Data Innovation Lab

Connecting the “internal and external perspectives” to build a holistic data strategy

As digital transformation accelerates and data acquisition becomes easier, corporates begin to recognise their past decision-making deficiencies and look to Big Data as the cornerstone of their future business capabilities. However, the lessons of the last decade have proven that getting the right Big Data mindset may be more important than rushing to find a sure-fire formula.

Corporates without a top-level design on the Big Data development that fits and adapts to a data strategy, data drivenness is likely to result in a partial win but overall failure.

If you are considering driving the digital transformation of your business, it is advisable to start by thinking about the following issues:

• What are the core business-related decision processes in the current state of the business in urgent need of optimisation?

• What kinds of decisions are suitable using a data-driven approach to improve accuracy and consistency?

• What kind of long-term value will the accumulation and development of data resources bring to the corporate?

• How do you measure and evaluate the value generated by digitalisation for your business?

• Is there an “implementation roadmap” to follow in the digital transformation process?

• What will happen to the management structure (organisation, people) before and after the digital transformation?

In the 21st century, digital transformation is not an option but a must for corporates wishing to improve their innovation capabilities.

Often, digitalisation is seen as an essential prerequisite for the future application of the Internet of Things (IoT), artificial intelligence and Big Data analytics in the enterprise. So, many corporates agree that digitalisation is an essential aspect of competitiveness in the future.

However, the truth is harsh: the success rate of digital transformation in corporations is not high. According to a survey report, more than 80% of digital transformation projects fail to achieve the expected results, which means there are deviations from planning to execution. Are the difficulties that corporates encounter in the process of digital transformation more technical or managerial?

In my experience, the problems in management are particularly evident in the following three areas:

• Lack of clear direction. Digital transformation, although identified by executives as one of the strategic objectives, has not been accurately articulated as an overall corporate endeavour. Digital transformation is long-term planning and must be complemented by a top-level design and implementation roadmap to strengthen the relevance of the business vision to the medium and long-term business.

• Lack of cross-functional resources. Most business leaders tend to focus on taking the lead in technology while neglecting that digitalisation is also a management reengineering process. Digital transformation efforts require strong management support, in addition to breaking down organisational structures that led to ‘data silos’ and introducing cross-functional data governance team work and technology platforms.

• Lack of readiness to adapt to change.  The goal of digital transformation is to lead companies to a new digital business model. Still, it also leads to a new way of internal operation and coordination, with implications for the relationship within the human organisation (including future human-machine collaboration) and a definite impact on the existing corporate culture. These impacts will intensify as the digital transformation deepens. It is important to note that the effects of the transformation may catch you off guard.

To ensure digital transformation success, leaders must have a deeper understanding of motivating employees during the transformation process.  Remember that transformation does not happen overnight; clear goals and sustained endurance are essential throughout the transformation process.

Most fatal: the absence of a corporate data strategy in the digital era

As data strategy becomes increasingly important in the strategic planning of corporations and governments, one of the critical points for successful digital transformation is to identify business needs and then use them to determine how data will be used in the data’s “Aggregation-Interconnection-Application” cycle. Here are some of my experiences:

• Establishing the scope of aggregation: effectively aggregating internal and external data, where the “aggregation” includes the import and merge of internal and external data. However, due to the external nature of Big Data, it isn’t easy to control the data’s integrity, authenticity and immediacy, so the ability to control the aggregated data must be enhanced, and the scope of data to be aggregated must be first determined.

• When there’s a blockage, it “hurts”. The “Connection” aspect is at the heart of data governance and has the role of carrying out the “aggregation” and initiating the “interconnection”. Data should be managed in such a way as to facilitate the flow of data, thereby facilitating the incubation of different levels, themes and details of data and improving the efficiency of data use. At the same time, this data interconnection has the meaning of connecting “Vertical and Horizontal function” internally within the organisation as well as connecting the organisation with the “external” world. I will elaborate on this point later.

• Data governance is for better application: The data application process is the best way to verify the effectiveness of the data “aggregation” and “interconnection”. This point is the best moment to introduce a data facilitation platform (DFP), which links data management and application to improve the efficiency of data governance, development, and usage efficiency.

• Optimise iterations with the outcome: one is simply the assessment of the effectiveness of data application; the other is whether the data’s “Aggregation-Interconnection-Application” architecture meets the long, medium and short term business needs. The key is to establish a mechanism to measure the continuous optimisation of “Aggregation-Interconnection-Application”.

The direction of the data strategy is to improve the ability to transform enterprise resources into organisational optimisation, smart business decisions, automated productivity improvements and ultimately, business transformation. The above data strategy framework of “Aggregation-Interconnection-Application” and “Application-Interconnection-Aggregation” must be based on the business strategy, corporate culture and business need through continuous communication between internal functions and promotion to all employees.    In the end, define the scope of data governance to achieve a strategic vision for data.

As you can imagine, the absence of a data strategy for business operations is akin to driving around an unfamiliar city without a map for guidance.

The essence: the way of data strategy for digital transformation in the enterprise

When talking about data strategy, the first question that needs to be answered is: What exactly does a company’s data strategy mean? Is it about increasing the stock of data assets and making up the reserve, or is it about filling the gaps in data technology? Or is data governance not being taken seriously, or are the organisation and culture not keeping pace?

For enterprises at different stages, these questions have very different implications. The strategic direction of data assets must first be determined from a knowledge of how data generates value, and management must have a deep understanding that data is a core asset for the future so that they can appreciate how to cut costs internally and actively develop resources externally to “accumulate” the data.

Don’t talk about data strategy until you know the direction of your business. When we were working at Alibaba, we had a saying, “if you are on the right track, you are not afraid of going far”. Alibaba’s data transformation process is also a “keep on fighting despite the continual setback” journey.

In Alipay, for example, the company understood before 2010 that establishing the practice of accumulating customer transaction and behaviour data for KYC (Know Your customer) as strategic direction. Still, it took six years to develop the data-driven capabilities finally. Throughout this process, the business strategy and the data strategy have been mutually supportive and coordinated.

A data strategy should focus on how data resources can be used in conjunction with technology and relevant business capabilities to maximise support for the high growth of the business. There are three ways in which data can drive support for business operations.

● Assisting operation staff in making accurate decisions through analytical tools, such as making forecasts for sales and optimising inventory.

● Using Big Data to optimise or automate production processes.

● Developing smart or automated products targeted directly at customers, such as chatbots.

In a broad sense, corporate data strategy encompasses business “efficiency and cost reduction”, the exploration of innovative business, and the accumulation of data that can help strengthen competitiveness and enable external cooperation.

From the perspective of a corporate strategist, these aspects should all be included in the framework of board and leadership discussions. The appropriate entry point should correspond to the stages, urgency and forms of corporate’s digital mutuality status.

Why do companies need a digital blueprint or data strategy? At a macro level, the value of a digital blueprint is to enhance the ability to coordinate between core data functions from a top-level design perspective to improve the competitiveness of the business that supports the data infrastructure.

The deployment of a digitalisation blueprint has great strategic significance and facilitates enterprises to find the best entry point from the current deficiencies. The digitalisation blueprint cannot be built without an understanding of the business strategy, clear planning of the data strategy, horizontal synergy (Cross Function or business unit) in the strategy implementation and execution process, and an adapted organisational capability and a sound mechanism for regular review and revision of the data strategy.

Here we propose an overall development plan of the digitalisation blueprint and its critical framework (see Figure 2-2).

Note: The meaning of the key steps in the figure:

1. “One strategy” is derived from the business strategy’s requirements and needs for the digitalisation of the enterprise and is one of the best supports for the realisation of the business strategy.
2. “Internal and external perspectives” refers to the outside-in and the inside-out viewpoint.
3. “Four key modus operandi” refers to the implementation based on data resources, data technology, organisational adaptation, and incentive mechanism establishment
4. “Overall integration” approach is to build a data-driven organisational structure, i.e., achieve horizontal resource synergy and joint decision-making across business units (BUs) through the Data Management Committee, and oversee the implementation effect. The Committee needs to implement the top-down approach, such as the “Aggregation – Interconnection – Application” mindset, or the bottom-to-top approach, such as the “Application – Interconnection – Aggregation” mindset.

In the future, the competitiveness of enterprises depends on their data-driven capabilities, so the strengthening of computing power, the mastery of data resources aggregation/utilisation methods and the training of business/technical personnel will all become key factors in determining success or failure.

A standardised, level-by-level, step-by-step implementation plan based on long-, medium- and short-term needs and emphasising scalability from a holistic perspective will only bear fruit. From experience, I have listed the following points for your reference.

• Top-level design is not an overnight project but a gradual improvement. The top-level design depends partly on experience and imagination but primarily on execution and continuous trial and error, so you should target specific scenarios and keep optimising the top-level design from actual practice rather than carved in stone. It is common in the Internet industry to use applications to incubate platforms and then use the platforms to feed the applications. However, if you focus on building an extensive and comprehensive platform too early, it will often waste resources. That is why many applications don’t realise that the actual requirements deviate significantly from the expectations until they have gone live.

• It is often easy for companies to give too little thought and practice introducing and using external data. Few companies are aware of external data’s role in complementing internal business data when doing top-level design. The purpose of the inventory checking should be to understand the relationship between applications and data, for example, what data is used most frequently within the enterprise, what data is most critical to the development of business, and what data is important but with low coverage. Big Data is not cast in stone, and we need to continually take stock of our resources and update and add to them according to current and future needs.

• In a Big Data scenario, the type and volume of data are continuously increasing. Without a dedicated system to track, monitor, analyse and collate data, it’s difficult to ensure quality, usefulness, and security. At the same time, data security policies are an essential standard for Big Data applications. As the volume and variety of data grow in size, automated monitoring will also become indispensable.

The successful implementation of a data strategy is not only a technological battle but also a test of a company’s management capabilities. Suppose a company believes that the digital economy is a megatrend but does not include data strategy in its management discussions.  Sooner or later, the company will fall into a scarcity of data or lack data management skills. I remember once being asked by an Alibaba board director, “Herbert, what do you think is the value that the Data Department brings to the company?” My response at the time was, “Without a properly functioning Data Department, we won’t see Gross Merchandise Value (GMV) tomorrow!”

With an end-game mindset, enterprise data strategy must be based on business strategy

A data strategy cannot be built on a castle in the air but only based on a business strategy that fits with it. Of course, we still need to have an end-game mindset, which means thinking through and figuring out what data resources and capabilities can match the vision of the business strategy.

The first thing to consider is whether there are enough data resources to support it, and then data governance and the application of machine learning and analytics according to business needs. Similar to developing a business strategy, companies need to constantly review their data strategy and can follow the following four steps:

• Take stock of data resources, governance capabilities, technology and the current state of the organisation.

• Have a deeper understanding of business plans and requirements.

• Select achievable strategic goals to drive the execution of the plan.

• Determine the gap between existing resources and capabilities with the objectives.

From the perspective of data strategy positioning, we should start with examining the current internal status. Clearly understand the existing data resources (or data assets) of the enterprise through data inventory checking, that is, to clarify how much leverage power you have in terms of data and then draw a data map expressing the relationship between supply and demand according to the match between business development direction and data resources. At the same time, it is necessary to fully understand the market, competitors, and industry trends externally and keep pace with the trends in the technology, resources, and data capability (Governance of data, data quality, data safety, data-driven application, etc.).

The four steps mentioned above are directly related to the success or failure of an enterprise’s data strategy.

In my opinion, they are the most critical and easily overlooked. If these four steps are overlooked, companies may end up in 2 bizarre situations.

• Blindly over-expansion;

• Driven by short-term gains, leaving behind the long-term work of digital transformation.

Once you have found the positioning and direction of your business data strategy, it is imperative to find the right route and approach to implement it.

From companies to governments, I’ve seen many digitalisation projects fail prematurely because the wrong entry point was chosen, especially in the initial stages of data strategy implementation, and often months pass before the organisations are still discussing the breakdown of the data strategy objectives and the division of labour between the various management units.

Even if this does not happen, it doesn’t mean things are working well, as people may not be aware of the problems yet. It is only after some running-in and practice that a cross-departmental organisation and effective implementation mechanisms (e.g. synergies, collaboration, incentives, etc.) can be developed to ensure the effectiveness of the investment of resources in the enterprise data strategy and the implementability of the strategic initiatives.

During the operational stage, the idea of setting up a Data Management Committee (DMC) came to the fore to bridge the gaps in coordination between the various parties involved to facilitate better cross-sectoral cooperation and to establish the relevant operational mechanisms.

The DMC was initially set up to address four key issues:

● Establishing data security mechanisms.

● Ensuring data quality.

● Promoting a data-driven culture.

● Recruiting and developing top talent.

Identify Strategic data is a prerequisite for an enterprise data strategy

The starting point of an enterprise data strategy is to clarify objectives and business development needs, then collect data (aggregation), refine the correlation between data and problems (interconnection), and then solve problems (application). If you think linearly, every business problem needs to go through the three stages of “aggregation-interconnection-application” before data value can be effectively unleashed.

It is common for many companies to have different business units collecting data, defining different standards, and building their technology platforms.  If not managed in a unified manner, it can lead to severe data quality and interconnection problems.

The essence of a data strategy is to find a starting point for data governance. So, is it reasonable to standardise all data? Of course not. The purpose is not to “unify the world”, but to pursue a better form of data utilisation!

Once the strategic objectives and the core issues to be faced are clearly defined, it’s time to focus on those data that can generate value, co-build targeted resources, and deploy major resources to solve the issues. This is conducive to ensuring the high-quality implementation of the data strategy.

“From the outside-in” and “from the inside-out”, both perspectives are indispensable

In developing a data strategy, corporates need to set the direction of their business strategy and look at past, present, and future problems in the digitalisation process from both inside-out and outside-in perspectives to develop future-oriented plans and concrete implementation measures.

• Use an outside-in perspective to explore inwards: optimise and align the practical issues of managing and applying data internally, and maximise the output of resources through a more effective and efficient organisation structure and culture.

• Use an inside-out perspective to explore the opportunity for “Co-opetition”: build on existing resource endowments to explore ways to respond to the external environment and gain advantages in the digitalisation process, thereby gaining greater market share, a favourable market position, and more quality data resources through a data-driven approach.