Synthetic Customer Flow Simulation and Marketing System for Telecommunications Dealer Networks

Abstract

The invention relates to a system used in telecommunications, physical retail network management, data-centric regional marketing strategy, pricing optimization, and customer behavior simulation that models and simulates a telecommunications operator's physical retail network and the recorded customer behavior associated with that network in a completely synthetic digital environment, without any real risk to personal data.

Description

1 TARIFF Synthetic Customer Flow Simulation and Marketing for Telecommunications Dealer Networks The system Technical Area The invention encompasses telecommunications, physical retail network management, and data-centered regional marketing. strategy, pricing optimization and customer behavior simulation areas used, a telecommunications operator's physical dealer network and recorded connections to that dealer network By anonymizing customer behavior, we create a completely synthetic system without the risk of disclosing real personal data. It involves a system that models and simulates in a digital environment. State of the Art Current telecommunications marketing and dealer network analytics tools contain real personal data. carries risks based on usage or the impact of the physical distribution channel (dealer) on customer 15 They are unable to adequately integrate behavior into their models. These shortcomings in the process and the resulting... The technical problems encountered are as follows:  Legal and Ethical Risks of Simulating with Real Data: Real data from a specific region Running "what if" scenarios on customer data is covered under GDPR / KVKK. Processing personal data without explicit consent constitutes processing without consent and carries a high legal risk. 20  Separate Customer and Dealer Behavior Models: Current simulation The tools generally analyze customer behavior or dealer / point-of-sale performance separately. models. A change in a dealer network (e.g., a dealer closing) affects customers. The realistic impact of tariff selection, complaints, or churn cannot be modeled.  Insufficient Integration of Geographic Factors: Customer behavior models often have 25 It is based on national averages. The socio-economic structure of a province, the density of competing dealers, or Geographic factors such as transportation habits profoundly influence the campaign response, however These factors are not addressed in sufficient detail in current systems.  Low Real-Life Applicability of Simulation Results: Only the customer The outputs of simulations that focus on metrics but do not include dealer channel dynamics are 30 This may not align with real-life applications, leading to misleading strategic decisions. There are no anonymized structures created using real data. In conclusion, due to the negative aspects described above and the current solutions being the subject of discussion... Due to its shortcomings, it has become necessary to make improvements in the relevant technical field. 35 2 Purpose of the Invention The invention was created by drawing inspiration from existing situations and overcoming the aforementioned drawbacks. It aims to solve the problem. The main purpose of the invention is to generate confidentiality-guaranteed synthetic data and to facilitate geographical-dealer-customer interaction. 5 incorporating modeling that preserves the statistical properties of real data but individual by creating an anonymous synthetic population and dealer network, an infinite number of people can operate within this virtual environment. The goal is to provide a system that allows for risk-free scenario testing. Another purpose of the invention is to apply it to telecommunications, physical retail network management, and data center applications. regional marketing strategy, pricing optimization, and customer behavior simulation usable in various fields, a telecommunications operator's physical dealer network and the connections to this dealer network By anonymizing recorded customer behavior, without the risk of actual personal data loss, The goal is to create a system that models and simulates in a completely synthetic digital environment. Another objective of the invention is to analyze the statistical characteristics of real customer data for each province / district. (demographics, tariff usage, churn rate, packet preferences / switches, data consumption, etc.) and physical dealer behavior (dealer location, service quality, promotion application habits, sales) a "Synthetic" record that preserves (performance) but where all individual records are entirely artificial. The goal is to create a "Telecommunications Universe"; within this universe, region-specific tariffs will be applied. changes such as campaigns, increases / decreases in the number of dealers, or changes in advertising budget By running scenarios, data such as churn, revenue, and subscriber numbers are analyzed before implementation in real life. The goal is to predict potential impacts on critical metrics; data privacy regulations such as GDPR and KVKK. without violating the laws, a special tariff for a specific geographical region (e.g., Eastern Anatolia) or Simulating campaign results securely without accessing actual customer data. 25 The goal is to facilitate the addition / closure of physical dealerships, changes to dealership locations, or customized dealership options. the impact of promotions on regional customer acquisition and churn The goal is to model it; multiple options such as "Price + Advertising Budget + Number of Dealers". the consequences of changing the marketing and distribution channel variable simultaneously The aim is to enable the prediction of outcomes; high-risk or costly regional decisions (e.g., the new 30 (a tariff launch) before implementing it, "playing" it in a virtual environment and potential The goal is to minimize the risk of failure; complex telecommunications dealers in the real world. Creating a privacy-guaranteed digital twin of the customer ecosystem and operating risk-free on that twin. The aim is to enable experiments to be conducted; to provide real historical subscriber data. by processing, a statistically realistic but completely artificial 35 for each region and dealer. The goal is to build a "Synthetic Telecom Universe"; no individual has access to real data. 3 millions of unconnected synthetic subscribers and hundreds who will interact with them The goal is to ensure that it consists of synthetic dealer agents; "In this province, we will change the tariff price by this much, and this..." What would happen if we added more dealers? What would the profit be if we, as XX dealers in Sivas, ran these campaigns? The goal is to enable the definition of a scenario such as "we can make a sale"; the defined changes The goal is to advance time by applying it to a virtual environment; synthetic customers, 5 at the new price. and respond to existing / new dealers, make churn decisions, new subscriptions to initiate; all these interactions are governed by the dealer agents' code of conduct. The aim is to ensure that complex simulation outputs are shaped within this framework; geographical maps and The goal is to enable it to generate clear metric reports; without ever touching the actual customer, The aim is to enable the individual to see the possible consequences of their decision and to improve the quality of their decision-making. 10 Figures that will help understand the invention. Figure 1 shows the general architecture of the system that is the subject of the invention. Description of Part References 15 1. Geographic-Demographic Synthetic Customer Generator 2. Physical Dealer Network and Behavior Modeler 3. Customer-Dealer Interaction and Flow Simulation Core 4. Scenario Management and Dynamic Variable Injection Interface 5. CBS Integrated Visualization and Metric Engine 20 Detailed Description of the Invention In this detailed explanation, the preferred configurations of the invention are not merely for better understanding the subject. This is intended to facilitate understanding and will not impose any limiting effects. The invention, telecommunications, physical retail network management, data center regional marketing 25 strategy, pricing optimization and customer behavior simulation areas used, a telecommunications operator's physical dealer network and recorded connections to that dealer network By anonymizing customer behavior, completely synthetic data is obtained without the risk of disclosing real personal information. It is a system that models and simulates in a digital environment. The working principle of the system described in the invention is based on complex telecommunications dealer-customer interactions in the real world. Creating a privacy-guaranteed digital twin of the ecosystem and experimenting on this twin without risk. It is based on enabling this. The process is primarily based on Geographic-Demographic Synthetic. Customer Generator (1) and Physical Dealer Network and Behavior Modeler (2) of the actual historical subscribers By processing the data, statistically realistic but completely artificial 35 for each region and dealer. It begins with the construction of a "Synthetic Telecommunications Universe." This universe has no individual access to real data. 4 millions of unconnected synthetic subscribers and hundreds who will interact with them It consists of a synthetic dealer agent. This ready-made universe is a Customer-Dealer Interaction and Flow Simulation. It forms the basis of the core (3). User (marketing strategist), Scenario Management and Through the Dynamic Variable Injection Interface (4), you can set the tariff price in this province to this amount. What would happen if we changed it and added this many dealers? What if we, as XX dealers in Sivas, ran these campaigns? It defines a scenario such as "how much sales can we make?". This scenario is inserted into the simulation core. It is transmitted. The core loads the synthetic universe and applies the defined changes to the virtual environment. It moves time forward. Synthetic customers react to the new price and existing / new dealers, churn. They make decisions and initiate new subscriptions. All these interactions are governed by the dealer agents' code of conduct. It takes shape within this framework. When the simulation is complete, GIS Integrated Visualization and 10 The Metric Engine (5) is activated, converting complex simulation outputs into geographic maps and clear It converts these into metric reports. The user can make decisions without ever interacting with an actual customer. This allows them to foresee the possible consequences and improve the quality of decision-making. The elements and functions of the system described in the invention are as follows: 15 The Geographic-Demographic Synthetic Customer Generator (1) learns from real regional dealer data, statistically realistic but individually anonymous synthetic customers for each region. It creates profiles (subscribers). It guarantees privacy using Differential Privacy. It forms the realistic basis for the simulation. Physical Dealer Network and Behavior Modeler (2), using real dealer data, for each 20 synthetic dealer as an "agent" with specific behavioral rules and performance parameters models. They allow you to simulate how a dealer will shape customer interaction. This, This is the most critical distinction of the invention. Customer-Dealer Interaction and Flow Simulation Core (3), synthetic customers and dealers 25 how agents interact with each other and with external variables (campaigns, prices) A simulation engine. Agent-Based Modeling (ABM) to simulate customer decision-making processes, Discrete Event Analysis. It models events (visits, complaints, churns) in the system through simulation. Scenario Management and Dynamic Variable Injection Interface (4) allows the user to, on a geographical basis multiple marketing, pricing and dealer operation variables ("what if... what happens?" (define the question) and dynamically inject these variables into the simulation core. 30 It is the control panel and API that enable it to do this. CBS Integrated Visualization and Metric Engine (5) makes simulation results understandable. It visualizes this on geographic maps (e.g., which districts have the highest concentration of churn growth). (showing heat map). It also displays regional and total metrics (revenue, number of subscribers) Accounts and reports. 35 The functions performed by the system described in this invention are as follows: National subscriber data includes all subscriber records / purchase behavior and population statistics across dealerships. By processing data, we have created statistically realistic data for all 81 provinces and all districts of Türkiye. It creates a database consisting of millions of synthetic subscribers. By gathering all dealer location and performance / sales data nationwide, each actual dealer receives 5 synthetic dealer with specific behavioral parameters corresponding to its location and type It creates agents and registers these agents in a geographic database. If we increase the monthly package offer fee of Tariff Y in the Eastern Anatolia Region by 15% and the same If we increase the number of bids in the region, the regional churn rate and net income will increase over a 12-month period. Engine and scenario 10 that allows defining a scenario such as "how does it change?" It works. The described scenario is implemented by the Customer-Dealer Interaction and Flow Simulation Core (3). It is processed. The core loads synthetic customer and dealer data for the relevant region (Eastern Anatolia), price It injects variables such as growth and new dealer additions into the model and all processes like 12-month periods. It simulates the simulation scenarios step by step. 15 When the simulation is complete, CBS Integrated Visualization and Metric Engine (5) will display the results. jobs. On the map of the relevant region, the provinces with the highest churn growth and the new dealers are shown. The districts that gained the most subscribers are visualized with heat maps. At the same time, regional Numerical indicators such as the absolute change in churn rate (+2.1%) and the estimated change in net income (-5.3%) Metrics are reported. The reported data is also used in the development of synthetic data. It is saved to the data field.

Claims

6 REQUESTS 1. Telecommunications, physical retail network management, data center-based regional marketing. strategy, pricing optimization and customer behavior simulation areas used, a telecommunications operator's physical dealer network and recorded connections to that dealer network By anonymizing customer behavior, completely 5% without any real risk to personal data. It is a system that models and simulates in a synthetic digital environment; its characteristic feature is:  By learning from real regional dealer data, statistically analyzing data for each region. It creates realistic but individually anonymous synthetic customer profiles (subscribers), Using differential privacy to guarantee privacy and making the simulation realistic The geographic-demographic synthetic customer generator (1) which forms the basis of the real dealer data 10 By using each synthetic dealer, specific codes of conduct and performance modeling a dealer's customer interaction as an "agent" with parameters physical dealer network and behavior that allows simulation of how it will shape modeler (2), synthetic customers and dealer agents interact with each other and external variables Agent-based modeling that simulates how (campaigns, prices) interact. 15 (ABM) analyzes customer decision-making processes and system events using discrete event simulation. Customer-dealer interaction and flow simulation modeling (visits, complaints, churns). core (3),  the user has multiple marketing, pricing and dealer operations based on geography Define the variable (the question "what if...?") and apply these variables to simulation 20 a control panel and API that allows it to dynamically inject into the kernel scenario management and dynamic variable injection interface (4), simulation visualizing the results on geographical maps to make them understandable, the same calculates and reports regional and aggregate metrics (revenue, subscriber count) over time. CBS integrated visualization and metric engine (5) 25 It includes.

2. The system complies with Request 1 and its features include: national subscribers, all subscribers across dealer networks. By processing records / purchase behavior and population statistics, all 81 provinces and all districts of Türkiye. for, millions of synthetic subscribers, each with statistically realistic characteristics It includes a geographic-demographic synthetic customer generator (1) that forms a database. 30 3. The system complies with Claim 1 and its feature is; all dealer locations nationwide and By gathering performance / sales data, corresponding to each actual dealer location and type, creating synthetic dealer agents with specific behavioral parameters and these agents It includes a physical dealer network and behavior modeler (2) that records in a geographic database.

4. The system complies with Request 1, and its feature is; in the Eastern Anatolia region, the monthly package of tariff Y is 35. If we increase the bid fee by 15% and increase the number of bids in the same region, regionally 7 A scenario like this: "How do churn rates and net income change over a 12-month period?" Customer-dealer interaction that runs engines and scenarios that allow for its identification. and includes the flow simulation core (3).

5. The system complies with Claim 1, and its features include: the defined scenario, customer-dealer interaction, and Synthetic 5 of the relevant region (Eastern Anatolia) operating by the flow simulation core (3). The model loads customer and dealer data, and includes variables such as price increases and the addition of new dealers. injecting and simulating all simulation scenarios step by step, such as 12-month processes It includes customer-dealer interaction and flow simulation core (3).

6. The system complies with Claim 1 and its feature is that it processes the results when the simulation is complete. On the map of the relevant region, the provinces with the highest churn growth and the top 10 new dealerships are shown. visualizing the districts that have gained the most subscribers using heat maps, and also regionally... absolute change in churn rate (+2.1%) and estimated change in net income (-5.3%) reporting numerical metrics and also developing synthetic data based on the reported data. CBS integrated visualization and metrics engine (5) that saves to the blend data field. It includes. 15

Images