Digital transformation is a process of learning from mistakes
In the course of consulting with a number of company executives, I have found that they all agree that companies must speed up their digital transformation, and there is no shortage of those who aspire to shape their businesses into data-driven enterprises. Readers who have gone through this process should understand that digital transformation is difficult, especially for some companies in traditional industries where rapid change in a short period is impossible.
Over the years, I have discussed the topic of digital transformation with many executives in my consulting work. Below, I have summarised some of the issues that resonated with me and hope readers can draw some ideas from these conversations for digital transformation in their businesses.
CEO Q1 : After much thought, I’m perplexed whether it’ll be too risky for the traditional industries to undertake digital transformation. Traditional and internet companies are fundamentally different, our business is mainly offline, and sometimes it is difficult even to collect data, not to mention data-driven.
Herbert: I think many companies tend to make the mistake of expecting digital transformation to be “big and comprehensive”, but this is what we need to avoid. Because transformation is a process of learning from mistakes, the key to transformation is not in the technology but in addressing the team’s resistance due to the lack of adaptation. Especially in the case of a cold start, it is better to prioritise a clearly defined business problem and then use the problem as a guide to finding data or select parts of the traditional business for a full digitalisation. The former is more conservative but less risky; the latter is riskier but can have a significant impact if successful.
CEO Q2: I understand that the person in charge of a digital transformation project plays a crucial role and a critical factor in determining the success or failure of the transformation. So who should be the head of this project, a business leader or a technical leader? If the business leader takes the lead, I am concerned that the digital transformation project will turn into a tool for existing business development, unable to think outside the current business logic. However, the change we need is to think outside the box.
Herbert: I think digital transformation isn’t a project but a change that ripples through the resources of the whole company. So the person you choose must report directly to the CEO and must have interdepartmental coordination skills, be technically savvy and familiar with the business. This all-round talent is best sourced internally and backed by executives and consultants with practical transformation experience.
In addition to selecting the person in charge, I believe it’s also important for top management to set data strategies, manage targets and review them regularly. Remember, digital transformation is a company-wide change and must be a team effort.
CEO Q3: Our Group has a wide range of businesses, and the data of each subsidiary is scattered; even within the subsidiaries, the data is distributed across dozens of different systems, so we often hear various complaints in using the data. Centralised management is challenging, but it must be done. How should the organisation be structured so that data within and across subsidiaries is managed in a unified manner? Should it be managed separately by each subsidiary, or should the Group as a whole manage it?
Herbert: Data governance must start with an overall plan. Even if it begins with a particular department, there should be an overall top-level design as a reference. It’s essential to set up a horizontal team for coordination. However, it is crucial not to see the coordination work as centralised management.
For instance, the scattered data is justified because subsidiaries naturally structure their data to match their business development to speed up their growth before considering the synergies if standardising or sharing data can lead to more opportunities for growth.
Therefore, integrating or segregating data resources depends firstly on business needs and then on strategic deployment. Attention should be paid to the proportion of strategic components, especially the integration of common data should not take up too large a proportion. This balance depends on the Group’s data maturity and the use with exchanged data. The prerequisite for all this is to increase the transparency of information about data resources among subsidiaries. If there’s no such trust, how can we talk about connecting the data!
CEO Q4: If a centralised DFP has already been established at the Group level, do subsidiaries no longer have to build their data platforms? What kind of relationship should the Group’s DFP have with the data platforms of the subsidiaries?
Herbert: The establishment of a DFP for corporations is not simply to centralise all data but to connect the channels for data sharing driven by business objectives and needs. Connecting data doesn’t mean that everyone can use it casually; data security, compliance, and governance are all encapsulated in this.
After confirming data interconnection as a strategic goal and the core issues to be faced, it is possible to define the scope of the data, focus on superior resources, and build the targeted data resources together.
Having understood this critical point, you should know that the Group’s DFP is to solve the problem of sharing data across subsidiaries, thus maximising the value of the data and using it to drive each subsidiary’s business development. At the same time, there’s no contradiction in that each subsidiary has its own data needs to deal with internally.
If a subsidiary feels that they need to set up their data platform, the way to look at it is to see whether it adds value to the Group’s DFP or is just creating another “data silo”.
CEO Q5: I’ve heard lately that data asset is an essential part of a company’s future assets. I agree with that, but only to a limited extent because data can also be a liability. I’d like to hear your thoughts on data assets, particularly how data can be managed to become a useful asset?
Herbert: You’re right! From the perspective of data strategy, the first step is to assess the current state of data within the company, conduct a data inventory checking to clearly understand the relationship between the existing data resources and the business and analyse the availability, usage, and sharing of data.
Suppose a corporation lacks an understanding of its data asset and lacks monitoring of the data lifecycle process. In that case, planning and implementation will eventually go in the opposite direction.
The situation is somewhat more complex when assessing data assets from a strategic direction. Due to the external nature of Big Data, companies also need to be aware of the external perspective of their data resources. Understanding the strategic complement of third-party external data to internal data could be the key to improving the competitiveness of your business.
As you say, not all data can be a useful asset in an organisation; only data that meets the requirements for data usability and that generates value can be called data asset. External data may have to be collected more carefully because of its cost. Otherwise, it may not be able to make ends meet, and don’t forget that data governance is a process that requires a lot of investment.
Corporates need to develop a data governance policy based on the strategic scope of the data (reasonable assumptions). But driving digital transformation across the corporation is challenging and requires planning and governance on every aspect of the data lifecycle so that data becomes an asset, not a liability, and turns into a core component of your competitiveness.
This kind of data project can take several years and requires a data governance implementation plan based on the preliminary data asset inventory, the actual needs of business development, clear identified objectives and the clear work plans to be achieved in the short, medium and long term.
Finally, there should be a quantitative metric to measure the data asset, one of the best aiming devices during digital transformation.
CEO Q6: As I mentioned in the first question, our business is mainly conducted offline by agents (intermediaries), and there are natural barriers to data collection. To improve the service standard of the agents, we’ve been working on using Big Data to empower them, but the results are not satisfactory. Reason being:
1. data is scattered across different systems in subsidiaries and has not been connected for sharing.
2. many data is not collected or cannot be collected and often requires manual analysis and processing.
I would like to hear your view on this.
Herbert: Taking this question as an example, we can look at it from 2 perspectives:
● From the outside in – based on business needs and objectives, can you identify the data that needs to be collected from the agents and used to solve what type of business problems and what benefits does this bring to the agents?
● From the inside out – correlate the data collected from customers to the company’s internal data to increase the value of the combined analysis of the data.
It is important to note that this ability to correlate needs to be considered from a data governance perspective. Managers need to be clear about the purpose and value of acquiring customer data and the quality control strategy when aggregating data both internal and external, including the source of the system, update frequency and data change checking mechanisms, etc.
I want to make the final point that the clearer the objective, the more targeted the data scope and the better the focus on collecting the required data. Sources of data can be: existing, integrated and formed, and identifying alternatives.
The alternative of data is a bit abstract; for example, if a competitor’s shop sales are hard to come by but can be predicted based on footfall outside the shop (e.g WIFI, infra-red sensor), I like to call this proxy data.
CEO Q7: Wouldn’t it be paradoxical if all subsidiaries want access to other subsidiaries’ data to develop new business, but none wants to share their data with others?
For example, the life insurance division has suggested that it would like the Group to co-ordinate with other subsidiaries to open up the sharing of data (e.g customer CRM) so that insurance agents can have a clearer picture of their customers’ needs for business development and to increase product sales, but when our subsidiary that provides Lending Universal makes a similar request, our life insurance division refuses for data security and regulatory reasons.
Lending Universal is the combined name for credit services and inclusive financial services. Inclusive financial services refer to providing appropriate and effective financial services at an affordable cost to all segments and groups of society in need of financial services, based on equality of opportunity and commercial sustainability. –Editor’s Note
Herbert: Going back to your previous question about the difficulty of digital transformation for traditional companies, I‘d like to use Amazon and Alibaba as examples to illustrate.
Driven by the customer first principle, these internet-based companies rely heavily on data to help acquire target customers, engage them and improve their experience. Under this premise, different subsidiaries are willing to share the data they collected to enhance their knowledge of customers. During the process of sharing, they rarely look at the overall interest of the Group, and they believe that the data is owned by their subsidiary.
The first step to changing this mindset is to start strategically and make all employees understand that all subsidiaries are interdependent with a common destiny. Naturally, subsidiaries have their little agendas; this won’t harm as long as they don’t affect the overall performance of the Group.
You can think about the following questions: What data does life insurance have? What data do you want? What data is already being used? What data do you think will be useful in the future? What data can be shared with other subsidiaries?
If the subsidiaries can answer these questions, I believe a solution to facilitate data flow among the subsidiaries will emerge. Because data sharing is essentially a matter of benefit distribution, the Group can assign commonly interested data under the Group’s management and then determine a benefit distribution mechanism among the subsidiaries.
CEO Q8: In recent years, we’ve been doing a lot of data-related projects one after another, and we have set up a special data management team in the life insurance division. Every year the cost of data collection, processing, development and storage is upsurging. How should we refine our approach to control costs and allocate and use resources more rationally?
Herbert: No company can endlessly expand the scope of data collection and increase the volume of data. However, many companies have an obsession: what if some data is useful? So they will back up or archive as much data as possible.
In addition, inappropriate cost allocation causes resources wastage. The most apparent behaviour is duplicating the same data stored by different teams, and the misuse of computing resources is rife. Not only does it raise the total cost of data storage, but it also increases the cost of maintaining this data.
The development of a data archiving strategy should not be overlooked as a breakthrough point for refined data management and cost savings.
The goal of data archiving is to move inactive data out of the data lifecycle and focus on optimising the efficiency of active data to save costs.
Organisations can consider storing static data in a more cost-effective hierarchical archive management system, leaving it in a retrievable backup state before physical deletion. Data archiving is not a new issue in data, but as the cost of storing data continues to decrease, the strategy for archiving data should remain flexible.
CEO Q9: The last question, and one that has been a tremendous headache to me. I am willing to invest resources in digital transformation, but how should I measure the return on investment? In particular, should I use company performance indicators as targets for the success of the digital transformation? I would like to see the changes that digital transformation brings to the business within a year. What is your view?
Herbert: Your question is also one that many CEOs are concerned about, and it’s impossible to answer this question simply because every company has a different focus at different stages of digital transformation. In principle, a data-driven approach starts from an analytics report to multi-dimensional analytics software, from man-made decisions to automated algorithms and automated operations. The clearer the closed-loop data collection (AKA data feedback loop), the easier it is to measure the success of the digital transformation. The key is for companies to get into the habit of “setting up success measuring criteria on quantifying projects before starting them”.
The most ludicrous situation I’ve ever seen is when an organisation has many data transformation projects but has no decent quantitative metrics.
Successful digital transformation requires employees to change their traditional behaviours and disrupt their old ways of thinking. The only way to fundamentally evolve a business is to change how the traditional business processes operate in an organisation and equip everyone with a data mindset.
Just as the information technology revolution in the 1980s and 1990s introduced business from paper to internet systems, digital transformation now requires a radical change in company culture and employees’ thoughts. Only then can the value and role of data asset be brought into play; otherwise, it remains as a local tool at scattered points
At the same time, it is essential not to see digital transformation as a campaign project that will achieve performance targets in the short term. From my observation, most companies pursue short-term benefits in digital transformation and overlook the traditional vertical division of labour and information management characteristics. Thus, they fail to reuse, co-create, and accumulate knowledge properly and are prone to lose innovation.
Suppose companies want to see their business grow in the short term because of data transformation. In that case, they must first look at the value points of their customers and examine from the outside to see if there are opportunity gaps that can be addressed using a data-driven approach. For example, when a new customer visits to gift shop, are there analytics tools to help understand the customer’s preferences, What they have seen recently, or if that customer buying for herself?
Finally, I would like to highlight the fact that over-eagerness and over-ambition in the data transformation process of a business are both equally fatal and result in failure.