Show last authors
1 (% class="box" %)
2 (((
3 **Contents:**
4
5 {{toc depth="5" start="3"/}}
6 )))
7
8 To increase the customer loyalty and build a strategy for retention, return and reduction of customer churn, Partners can use various recommendation systems to generate [[personal offers >>doc:Main.General_information.Loymax_Loyalty.Offers.Individual_offers_in_LP.Personal_offers.WebHome]]for Loyalty Program (LP) members and send mailings through the most effective [[communication channels>>doc:Main.Using.Communications_ways.WebHome]].
9
10 === How recommendation systems work ===
11
12 Recommendation systems collect and update data based on changing customer preferences, while also analyzing existing historical data. The following groups of parameters are considered when collecting and analyzing data:
13
14 * personal data of LP Members (e.g. gender, age, presence and age of children, etc.);
15 * purchasing power (e.g. frequency and number of purchases, receipt amount and length, producer-related preferences, promotional sensitivity and communication channel preferences, etc.);
16 * products and product categories (e.g. frequency and number of purchases, product attributes, price niche, presence of product categories in the receipt, etc.);
17 * retail outlets (e.g. attributes of retail outlets, price segment of buyers and goods, availability of manufacturers, etc.);
18 * calendar and seasonal factors (dependence of purchases on the season, time of year, day of the week, holidays, etc.).
19
20 The results of data collection and analysis are processed and then the optimal model for the specific personal offer is selected. As a result, the LP members receive unique offers through the communication channels that suit them best.
21
22 === Integration of Loymax Loyalty with recommendation systems ===
23
24 The **Loymax Loyalty** module can be integrated with any recommendation system to automate processes and free up marketing team resources. This will allow to optimize the approach to setting up offers with account of offer personalization, reduce the load on calculation of offers, provide an instant System response, even if there are complex rules in the offer, eliminate delays, etc.
25
26 To integrate the recommendation system with **Loymax Loyalty**, the following settings should be made:
27
28 |(% style="border-color:#ffffff; text-align:center" %)[[image:attach:Recommendation_systems_en.png]]
29
30 ==== Integrations ====
31
32 During the initial stage, the Partner and Loymax representatives carry out preliminary system settings, including [[integration with the Partner's ERP system>>doc:Main.Integration.Loading_data_into_the_system.ERP_system_integration.WebHome]]. This involves uploading of the product catalog with all product attributes. Integration of the [[cash register protocol>>doc:Main.Integration.Cash_register_integration.Integration_methods.Exchange_protocol_with_cash_register.WebHome]], the functionality of displaying messages to the cashier and message printing on the receipt are checked for completeness.
33
34 ==== Data collection ====
35
36 Information for analytics and generation of personalized offers is downloaded from the data warehouse (DWH). There are also direct data loadings from the Partner's database.
37
38 ==== Uploading personal offers and selecting communication channels ====
39
40 At this stage, the external recommendation system is integrated (via a special loader on the Partner's side) using the **Loymax Loyalty API methods**. This stage includes creating and uploading customer attributes, uploading target audience lists, and configuring all basic parameters and restrictions for setting up offers, as well as coordinating the mechanics used, etc. For example, if the **Loymax ML** module is used as a recommendation system, it is possible to select up to 8 [[mechanics for generating personal offers>>doc:Main.Installation_and_configuration.Extra_modules.CommunicationService_ML.WebHome]]:
41
42 1. Discount, % per receipt,
43 1. Points, % per receipt,
44 1. Discount, % of receipt amount,
45 1. Points, % of receipt amount,
46 1. Discount, % for category,
47 1. Points, % for category,
48 1. Discount, % for product,
49 1. Points, % for product.
50
51 ==== Steps for uploading personal offers ====
52
53 |(% style="border-color:#ffffff; text-align:center" %){{lightbox image="Personal_Offer_Uploading_eng.png" width="800"/}}
54
55 To upload personal offers to the customer's attribute, this attribute should first be created. The type of attribute for personal offers must be JSON. The attribute structure may differ depending on the mechanics used. To learn more about [[creating and uploading customer attributes>>doc:Main.Using.MMP.Marketing.Offers.Iterator.Attributes_update.WebHome]] and [[uploading personal offers>>doc:Main.Integration.Ways_to_use_API.System_Api_Methods.Sapi_examples.Batch_update_of_customer_attributes.Personal_offers.WebHome]], please refer to the relevant articles. Next, perform [[batch uploading of attributes>>doc:Main.Using.MMP.CRM.Load_client_attributes.WebHome]] and [[import offer templates>>doc:Main.Integration.Loading_data_into_the_system.Importing_offers_API.WebHome]] with the required mechanics for the created attribute.
56
57 Also, if a special **Loymax Loyalty** module is used, [[channels for communication>>doc:Main.Using.BI-analytics.Reports_Excel.BI-analytics_Reports_Excel_2022.Communications.WebHome]] with customers are selected and configured, and mailing templates are formed at this stage.
58
59 ==== Additional information ====
60
61 The results of calculations with personalized recommendations are transferred to the **Loymax** **Loyalty** module ([[Loymax processing>>doc:Main.General_information.Loymax_Loyalty.Processing.WebHome]]), where the following actions are carried out:
62
63 * creation of [[offers>>doc:Main.Using.MMP.Marketing.Offers.General_information.Offers_life_cycle.Creating_offers.WebHome]] and [[mailings>>doc:Main.Using.MMP.Communications.Mailings.Mailing_creating.WebHome]] from templates,
64 * generation of [[target audiences>>doc:Main.Using.MMP.CRM.Target_groups.WebHome]] (if a control group takes part in the promo/campaign),
65 * [[creation of counters>>doc:Main.Using.MMP.Marketing.Counters.Counter_creation.WebHome]] to keep track of the number of triggered offers when using mechanics with limitation,
66 * calculation and provision of preferences ([[calculation of direct discounts>>doc:Main.General_information.Loymax_Loyalty.Processing.Typical_processes.Discount_in_LP.WebHome]], calculation and accrual/deduction of bonus points).
67
68 The results of processing can be further used by the recommendation system for generation of new personal offers.
69
70 (% class="box" %)
71 (((
72 **Read also:**
73
74 * [[Setting up entities in Machine Learning>>doc:Sandbox.Настройка сущности Machine Learning.WebHome]]
75 * [[Personal Offers configuration using Machine Learning mechanics>>doc:Main.Installation_and_configuration.Extra_modules.CommunicationService_ML.WebHome]]
76 * [[ML report on personal promo effectiveness>>doc:Main.Using.BI-analytics.Dynamic_reports_Power_BI.Version_2022.ML_Personal_Promo.WebHome]]
77 * [[Uploading customer attributes>>doc:Main.Using.MMP.CRM.Load_client_attributes.WebHome]]
78 * [[Examples of Personal Offer uploading>>doc:Main.Integration.Ways_to_use_API.System_Api_Methods.Sapi_examples.Batch_update_of_customer_attributes.Personal_offers.WebHome]]
79 )))
Created 09/01/2024 13:56
Last edited 06/05/2024 09:37