Hartnell Ndungi, Chief Data Officer at Absa Group, talks about how data leaders play a critical role in measuring the success of digital transformation efforts across the enterprise.
Data is at the core of every digital transformation strategy, so it’s no surprise that data leaders are increasingly being called upon to support digital programmes and measure transformation effectiveness. Assessing your transformation journey requires a digital maturity model, of which consultants around the world offer a dozen. “To effectively measure the success of digital initiatives, it’s important to establish measurable KPIs, data-driven metrics, and relevant dashboards to showcase the business’s stepwise growth from digital use cases and performance against targets,” says Hartnell Ndungi.
Ahead of his session at Velocity – Data and Analytics Summit – taking place in South Africa on March 07-08, 2023, Ndungi spoke about how the industry is steadily shifting toward data partnerships and reciprocity as organisations become more aware of the use cases that can be delivered from internal data
Excerpts from the interview;
What are the key metrics you prioritise to measure data quality?
The key metrics to monitor enterprise data quality are accuracy, completeness, consistency, timeliness, and validity. Automated quality reports should be deployed to ensure timely action on any exceptions. Data stewards and owners should be appointed in each enterprise department to guarantee the highest quality of data. Bi-weekly stewardship meetings are essential for discussing quality findings and exceptions, and data stewards should receive support from data owners to resolve data issues in accordance with the data quality metrics.
Tell us about your data vision for 2023, and how you see the industry changing.
The industry is steadily shifting towards data partnerships and reciprocity as organisations become more aware of the use cases that can be delivered from internal data. Data partnership agreements are rising, with banks partnering with FinTechs to explore shared value in unbanked ecosystems. This year, data literacy remains an area of focus as organisations strive to ensure that their business colleagues have the skills to run self-service solutions and participate in data projects. Artificial intelligence (AI) has become the engine behind value engineering as well as data commercialisation, making it essential for organisations to have the right mix of data scientists and AI engineers to maximise their use cases. Senior executives are increasingly focused on achieving a return on investment from data projects, necessitating the inclusion of a data monetisation excerpt with value drivers in their data strategies.
How can data leaders help measure the success of enterprise-wide digital transformation?
Data is at the core of every digital transformation strategy, so it’s no surprise that data leaders are increasingly being called upon to support digital programs and measure transformation effectiveness. Assessing your transformation journey requires a digital maturity model, of which consultants around the world offer a dozen. Most models are organised into five stages, from the least mature to a stable digital maturity. To effectively measure the success of digital initiatives, it’s important to establish measurable KPIs, data-driven metrics, and relevant dashboards to showcase the business’s stepwise growth from digital use cases and performance against targets. The most important digital metrics to track include return on digital investments, employee productivity, adoption, digital reliability, cost-benefit ratio, active users, and revenue from digital.
What are the key considerations when investing/ building solutions for your technology stack?
As cloud computing continues gaining popularity, organisations must consider several key decisions before starting. The first is whether to utilise cloud or on-premises solutions. All data analytics and data science tools are now available on the cloud, making this an increasingly popular option.
The second decision is to purchase an existing solution or build one from scratch. With the right data skills, enterprises can build a comprehensive data and analytics platform with descriptive and prescriptive capabilities. However, working with consultants or outsourcing may be the best approach for more complex solutions and platforms if the organisation is at a low data and digital maturity level. Furthermore, organisations should consider their business strategy, existing infrastructure, existing skills, use cases, data privacy guidelines, regulation, and leadership when making their tech deployment decisions.
What advice would you give enterprises beginning their AI journey? What should data leaders know as they start and when they scale?
For any effective AI strategy, it is essential to embed data management principles properly. Poor data quality is one of the major reasons why AI models fail, making metadata management and data lineage critical considerations for providing well-tagged data for modelling. It is also important to have the right skills to ensure the successful deployment of AI solutions to production. When using AIOps approaches to model building, working with engineers is a must. To avoid falling into the AI hype, each AI use case should be backed by a business need, and the right hypotheses should be set for data studies to simplify proof points and model validation. When scaling AI projects, the focus should be on delivering value. As businesses increasingly recognise the value of AI and data science, more use cases are being requested. Value can be measured across four dimensions: cost reduction, revenue generation, asset growth, and operational efficiency.
How do you approach search and navigation within organisations to make data more accessible for non-technical teams and users?
Knowledge bases are an excellent way to disseminate technical knowledge to business colleagues. Structuring data into categories and specialities makes accessing information simpler. However, data literacy is a more effective way to share data knowledge throughout the organisation. With tiered training programs tailored to different types of audiences, data skills can be imparted to every non-technical user. Self-service analytics is also an excellent way to decentralise insights to business users. Users can create ad hoc reports and dashboards for quick insights using simple drag-and-drop tools.
If you liked reading this, you might like our other stories