한국어 한국어   Englilsh English

.

Skip to end of metadata
Go to start of metadata

Q. How do you use RecoPick and what is the pricing policy?

A.  A brief guide on getting started with RecoPick is available at  recopick.com > Guide and pricing information at  recopick.com > Pricing.

Once registration is complete, refer to  Docs > RecoPick installation guide  for information on installing RecoPick recommendations. For more information on usage and pricing policy, etc., please contact us via e-mail at  recopick-help@skplanet.com.

Q. How do you apply recommendations?

A. To apply RecoPick recommendations, you must first insert 1) a log collection script to analyze item/user relationship, 2) a metatag to display product information on the widget, and 3) a recommendation widget to place the widget at the desired location on your website. Login using your Admin Dashboard account to access the Recommendation installation guide page and follow the guidelines provided. For further details, refer to Docs > RecoPick Installation Guide. Feel free to contact us in case of any difficulties.

Q. Are recommendation widgets applied immediately?

A. In order for optimal recommendation performance, we provide recommendation results 2~3 days after thorough analysis of user behavior. For larger-scale websites, a sufficient amount of correlation can be collected in a shorter period of time, however, a log collection period is necessary for data accumulation to make possible a more precise calculation of recommendations. You will be notified via e-mail when the results will be implemented on the recommendation widget. Once you receive the e-mail, you may select and save the desired type of recommendation, which is disabled at default, under Admin Dashboard  > Recommendation widget configurations and start viewing recommendation results on your website.

Q. What types of logs are collected?

A.  RecoPick recommendation requires purchase, my cart, and item view logs weighted in the order of purchase > my cart > item view.

Q. How do you monitor recommendation performance?

A. Recommendation performance can be readily monitored using Admin Dashboard’s recommendation performance analysis. You can monitor PV / number of items purchased / conversion rates of users visiting through recommendation as well as all users. If purchase price data is collected, you can also monitor revenue made via recommendation in the last 24 hours (customizable).

Q. What happens when changes are made to product information of the recommended items?

A. Product information is renewed and recollected every 24 hours at nighttime. You may designate a specific time to recollect data if you wish to update product information or carry out promotions at a specific time (morning, etc.) of the day. In case a sudden change has been made regarding product information of 1~2 items, you may recollect product information of individual items by clicking on product information in Admin Dashboard.

Q. If recommendations are not made due to a server error, how will it affect my shopping mall?

A. RecoPick currently runs on an AWS(Amazon Web Service) server and we have not yet experienced any sort of server error while providing recommendations to large-scale websites such as 11th street. Because we utilize an external call method, if in any way a server error occurs, the widget may disappear but the shopping mall’s design and functions will not be affected in any way.

Q. Are recommendations also useful in fashion-related shopping malls that are fast-changing and trend sensitive?

A. Definitely. Fashion/accessory shopping malls make up the largest portion of our client pool and are showing great results. Our exclusive real-time personalized recommendation, which analyzes user behavior in real-time, is best suited for trend-sensitive fashion/accessory shopping malls that need to cope with fast-changing customer interest.

Q. How do item-based and user-based recommendations differ?

A. Item-based recommendation utilizes individual item logs to suggest ‘users whoviewed this item also viewed(ViewTogether)’ or ‘users who purchased this item also bought(BuyTogether)’. On the other hand, user-based recommendation utilizes user interest and purchase patterns to calculate recommendations based on items that a user has recently viewed.

Q. What is a channel seen in recommendation click-through rate and revenue via recommendation?

A. A channel is an identifier of a widget used for monitoring recommendation click-through rate or revenue via recommendation of each widget when multiple widgets are inserted at different locations of the website.

By designating a channel when using Recommendation API or widget, statistics for each channel can easily be monitored in Admin Dashboard.

For more information on how to designate channels, refer to Client Library APIs for recommendation widgets and GET v1/recommendations/item/:service_id/:user_id/:item_ids for Recommendation API.

  • No labels