Data strategy begins with an understanding of your business goals. What capabilities do you need to develop to realise your strategic objectives? In this blog I continue to build on the data analytics concepts to outline how to improve the analytics capability in your organisation.
One of the best ways to assess analytics maturity is the framework developed by Davenport and Harris. It covers Data, Leadership, People (Analysts), Enterprise and Targets across five stages of maturity.
| Data | Enterprise | Leadership | Targets | Analysts | |
| Stage 5 Analytical Competitors | Relentless search for new data and metrics | All key analytical resources centrally managed | Strong leadership passion for analytical competition | Analytics support the firm’s distinctive capability and strategy | World‐class professional analysts and attention to analytical amateurs |
| Stage 4 Analytical Companies | Integrated, accurate, common data in central warehouse | Key data, technology and analysts are centralized or networked | Leadership support for analytical competence | Analytical activity centered on a few key domains | Highly capable analysts in central or networked organization |
| Stage 3 Analytical Aspirations | Organization beginning to create centralized data repository | Early stages of an enterprise‐wide approach | Leaders beginning to recognize importance of analytics | Analytical efforts coalescing behind a small set of targets | Influx of analysts in key target areas |
| Stage 2 Localized Analytics | Data useable, but in functional or process silos | Islands of data, technology, and expertise | Only at the function or process level | Multiple disconnected targets that may not be strategically important | Isolated pockets of analysts with no communication |
| Stage 1 Analytically Impaired | Inconsistent, poor quality, poorly organized | None | No awareness or interest | None | Few skills, and these attached to specific functions |
You can use this framework to determine the current level of data analytics maturity of your organisation. Let’s go through a case study to demonstrate how to apply this framework in practice.
Consider you’ve been put in charge of data strategy in your company. Let’s imagine your analysis revealed that’s it’s at Stage 2 – Localised Analytics. For example, only the Product team uses data analytics consistently for specific use cases, mainly to track customer metrics (e.g. onboarding and retention). The Finance department also uses some data analytics to track budget and other financial metrics. In both cases, these are limited to descriptive analytics, focusing on point-in-time reports and dashboards with no forecasting capability. Both teams use different tools, there is no centralised storage and analytics capability, processes are ad-hoc, inconsistent and not well documented. There is high dependency on the individual knowledge of a few key employees.
How do you go about lifting the analytics maturity to the next level?
To get to Stage 3 – Analytical aspirations, you need perform a gap assessment for every business function to define integrated data analytics capabilities required to support the business objectives.
You can then develop specific plans to close capability gaps, outlining initiatives required to drive business decisions. These include the type of data needed and identification of data sources.
This strategy should also include data governance requirements and cover data quality, compliance, security and privacy to ensure lawfulness, fairness, transparency and accountability when processing sensitive (including personal) data.
Next, you can develop an enterprise architecture to support the data strategy. For example, within cyber security we can use NISTIR 7756 as described in my previous blog:
Following it bottom-up, we can trace data collection from various sources, including people processes and technology, through to the storage and presentation layers which in turn help drive timely and informed decision-making. A similar technology-agnostic architecture can be used for the organisation as a whole.
As outlined by the high-level architecture above, the organisation will need to invest in infrastructure and tools to support capabilities, such as data collection and storage (data warehouse), analysis and visualisation. Such tooling uplift needs to be supported by the corresponding investment in talent. This can range from hiring new staff, upskilling existing employees and scaling through partners (e.g. external consultants).
Multiple skills need to be covered (e.g. data wrangling, data analysis, design, storytelling, project management) by multiple people. It’s not about hiring a single data scientist and expecting them to solve all the problems.
Engaging business stakeholders is key for this to succeed. Throughout this process you’ll need to get buy-in from top management on the data strategy and the roadmap to get to the next stage of analytics maturity.
The tone from the top is important in building the culture of data-driven decision making. Processes, tools and behaviours should be embedded across the organisation with the right incentive structure and development opportunities (e.g. training courses, lunch and learn workshops, knowledge sharing sessions). Leading by example, showcasing and celebrating wins, however small, is key to maturing the data analytics capability in the company. You should also anticipate some challenges and prepare to manage them appropriately.
Culture change is difficult and the amount of change required can be significant. Some people are used to making decisions based on their ‘gut feel’ and not everyone will welcome the new way of doing things. Staff will need to adopt new ways of thinking, working (including potentially learning to use new tools) and making decisions. To overcome this, involve key stakeholders (heads of departments and top management) in co-developing the strategy and highlight the benefits of developing the data analytics capabilities.
Integration and data quality is another concern. Availability and quality of data will vary across the company. Many organisations evolved through a series of mergers and acquisitions, so breaking the silos and harmonising tooling and ways of working can be challenging. There might be multiple legacy systems in use and some processes may still be manual. It will be challenging to automate some of the immature processes and improve communication across departments. The ‘garbage in, garbage out’ concept is relevant here too, as there might not be a single agreed data format or quality standard.
This initiative can be quite expensive for two reasons:
- Data analytics talent is hard to recruit, develop and retain in the current market.
- Tooling, such as data warehouses, can be expensive to implement and maintain.
The top management will need to be presented with a clear business case with measurable return on investment if such an initiative is to be approved. The level of investment in people, process and technology required can be significant. Same goes for the level of commitment to changing the organisational culture, particularly if the gains will not be quick and potentially not immediately obvious.
With this in mind, it’s important to identify quick wins and make incremental improvements. For example, can you extend the existing Product team’s data analytics capability to other teams, starting with Engineering? Keeping your data strategy agile, with a balanced focus on long-term outcomes all the while delivering value quickly is key to success.
Cyber Security Leadership 