the Event of Equal Rating
Where the overall score is comparable, the decisive factors are — in this order — the currency of the most recent review, the completeness of the profile and, finally, the time of first listing.
4. Logarithmically Flattening Number of Reviews
So that a large number of reviews does not dominate the ranking on its own, the number of reviews is factored into the score in a tier-based manner with a diminishing marginal return:
| Number of reviews | additional score contribution |
|---|---|
| from 1 review | +5 |
| from 5 reviews | +5 |
| from 20 reviews | +5 |
| from 50 reviews | +5 |
| from 51 reviews | 0 (saturation) |
Effect: A listing with 50 reviews receives the same volume contribution as a listing with 500 reviews. The ranking is thereby less distorted by the sheer volume of historical reviews; current review quality and authenticity markers gain in weight.
In its effect, this logic corresponds to a logarithmic review curve with hard saturation and is deliberately transparent — it protects both established and new listings from purely quantitative displacement.
5. Influence of Remuneration on the Ranking (Article 5(3) P2B Regulation)
(1) Business customers may book paid Plan levels that can favour the ranking — in particular in teaser areas and prominent list positions.
(2) The purely substantive order of results is not overridden by a booking: a more highly rated basic listing does not drop behind a more weakly rated premium listing where the search query is a better match for the basic listing.
(3) Placements attributable exclusively to the Plan level — for example in city teasers labelled "Recommended providers" — are marked as such in the user interface.
6. Equal Treatment of the Provider and Affiliated Undertakings (Article 7 P2B Regulation)
The Provider does not treat its own content and the content of undertakings affiliated with it more favourably in the ranking than that of other business customers, unless there is an objectively justified reason. Any such difference is disclosed here within the meaning of Article 7 P2B Regulation.
The Provider currently operates no listings of its own in the directory.
7. Machine Learning in the Ranking
(1) Current status. The ranking is currently calculated exclusively rule-based according to the parameters set out in Section 2. A machine learning model is not yet active in the inference path and does not influence the order of listings.
(2) Training data collection. In preparation for a later learning model, pseudonymised search signals are already being collected now. Per session, the Provider processes:
- the search term entered as well as the category and city context of the list view,
- the order of the results displayed (impression list, at most top 50),
- the result clicked and its list position,
- subsequent conversions (call, website visit, review) triggered within the same session on a previously clicked result.
(3) Pseudonymisation. Sessions are grouped via a randomly generated identifier that exists exclusively in the browser's sessionStorage, is reassigned after 30 minutes of inactivity and expires when the tab is closed. A linking with the user account, the IP address or a device feature takes place neither when writing nor when analysing.
(4) Minimum aggregation. In the daily aggregation, search cohorts below five distinct sessions per day are discarded entirely in order to exclude re-identification risks. Only cohorts of five or more sessions are factored into the long-term analysis.
(5) Storage period. Raw events are stored for 30 days and then deleted; the aggregated daily metrics remain indefinitely as a training basis. All processing is carried out exclusively on the Provider's own servers within the European Union.
(6) Planned use & oversight. A later learning model will be trained on these aggregated metrics and applied exclusively as a re-ranking layer over the heuristically calculated top list; the main parameters set out in Section 2 remain the authoritative basis. Before commissioning, the specific input signals, the model class used and the specialist oversight function will be added here; the change will be announced, in accordance with Article 5(5) of the P2B Regulation, here and in the customer account at least 15 days before it takes effect.
8. Influencing the Ranking by the Business Customer
The business customer can improve its ranking through the following measures:
- Complete maintenance of the profile (address, opening hours, specialties, description, logo, images);
- Active responding to reviews received;
- Recommending the reception display (as of the "Klassik" Plan), which generates "Display-verified" reviews with a higher authenticity weight;
- Regular updating of job postings;
- Where required, switching to a higher Plan level.
Insofar as the Provider additionally offers further remuneration models that directly influence the ranking — for example advertising slots acquired against individual payment, ad bookings or highlighted placements — these are clearly marked in the user interface as "Advertisement", "Advertising" or comparable and are separately described in Section 5 of this transparency page.
9. Complaints Regarding the Ranking
Business customers may submit complaints regarding the ranking via the internal complaint management system pursuant to Section 17 of the Terms — via the customer account in the "Support" area as well as by email to info@mcgesund.de.
Provider
Organon Informationssysteme GmbH
Karlstraße 31
63571 Gelnhausen
Contact
info@mcgesund.de · Tel. +49 (0) 69 9043 1680
gültig ab 28.04.2026