Introducing a 3 posts series on make-or-buy decisions, let’s start with the key capabilities to master along the data value chain to secure performance improvement.
Why make-or-buy matters now
There are two main arguments for explaining why make-or-buy decisions regarding the data value chain are important:
- Data assets and capabilities redefine market boundaries and competition dynamics. Technology companies are challenging non-digital companies with a value proposition based on analytics (IBM is a good example of that). Platforms challenge incumbents with usage-based value proposition (Uber and Airbnb are good examples).
- Data and analytics are a significant value creation driver. Costs can be reduced, revenues increased, as margins. We described this in previous posts.
So the question is for all non-data natives companies to decide on their positioning in the data value chain: which activities are key, which should be mastered internally, what to outsource? which partnerships to forge?
Depending on how they answer these questions companies have a higher or lower ability to leverage data assets and capabilities for value creation.
OK, positioning is important, but why is it particularly NOW? I think we are in a time when several parts of the puzzle are assembling:
- The incentive to leverage data and analytics assets is higher: incumbents are being aggressively challenged by digital players and one key resource in this new battle is data.
- The time of experiments and proof of concepts ended: in many industries, incumbents are redefining their physical-digital balance and the experience of the last 5-10 years helped them to define a position and a strategy regarding their digital capabilities. COVID 19 has accelerated this shift for many companies. See for example this recent McKinsey study which reports that the adoption of digital services has increased as much in 6 months that it would have been in 7 years.
- The market for solutions is now well structured: thanks to progress in research and teaching, investors poured a significant amount of money in AI startups in the last 5 years. Now the market for AI solutions is flourishing. Similarly, the Tech giants have scaled up their capabilities.
Value chain 101
Value chain is an essential concept in strategy, described by Porter at the beginning of the 1980’s. The value chain describes the categories of activities within an organisation to create a product or a service. It is used to model the value generation of an organisation and its competitive advantage.
« Strategy is the creation of a unique and valuable position, involving a different set of activities. The essence of strategy is to choose activities that are different from rivals » (Porter, 1996)
The value chain of a company is its recipe for value creation:
- what are the key activities to realise to perform in one market
- which activities are done internally and which are outsourced
- in which activities the company performs better than its competitors
Data value chain
The connection between data assets and capabilities and the value chain is twofold:
- directly as a support activity (“Technology”)
- indirectly as they have an influence on how the primary activities are realised and their performance
If we want now to be more specific about the activities in the data value chain, we can split the data value chain into 4 components:
- Generation: activities and assets necessary to capture and record data (structured, semi-structured and unstructured). E.g.: Web applications, ERP, IoT and connected devices, Social media, …
- Collection: activities and assets to collect, validate and store data. E.g.: stream, cleansing, reduction, integration, storage infrastructure and models, security, …
- Analysis: activities and assets to analyse and generate insights. E.g.:Semantic analysis, models (Predictive, Descriptive, Prescriptive), visualisation (graph, maps, 3D, …), …
- Exchange: activities to expose outputs internally and externally. E.g.: decision making, trading
Key resources and capabilities for value creation
So now that we know why it’s important and what is it, let’s concentrate on how it works.
1) It’s a matter of data assets and capabilities
Quite evidently if a company wants to use data as leverage for value creation, it needs to master data assets and capabilities:
- tangible assets: data, technology, infrastructure
- human assets: technical skills, managerial skills, domain expertise, relational knowledge
- intangible assets: data-driven culture, organizational learning
- data collection capabilities: text mining, web mining, social networks analysis
- analytical capabilities: statistics, optimisation, modelling, machine learning
- interpretation capabilities: data visualisation
- predictive analytics capabilities: forecasting, simulation
These are the necessary assets and capabilities. You cannot mass-customize your marketing messages unless you have a dataset on the past behaviour of your customer and prospects + an infrastructure to analyse it in near-real-time + talents to fine-tune recommendation algorithms.
2) It’s a matter of IT and design capabilities
Data assets and capabilities are necessary but not sufficient, other IT and design capabilities are necessary.
Recent research (MULLER, FAY, Vom BROCKE (2018), The Effect of Big Data and Analytics on Firm Performance: An Econometric Analysis Considering Industry Characteristics) showed that data assets are associated with an average of 3%–7% improvement in firm productivity. Yet differences in returns are substantial when the industry is considered. In IT-intensive industries, the average productivity gain is 6.7% productivity gains (vs 0 for other firms). Which means that data solutions require complementary IT assets and capabilities (transactional enterprise systems, data scientists, …) which can provide the necessary data and skills to extract knowledge out of this data.
On a different note, a lot of the value creation associated with data assets is related to web services and applications for customers: customized content, online services, usage-based service, … The value is captured only if the customers massively adopt these services. In that matter, design capabilities (UX and UI) are key to trigger massive adoption. If a company has the perfect dataset, the good infrastructure and talented analysts but fails at designing an attractive application for its customers, the value is not captured. This is partly why it has been so difficult for manufacturing companies to thrive in the Internet of Things environment, they lacked the design capabilities to produce an attractive solution to their customers.
3) It’s a matter of ability to respond to the signal
To illustrate this in a previous edition, I used the example of a bank. They had a problem with churn, so they implemented a data-driven approach to predict which of their current clients would churn in the short term. The model worked pretty well, but the problem was not to know they will churn, the problem is to have the good services to keep them.
Data-driven insight is only a component of a firm’s ability to sense, seize and reconfigure, the organization must be designed to be able to respond to changes that insight indicate. Big data generated insight is only one component of gaining value from big data investments, the other is responsiveness.
Conclusions on data value chain
- Big data resources and capabilities are a necessary component to master but are not sufficient for capturing the value (other IT components, design, community and ability to respond)
- According to which service is offered, various stacks may be required (threshold capabilities)
- There is little use building a robust technical stack without building a strong operational capability