Paper Sharing:Theoretical Foundations of Data Assetization Governance

Previously, we shared the first installment on the theme of "Building Data-Driven Business Ecosystems." Today, we present the second installment, focusing on the core topic of "Data Assetization Processing." This session provides a comprehensive overview—from theoretical foundations and problem analysis to practical applications—to help you gain valuable insights. We hope you find it beneficial.

1.1 The Connotation of Data Assets

To understand the concept of data assets, it is essential to start with their distinction from ordinary data. According to policy documents from China's National Data Administration, data assets are not merely digital files stored on hard drives. They are information resources legally controlled by enterprises that can generate tangible economic benefits. For example, user browsing records on an e-commerce platform may initially appear as raw data. However, after undergoing desensitization and analysis, they can guide merchants in adjusting marketing strategies—transforming into data assets capable of creating value. This process is akin to refining iron ore into steel: data resources must undergo steps such as cleaning, processing, and rights confirmation to evolve into assets.

Data assets can be categorized into three tiers based on the depth of processing:

Basic data serves as the raw material, such as raw temperature data collected by sensors.

Processed data functions as semi-finished products, like annotated image datasets.

Applied data represents the final product, such as precision recommendation models generated from user behavior.

This classification facilitates orderly data management by enterprises and helps prevent "asset sinking"—a scenario where vast amounts of data remain idle on servers, neither generating value nor justifying storage costs. For instance, a manufacturing firm accumulated 20TB of production data over five years without clear applications until it created a data asset inventory. This revealed that 60% of machine operation data could be used to optimize equipment maintenance protocols.

1.2 The Characteristics of Data Assets

The most prominent characteristic of data assets is their "multifaceted nature."

Non-exclusivity in sharing: The same customer data can be used for both product R&D and market analysis, enhancing data utilization efficiency. However, this also complicates rights attribution.

Dynamic value appreciation: Similar to brewing wine, the value of data assets grows over time and with expanded application scenarios. For example, a bank’s risk control model initially designed for credit card approvals was later found applicable to small business loan assessments, doubling its value.

Scenario dependency: Medical imaging data is a critical asset in diagnostic contexts but holds negligible value in advertising targeting. This necessitates that companies identify "the scenarios where data delivers the highest value."

The privacy attribute of data assets is a double-edged sword. An internet hospital aiming to share medical data must simultaneously protect patient privacy and extract research value. Privacy computing technology achieves this by enabling "data usability without visibility." This is particularly evident in the financial sector: credit data is used for loan approvals but must not be misused. Consequently, the concept of data sovereignty credentials has emerged—akin to equipping data with GPS, enabling usage tracking and rights revocation.

Another challenge lies in circulation barriers. In manufacturing, design and production departments often face "data dialects" (mismatched data formats), creating information silos that hinder cross-department collaboration. Cross-organizational circulation is even more complex, with high compliance costs. For example, two companies sharing supply chain data may spend months resolving data ownership disputes. Here, blockchain notarization technology proves invaluable, acting as a tamper-proof "black box" for data transactions that ensures transparency and reduces disputes.

A deep understanding of these characteristics is crucial for enterprises to establish a data asset governance framework. Beyond data collection, companies must learn to "label," "value," and "market" their data—much like a pawnshop skilled in receiving items, appraising worth, and finding buyers. Only by comprehensively mastering data asset traits can enterprises:

Design intelligent decision-support systems with accurate key metrics;

Develop dynamic pricing models for market-oriented data trading mechanisms.

Data asset governance is not merely a technical task but a transformative battle requiring both commercial acumen and technical expertise.

2. The Construction Logic of a Data Assetization Governance Framework

Establishing a data assetization governance framework equips enterprises with a standardized system to manage data resources in the digital economy era. It addresses the transition from chaotic to orderly data management and unblocks bottlenecks in the market-oriented circulation of data elements. To illustrate, consider data as oil buried underground—the governance framework is the complete infrastructure for exploration, extraction, refining, and transportation, where every component is indispensable.

Necessity of the Framework

Many enterprises invest heavily in data platforms only to encounter "data dialects" (incompatible systems) that prevent communication between business units, resulting in data silos. Approximately 70% of data assets lie dormant, generating no economic returns while occupying storage space. More critically, circulation barriers have led to a "loud thunder, little rain" scenario in data element markets—akin to building a marketplace with no vendors. Surveys indicate that supply chain delivery delays in manufacturing due to data incompatibility can reduce overall profits by 30%.

Logical Starting Point: Resolving Rights Attribution

Traditional asset rights are relatively clear, but data asset rights are fluid, like honey—difficult to define and prone to spillover. Inventory management becomes a pivotal solution, assigning a "digital ID" to each data asset that records details such as collection time, usage permissions, and circulation paths, enabling end-to-end tracking. For example, an e-commerce platform uses blockchain notarization to divide user behavior data rights into three tiers: original data ownership to users, desensitized data usage rights to the platform, and shared revenue rights for processed data. This transforms data assets into standardized, tradable products.

Technological Support: Intelligent Decision Systems

Traditional market decisions resemble gambling in casinos—highly speculative. In contrast, data-adaptive optimization technologies function like autonomous vehicles, dynamically adjusting strategies. A bank’s real-time decision-support system reduced credit approval times from three days to 15 minutes by leveraging algorithms to identify data value density. When user-provided data reaches a critical volume, the system automatically initiates approvals. Such breakthroughs accelerate decision-making and create a virtuous cycle: "The more data is used, the greater its value."

Market Mechanisms: Dynamic Pricing Models

Designing market transaction mechanisms requires meticulous planning. Dynamic pricing models, akin to "smart calculators," adjust prices based on data freshness, application scenarios, and market demand. A medical tech firm facilitating cross-organizational data circulation used privacy computing to ensure "data usability without visibility," safeguarding patient privacy while quadrupling the value of tertiary hospital clinical data. This compliant, risk-controlled system update transformed data commercialization from a high-risk experiment into a replicable, mature model.

In summary, constructing a data assetization governance framework resembles implementing a "South-to-North Water Diversion Project" in the digital realm. It requires breaking down data silos, unblocking circulation bottlenecks, and establishing fair market pricing. Technology breakthroughs provide the driving force, institutional innovation sets the direction, and market mechanisms fuel the process—all three are indispensable. When enterprises master transforming data into assets, managing for efficiency, and utilizing for value, the journey toward data element marketization has only just begun.

3. Root Cause Analysis of Data Silos and Circulation Barriers

The causes of data silos and circulation barriers can be categorized into four core dimensions: technology, management, policy, and culture.

(1) Technical and Standard Barriers

Divergent technical architectures (e.g., database models, storage methods, interface protocols) across departments/systems, lacking unified data standards and exchange protocols.

Inconsistent data formats and interface standards between internal systems (e.g., production vs. sales).

Generational gaps between legacy systems (e.g., traditional databases) and modern platforms (e.g., cloud computing).

(2) Organizational and Management Mechanism Obstacles

Departmental self-interest: Some units treat data as "private property," withholding it due to resource control or profit concerns.

Government reluctance: Public agencies fear data sharing may erode jurisdictional authority, reducing motivation.

Process fragmentation: Isolated internal workflows, absent cross-department collaboration mechanisms, and fragmented data governance strategies lead to data accumulation.

Cultural and institutional conflicts: Large enterprises struggle with data flow due to cultural differences and systemic contradictions.

(3) Policy and Security Conflicts

Regulatory lag: Ambiguous policies fail to define data-sharing boundaries, causing departments to "hesitate to share" (e.g., unclear public data authorization responsibilities) or "lack the capability to share" (e.g., inconsistent classifications of important vs. general data).

Privacy and security concerns: Data openness risks privacy breaches, especially for personal/sensitive information. The absence of anonymization standards creates a dilemma between sharing and protection.

(4) Governance and Cultural Ideological Clashes

Incomplete governance systems: Lack of unified data lifecycle management norms leads to inconsistent data quality and unclear rights/responsibilities (e.g., absent ownership/authorization mechanisms, forming "logical silos").

Conservative culture: Traditional management views data as a power symbol, fostering a closed, non-collaborative culture. Some institutions even perceive sharing as a loss of competitive advantage.

Integrating these factors reveals that resolving data silos requires addressing technical, managerial, policy, and cultural dimensions. Solutions include:

Creating unified data exchange protocols and classification/grading standards.

Implementing incentive policies for cross-department collaboration.

Leveraging technologies like trusted data spaces to ensure secure data sharing.

4. Pathways for Data Rights Attribution and Inventory Management

In the reform of data element marketization, defining data rights and implementing inventory management are pivotal to resolving data silos and circulation barriers. Unclear data rights deter inter-enterprise sharing and increase transaction costs, while inventory management provides technical support for data asset visualization and standardized circulation.

Methodology for Data Rights Attribution

Adhering to a "classified attribution, dynamic adjustment, scenario adaptation" framework, rights are assigned based on data originators and value contributions. This approach has become mainstream:

China’s "Provisional Regulations on Accounting for Enterprise Data Resources" divides enterprise production data along the value chain ("raw data → derived data → fused data"):

Original data producers retain foundational rights.

Data processors hold usage rights for derived data.

Composite rights for fused data are allocated via contracts.

South Korea’s K-Data Project employs a tiered attribution model with a "data contribution index," assigning equity weights (35% for transaction data, 25% for logistics data, 40% for credit data) to supply chain enterprises. This improved cross-border trade data sharing efficiency.

Dynamic Attribution Mechanisms

Incorporating time and scenario variables, the European Data Act’s "rights decay function" links data rights validity to value half-life. When data timeliness falls below a threshold, rights are automatically reset—a model successfully applied in financial risk control data.

Building an Inventory Management System

This requires technological and institutional synergy:

Blockchain notarization provides an immutable registration system for data asset inventories. Smart contracts enable automatic updates to data directories and rights changes.

Example: A Chinese automotive group’s supply chain data hub, built on Hyperledger Fabric, encoded 162 data asset types from 12,000 suppliers, creating a dynamic inventory.

Metadata management (MDM) resolves heterogeneous data standardization challenges by constructing a "data fingerprint" system to classify and encode structured/non-structured data.

Example: Shanghai Data Exchange’s pilot "data passport" system uses NLP to extract 128 metadata items (e.g., source, format, sensitivity), boosting data asset identifiability by 76%.

Operational Implementation: A "Trinity" Framework

Foundation Layer: Data lineage analysis tracks the entire data lifecycle, creating end-to-end data maps.

Middle Layer: Knowledge graphs build data asset directory trees for multi-dimensional retrieval and correlation analysis.

Application Layer: Visual dashboards enable dynamic monitoring of data asset status.

Medical data sharing exemplifies this: A tertiary hospital’s "data asset cockpit" system categorized 12PB of medical imaging data into 23 dimensions (e.g., examination type, diagnosis, patient traits), reducing data retrieval response times to seconds. Privacy computing ensured secure access to inventory information.

5. Practice: Synergistic Innovation of Data Governance Technology and Institutions

In the implementation of the data assetization governance framework, the collaborative innovation of technology and institutions acts like installing "dual engines" for the data element market. However, some enterprises have invested heavily in data governance, only to exacerbate issues such as data silos and asset idling due to the lack of alignment between technological tools and institutional rules—akin to purchasing state-of-the-art agricultural machinery without knowing how to farm. This raises a critical question: how can technology truly serve governance objectives?

Privacy Computing: A Breakthrough in Balancing Security and Sharing

The significant advancements in privacy computing offer an important solution to this challenge. Dubbed the "safety valve for data circulation," this technology enables value exchange while ensuring raw data remains within its original domain through cryptographic algorithms. For instance, in medical data sharing scenarios, tertiary hospitals and research institutions can jointly develop disease prediction models using federated learning technology, mitigating privacy risks while generating scientific value. This technical approach resolves the long-standing dilemma in traditional data transactions—where data security and usability were often mutually exclusive. More importantly, it provides a new perspective on rights attribution: when the transfer of data usage rights no longer relies on physical duplication, the design of attribution mechanisms can focus more on value distribution.

Nevertheless, the deployment of technological tools still requires supporting institutional frameworks. For example, a provincial medical insurance bureau piloted a "data sovereignty credential" system, leveraging blockchain notarization to record critical information such as data access permissions and usage logs on-chain. This not only enabled dynamic updates to inventory management but also provided a reliable basis for subsequent benefit distribution.

Institutional Innovation: Transforming Abstract Rights into Operational Processes

Institutional reforms have also yielded notable progress. The pilot "data asset registration system" represents a breakthrough by converting abstract rights attribution into actionable registration procedures. Take an automotive manufacturing enterprise as an example: by packaging its accumulated autonomous driving data into standardized formats and registering them on a government-established data element market, the enterprise obtained a legitimate "passport" for circulation. Furthermore, it leveraged dynamic pricing models to realize value appreciation of its data assets. This institutional arrangement effectively established a conversion channel from data resources to data assets, addressing circulation bottlenecks. However, practical challenges emerged—for instance, an e-commerce platform once faced disputes over the valuation of user behavior data, where the same dataset could be priced differently by orders of magnitude depending on whether it was used for precision marketing or macroeconomic analysis. This highlights the need for more granular institutional arrangements to coordinate stakeholder interests.

Collaborative Innovation: Achieving "1+1>2" Synergy

Case studies demonstrate that the integration of technology and institutions often yields synergistic effects exceeding the sum of their parts. For example, a commercial bank adopted data-adaptive optimization technology through its intelligent decision-support system, reducing credit approval time from five business days to real-time decisions. This breakthrough stemmed not only from machine learning algorithms mining customer data deeply but also from regulators' compliance approval of the new credit risk model. Such collaborative innovation lowers compliance costs for enterprises and fosters new business models—the bank provided its risk assessment capabilities to smaller financial institutions via API interfaces, creating a cross-organizational data circulation ecosystem. However, the success of this model hinges on the reliability of real-time decision-support systems: last year, a fintech platform experienced a mass loan rejection incident due to model bias, underscoring that efficiency gains must not come at the expense of ethics.

Currently, the most noteworthy trend is the spiral development of "institutional technologization" and "technological institutionalization":

Expansion of technical standards: Privacy computing standards have evolved from internal corporate applications to industry-wide specifications.

Integration of technical elements: Data market entry requirements now incorporate emerging technologies like blockchain notarization.

This deep integration is reshaping the foundational logic of data governance: previously idle data assets can now be monetized automatically through smart contracts, while the rigidity of traditional inventory management is being replaced by dynamic, visualized processes enabled by digital twin technology. However, achieving substantive breakthroughs requires a fundamental shift in mindset. As one data administration official noted: "Data governance should not merely focus on 'control' but on 'activation'—letting technology serve as a catalyst for institutional innovation, not a constraint."

Deeper Implications: Building a "New Production Relationship" for the Digital Era

Fundamentally, data governance constructs a "new production relationship" for the digital age. When privacy computing resolves trust issues in data circulation and data sovereignty credentials clarify benefit distribution rules, the data silos that once hindered value realization will naturally dissolve. This aligns with the core objective of the National Data Administration’s "three flows" (data flow, value flow, and ecosystem flow) initiative—using technology and institutions as dual drivers to mobilize data elements, generating multiplier effects far surpassing those of traditional production factors.

Yet, the path ahead remains challenging. Advancing data assetization demands both "hard power" (e.g., breakthroughs in secure privacy computing algorithms and efficient dynamic pricing models) and "soft environment" (e.g., flexible regulatory frameworks and equitable benefit distribution mechanisms). Only through the synergistic advancement of both can obstacles in data assetization be overcome, fully unleashing the potential of data elements.

Today, we introduced the connotation and characteristics of data assets, the construction logic of the governance framework, as well as an analysis of the causes of issues and practical cases. In the following sessions, we will delve deeper into the practical exploration of data asset governance.