Listing Search Advanced Feature
Allows users to search listings using a predetermined algorithm
Last updated
Was this helpful?
Allows users to search listings using a predetermined algorithm
Last updated
Was this helpful?
Search relevance is key to maximizing matching potential and transactional performance. Second integrates a built-in machine-learning algorithm to provide the best results possible.
Search results can be sorted by different means :
Platform recommendation
Price
Distance
Sorting by the platform recommendation (used by default) seeks to provide the asker with the results that are most likely to lead to a transaction. Platform recommendation is managed by the Second search engine which makes use of machine learning to maximize the transaction rate between the lister and the asker.
Find more details about the Second search results in the Special Rules section below.
All services
Please, find all interface documentation related to the Listing search advanced bundle, here:
Front:
In the following documentation, you can find configurations related to the listing search advanced bundle:
There is no special workflow for the listing search advanced bundle.
The listing search constructs its results in the following way:
A global parameter of the application defines the minimum number of desired results to be returned for each search. This setting is set to 10 by default.
The purpose of this functionality is to provide a minimum number of desired results for a given search query by enlarging the covered geographic area.
Initially, the covered area is determined by a circle whose diameter is equal to the diagonal of the viewport selected by the Google Maps API for the searched location. If the desired number of results is not achieved within this area, the functionality automatically expands the covered area to the next zone. The radius of the next zone is twice the radius of the previous zone.
This process is repeated until the desired number of results is achieved or until the covered area reaches a maximum distance of 2000km from the searched location.
By implementing this functionality, the user can obtain more results and increase the chances of finding the desired information. The functionality is flexible and can be adjusted to the user's needs, allowing them to achieve the desired results without manually adjusting their search criteria.
The results obtained by the previous operation are then divided into groups and subgroups by distance. The groups and subgroups are formed as follows:
i. For results within one kilometre of the searched location: no subdivision is performed and no subgroup is formed
ii. For results ranging between 1 and 10 kilometres from the searched location, and if the selection of the results has been extended to this distance, the results are divided into 4 equidistant subgroups
iii. For results that are between 10 and 50 kilometres from the searched location, and if the selection of the results has been extended to this distance, the results are divided into 6 equidistant subgroups
iv. For results that are more than 50 kilometres from the searched location, and if the selection of the results has been extended to this distance, the results are divided into 8 equidistant subgroups
The search area is expanded until the minimum number of results defined for the platform are reached (for example 10 results).
When searching by date, each of the 1 to 19 search result groups mentioned above are again divided into two subgroups:
Listings with confirmed availability dates matching the dates of the search are displayed first.
Listings with unknown availability dates are displayed second.
Once in the day, an algorithm sets a score for each listing and this score is used to sort the listing results. The algorithm scoring is calculated as follows:
This rating is given by the administrator in a free and subjective manner. The rating given by the administrator ranges from 0 to 10 with 0.5 point increments.
Listing completion takes into account the amount of information provided by the lister about the listing: is at least one attribute selected? Does the description contain more than 250 characters? Does it have more images than the minimum required?
Profile completion takes into account the amount of information that the lister has completed about himself: does his presentation text contain more than 250 characters? Has he added more images to his profile than the minimum required?
The reviews received by listers are an important part of our platform, as they help us ensure that our users have access to high-quality service providers. We use these reviews primarily to identify poorly performing suppliers and remove them from our platform, rather than to promote good suppliers in the search results.
Our approach is based on the idea that good performance is the minimum expectation from a service provider, and that poor performance must be heavily penalized. To achieve this, we take into account the average score of the lister, which is calculated by dividing the total number of ratings they have received by the number of reviews.
By using this method, we can ensure that our platform remains a trusted source of high-quality service providers, and that our users can make informed decisions when choosing a supplier. We are committed to providing a fair and transparent review system that benefits both our users and our listers, and we continually review and update our policies to ensure that we achieve this goal.
Updating the calendar provides increased assurance regarding the dependability of the listing's availability, thereby enhancing its score. A more recent calendar update leads to a higher score due to greater reliability.
If a service provider has had a high rate of providing services in the past 30 days, it indicates that they are likely to be open to accepting booking requests. A positive rating is given based on the number of bookings that were paid for and not canceled within the last 30 days.
The response rate can be calculated by dividing the number of messages received by the lister in response to a specific communication by the total number of messages they sent as part of that communication.
The acceptance rate of a lister refers to the proportion of requests they have agreed to, irrespective of whether the booking was ultimately confirmed or not.
The calculation of successful transactions is determined by dividing the number of authorized wire transfers by the total number of accepted bookings.
Considering response times enables us to prioritize the most prompt responders. Response time is assessed for each thread by measuring the duration between the receipt of the last message and the provision of a response. If no response is provided, no time is factored in. A shorter response time results in a higher score being assigned.
The administrator has the option to certify each listing, which can result in an improved position for the listing in the search results.
In order to discover and support top-performing suppliers who are new to the platform, a bonus is provided to those who have submitted their listings within the past 30 days.
To prevent search results from becoming stale and to provide a fair chance to listings that did not initially succeed, a daily random bonus is given to 5% of the listings on the platform. The bonus is awarded randomly and can improve a listing's rating, which ranges from 1 to 5.
In situations where two listings have the same rating, their order is determined by their distance from the searched location. The default sorting of search results is updated once per day.