Shulou(Shulou.com)11/24 Report--

After many enterprises launch the BI and complete the basic data display, they are often confused about the follow-up construction path of BI and do not know where to start, or do a lot of BI development with great fanfare, resulting in a fierce operation and a gripping business evaluation.

The core value of BI lies in the use of enterprise data assets to assist business decision-making, and it is an important tool to realize the value of enterprise data assets. Like enterprise management, a more reasonable BI construction path should also be designed around the goal of ROI maximization, but how to evaluate the ROI of BI construction has been a difficult problem for a century.

Guanyuan data refines a set of simple analysis framework by summarizing the experience of serving advanced customers in various industries for many years and the phenomena observed from some of the customer samples, so as to try to provide reference BI construction path suggestions for different types and stages of enterprises from the perspective of ROI.

I. the classification of BI value and its influencing factors.

The ROI level of BI construction depends on both value and cost. First of all, analyze the value aspect.

The core value of BI lies in the use of data assets to assist business decision-making and improve decision-making efficiency.

According to the summary of past practical experience, we refine the two kinds of values of BI construction to business decision-making, divide them into basic value and enabling value, and summarize two major factors that have great influence on the data enabling value.

1. Basic value

The most basic data display, which enables business departments to see business data from different sources more easily, is the basic function of BI tools, which has a certain value, is the easiest to achieve, and is most likely to produce "aesthetic fatigue", which belongs to the "drooping fruit" in the construction of BI.

2. empowering value

✦ performance growth (explicit value)

Performance growth generally refers to business results, and the result indicators in the corresponding index system (GMV, profit, cash flow, number of users, etc.) are the most concerned and dominant value of all for-profit organizations, and are also given priority in the process of data construction.

At the same time, this kind of value is also the most difficult to evaluate and verify. On the one hand, because performance growth is usually the result of the joint action of many factors, it is difficult to evaluate the value of a single factor in data construction; on the other hand, because there is usually a clear division of roles in enterprise organizational planning, data construction positions are usually middle and back-end departments, not directly responsible for performance results, resulting in lengthy data value chain and high traceability cost. Some enterprises use the way that the business department binds the goal of the data department, let the data serve the business directly, and reflect the data value by shortening the value chain.

✦ business efficiency (hidden value)

Business efficiency generally refers to the process management in the business process, which corresponds to the process indicators in the index system (conversion rate, ROI, turnover rate, quality rate, etc.). The upper limit of this kind of value is usually limited, and it is usually not the most priority concern in the rapid growth stage of the enterprise, but it will become an important concern in the stage when the enterprise develops steadily and begins to seek internal benefits.

Paradoxically, data construction itself is also a work that requires the input of enterprise resources. at the stage when the growth rate of enterprises tends to stabilize and begins to seek internal benefits, the investment in data construction will also be affected. From this point of view, whether enterprises can layout their own digital construction in advance in the stage of rapid growth, match the enterprise life cycle and invest data construction resources reasonably, which is a great test of the data awareness and strategic pattern of the enterprise decision-making team.

3. Influencing factors

The basic data display value is very limited to the actual business activities of the enterprise, and it is more used as a tool to enhance the internal data awareness and show the digital development level of the enterprise. the form is greater than the content, and there is little difference among different enterprises. however, the data enabling value of different types of enterprises is very different, and the main influencing factors are as follows:

The matching degree between the data assets available to ✦ and its important business activities

Take the "repurchase rate" index as an example, if the goods or services provided by an enterprise have a high repurchase value, the repurchase rate will be an important index to measure its business health, and user operation is one of its important business activities; user data is easier to obtain than external latent customer data, as long as the construction of user-related data assets can quickly play a data value in the user operation link.

On the contrary, if the demand for repurchase of goods or services provided by an enterprise is low, the growth value of the data of regular users to its core business is limited, and external data that are more dependent on performance growth are difficult to obtain, it will cause a mismatch between data assets and important business activities, and it is difficult to give full play to the value of data assets.

The dependence of various decision-making processes on data assets in ✦ business activities

Overall, the order of dependence of different types of decisions on data assets is as follows:

High-frequency decision-making is higher than low-frequency decision-making: for example, mergers and acquisitions, new business development, organizational structure adjustment and other types of decisions usually belong to low-frequency decisions, with little reference significance for historical data, while investment channel selection and user operation strategy adjustment usually belong to high-frequency decision-making, which can be used to discover rules and guide business actions through historical data.

Standardization decision is higher than customization decision: standardization decision means that the logic of decision link is clear and can be described, such as store goods replenishment decision; customization decision refers to complex decision link and many factors to be considered, such as quotation decision of large IT project; usually, the proportion of customization decision of 2B enterprises is higher than that of 2C enterprises.

Analytical decision-making is higher than organizational decision-making: the core difference between the two lies in the proportion of "human perceptual factors" in this decision-making activity.

The proportion and value contribution of various decision-making types in business activities are the underlying driving factors for enterprises to attach importance to data construction. The importance of data to the core links of business activities determines the ceiling of the contribution of BI construction to enterprise value.

II. Classification and influencing factors of BI construction cost

The other end of ROI is cost. BI construction needs to be based on data construction. The cost of data construction is included here, which can be roughly summarized as data acquisition cost, data governance cost, and data application cost (BI construction is a part of data application). The corresponding influencing factors include:

1. Data acquisition cost

Influencing factors: external data, or internal data, To B or To C; online, or offline.

The cost of obtaining internal data is usually lower than that of external data, but there is a special case, that is, e-commerce brand enterprises. Due to the rapid development and digital construction of domestic e-commerce platforms (JD.com, Amoy, Douyin, etc.), e-commerce enterprises can rely on the transaction system capability of the e-commerce platform to obtain transaction data without building a transaction system. the ownership of this part of the data is the e-commerce brand, but it comes from the external e-commerce platform system. It is precisely because of this characteristic of e-commerce brand that many new e-commerce enterprises can still obtain very important transaction data when their own information level is very weak, and then enlarge the ROI of BI construction, which makes the development path of "digitization first, informationization later" possible.

The essential problem corresponding to the difference between business models (2B / 2C, online / offline, etc.) is to see the degree of dependence of business activities on informatization. The more dependent on online, 2C enterprises (such as retail banks), the higher the demand for information level (enterprises with insufficient information level have been eliminated), and the larger the scale of available data assets (regardless of asset quality). The lower the cost of data acquisition. On the contrary, the more we rely on offline and 2B enterprises, the lower the importance of their informatization level. If we want to obtain sufficient data assets, we need to improve the informatization construction (its necessity for business activities is not high). That is to say, it increases the relative cost of data acquisition in disguise.

2. Data governance cost

Influencing factors: one / two / three party data; the degree of business closed loop.

The impact of data sources on data governance costs, we still take the above e-commerce enterprises as an example: although e-commerce brand enterprises that rely on e-commerce platforms to complete transactions do not need to build their own transaction systems, and the cost of obtaining transaction data is low, however, because its data source comes from the platform and is two-party data, it will lead to an increase in the cost of data governance.

The influence of business closed-loop degree on data governance cost, we take manufacturing enterprises as an example: manufacturing enterprises with self-built factories can obtain supply data by improving production information system. Enterprises that rely on contract manufacturing can obtain supply data through the way reported by contract manufacturing enterprises, and each has a certain data acquisition cost. However, once the enterprises with self-built factories have completed the informationization of the production system, the cost of data governance will be much lower than that of enterprises that rely on OEM.

3. Data application cost

Influencing factors: data awareness; regulatory restrictions; data tools; organizational structure …...

The analysis in this part can be used independently as a topic, briefly summarizing the following points:

Generally speaking, when enterprises are affected by their own business model, enterprise information development, internal data culture and other factors, the cost of comprehensive data construction is low enough, it will improve the ROI of BI construction accordingly.

It is worth mentioning that when considering the cost of data tools, in addition to the "explicit costs" such as the procurement cost and deployment cost of the tool itself, the ease of use and maintainability of the tool itself will also affect the "hidden cost" of the enterprise, such as learning cost, construction cost, maintenance cost and so on. If the tool is difficult to use or maintain and does not play its due value, the waste of opportunity cost is usually huge.

Third, the subdivision path of enterprise ROI construction on the premise of reaching the BI standard.

The core value of BI is to improve the efficiency of decision-making, but not all decisions are most suitable to be solved by BI, and there is not a set of BI construction methodology suitable for all enterprises. Guanyuan data tries to provide a simple analysis framework to judge what path different enterprises should refer to and in which links to develop BI.

1. What kind of decision scenarios is more suitable for BI?

Similar to the matching relationship between the value of data assets and decision types, BI is more suitable for high-frequency, standardized, analytical decision-making scenarios.

✦ High Frequency decision scenario

If the frequency of decision-making is not high, such as quarterly and annual overall business analysis scenarios, even if there is no BI, analysts can use other analysis tools (Excel, Python, etc.) to meet the analysis needs, or even more flexible. However, if this kind of comprehensive analysis can be realized on BI, it shows that the quality of data assets within the enterprise is high, and it can also drive the managers of various departments to pay attention to the construction of BI, which is conducive to promoting the overall popularization of BI.

✦ standardization decision scenario

If the standardization of the decision scenario is not high, such as a targeted analysis of an occasional event, you may need to rebuild the data model, clean a large amount of "dirty data", and rebuild the BI page, but the reuse value of this scenario is limited, and the ROI implemented with BI is low. However, if the demand frequency of customized analysis in the enterprise is high, the difficulty of analysis is low, and the data assets are relatively complete, it is more suitable to develop BI by popularizing "self-help analysis".

✦ analytical decision-making scenario

Typical analytical decision-making has the characteristics of quantification, traceability, easy attribution and short decision chain. Action decisions can be made directly according to the results of data analysis, such as the promotion decision of fresh stores, which can dynamically adjust prices according to the temporary situation and inventory situation of different fresh products. After the algorithm rules are clear, the action suggestions can be directly pushed to the store manager through the BI application to guide the promotion action of the staff. This kind of scenario is suitable to bind the data flow to the business flow and realize the closed loop of data-analysis-action-feedback through the production-oriented BI application.

2. How to make a better BI construction path.

Ideally, the construction path and promotion mode of BI need to comprehensively consider the level of data assets in different stages, the proportion and importance of different types of decisions, organizational structure and decision-making context targeted planning.

✦ construction path

The answer to the construction path is "do what?" Here is a reference idea:

① first combines the enterprise's business model and strategic objectives to find out what are the key business activities.

② identifies the main decision types (high frequency or low frequency, standardization or customization, analytical or organizational) in these key business activities, and combines the available data assets to identify those decision scenarios that are more suitable for BI enabling and data assets available as priority BI construction goals.

③ keeps repeating ① and ②.

Of course, in reality, we should also take into account the construction planning of data assets in the enterprise, the priority of the needs of different business departments, management preferences and other factors to develop a specific BI construction plan.

✦ promotion mode

The answer to the promotion is "who will do it?" The concept of "decision context" is introduced here:

Observe the "decision-making context" of the enterprise, that is, examine how the "decision rules" of different levels and positions in the enterprise are made as a whole, and find out the scenarios that are suitable to be empowered by BI. If the way of making decision rules is more centralized (suitable for "across-the-board" approach), it is more suitable to adopt the way of centralized construction of BI; if the formulation of decision rules is more decentralized (suitable for "people who hear the sound of gunfire to make decisions"), then it is more suitable to use self-help analysis to develop BI.

Generally speaking, the more focused the type of business, the higher the degree of standardization of goods or services provided, and the smaller the size of personnel, the more concentrated the context of decision-making, such as the cutting-edge e-commerce enterprises of a single brand; the more diverse the types of business, the higher the degree of customization of goods or services provided, and the larger the scale of personnel, the more scattered the context of decision-making, such as global multi-brand super-large consumer goods enterprises.

3. Other factors affecting the construction path of BI.

In reality, the construction of BI, like other things, is affected by multiple factors such as climate, geography and human harmony.

Conclusion: the "potential, Tao and skill" of BI construction.

The BI construction in the enterprise is affected by various factors. Whether the BI builders or BI manufacturers in the enterprise often feel confused in the process of serving customers, which is the priority of the different factors, how to set the reasonable goal of BI construction and find the appropriate path? in summary, it can be summarized according to the structure of "potential, Tao, skill":

Potential: enterprise business model, social information technology development level and other factors are the highest dimensional factors, which define the difficulty level of BI construction to a great extent. at the same time, the iterative cycle of these factors is long and difficult to change in a short time, so they can only comply with the trend from the point of view of BI construction.

Tao: reasonably evaluating the ROI of BI construction, paying attention to the matching of data flow and business flow and decision context is the underlying methodology of BI construction.

Technology: specific work such as information construction, data asset construction, data governance, BI technology architecture, data department construction and management, BI project construction, BI self-help analysis and promotion all belong to the category of "technology" and need to be used flexibly according to the respective stage characteristics of different enterprises.

In the future, with more diverse and rich customer samples, Guanyuan data will continue to verify the rationality and prediction accuracy of the theoretical framework. We also look forward to discussing with you to explore new ways to improve enterprise BI and build ROI, so that data can create more value for business growth.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.