Empowering users to join the data value chain
Data self-service is undoubtedly key to business agility, but the process must start with building a solid data culture in the organisation.
Building a culture of data-driven decision-making puts data at the centre of organisational strategy. When data is used to drive business decision-making, companies proactively make resources available to ensure this important asset is developed and protected.
A well-established data culture ensures users at all levels are empowered to use available data to make informed decisions − from the CEO all the way down to the customer service representative within the call centre or store front.
An organisational data management strategy must be supported by a sound and scalable architecture.
The data architecture must be set up in such a way that data collected from source systems is made available promptly and presented in a format that is easy for business users to consume. Therefore, there is a need for business users to work closely with solution designers and architects during the solution design phase so that their input can be factored in regarding the kind of data they expect out of the systems, how to identify such data and what KPIs are important for their business areas.
For a data culture to thrive in an organisation, data users must trust the data they are working with. The process of extracting data and making it available to the user must be reliable. Data users need to play an active role in the data value chain to ensure the data they need is available in the reporting environment.
Continuous training is necessary to help business users transition from being data content users to data content creators. It is important to have up to date solution documentation, including a searchable user-friendly, enterprise-wide data dictionary to help users work independently with data.
Training should be tailored for specific needs of the target audience. Some business users may only be interested in analysing data using drag-and-drop functionality available in most visualisation tools, while the technically inclined may use advanced techniques to analyse data, which may include coding or scripting.
Organisations often standardise analytics or data visualisation/presentation tools that can be used for data self-service. To derive maximum capability, users need to be trained and be comfortable working with these tools.
Organisational benefits of data self-service
Less reliance on IT
The role of IT is rapidly evolving due to digitisation and cloud adoption, and IT departments are getting smaller. This has made a case for data self-service even stronger, as there is now this small number of IT personnel that is required to service the reporting requirements of the entire organisation. This leads to longer waiting periods as requests are queued and prioritised.
Data self-service reduces reliance on IT and places the power in the hands of the user. This allows business to continue without unnecessary delays. Empowering business users with self-service frees IT to focus on data engineering and maintaining the infrastructure that supports data self-service.
Users work with data they are familiar with
Business users are often familiar with the business process that generates the data, and they are likely to have been involved or had an input in the design of the upstream source systems. They can interpret and understand departmental or industry acronyms and jargons that may not make sense or be obvious to the data engineers. Users are also able to provide useful input to the data engineers to enrich the data to serve their needs better.
Data self-service ensures business can access data quickly without delay, ensuring important decisions can be made quickly, which could give business a competitive-edge. In a traditional environment, requests are logged with the IT department, which then assigns priority according to set criteria and this determines the turnaround time. Delays are often inevitable as IT works through a long list of requests according to their priority. With self-service, business users can quickly extract their own data, analyse it, and make decisions.
Overcoming challenges of moving towards data self-service
A social media model can be adopted for data self-service within organisations where business users create their own content using available data and be able to share it with other privileged users. Content re-use is important to reduce rework, which increases organisational efficiencies.
For a data culture to thrive in an organisation, data users must trust the data they are working with.
One of the ways to increase user adoption is creating self-service templates. For example, a finance department can have a set of templates to report on revenue. Templates will ensure a standardised way of reporting where consistent business rules are applied by all users.
To ensure a wider adoption across the organisation, data self-service tools need to be intuitive. Most business users are familiar with Excel or web-based interfaces; it makes adoption easier if the self-service tool uses a familiar interface. For example, a web-based interface of PowerBi and Excel interface of Microsoft SSAS Multidimensional cubes allows data to be presented using an interface that users already know.
Most technology-related problems can be solved by migrating most or all of an organisation’s infrastructure and services to the cloud. Cloud service providers usually offer built-in tools, which can help eliminate blockers and facilitate implementation of data self-service.
Factors to consider before implementing data self-service
Current state of the organisation
What is the current technological landscape of the organisation and the maturity level of the current data management processes – versus what is desired? What are the current challenges the organisation is facing? Looking at various departments within the organisation, what are the important business questions that each department would like to answer?
Defining the data self-service audience
Self-service may mean different things to different users. For example, a finance department superuser may prefer to use an Integrated Development Environment tool to query the database directly using SQL, while a marketing department user may be more comfortable accessing data using the Excel interface of an MS SQL Server Analysis Service multi-dimensional cube to access data.
Data scientists, on the other hand, may want to access and analyse data by employing advanced techniques using statistical packages like R and Python to deliver actionable insights. A data self-service strategy must seek to address the need for all types of users.
Security and data governance
There are security concerns around data access – how to ensure each user has access to the right data. If the self-service strategy involves moving on-premises data to the cloud – how does this affect existing data management and data governance strategies? How does it affect existing statutory and regulatory requirements?
It’s important to establish a role or function that focuses on data governance. The role should be responsible for defining and implementing necessary controls and auditing measures to ensure the integrity of the reporting environment.
To successfully implement a data self-service strategy requires an organisational cultural shift. It is important to ask the right questions to determine the state of the enterprise data landscape within the organisation. This will help formulate a plan to get to the desired state to fully support data self-service.
Both IT and business users need to work together to create a solution that works for the organisation.
Data engineer, PBT Group.
Nathi Dube is data engineer at PBT Group. He is a data engineering consultant with local and international experience spanning telco and broadcast media industries, and large-scale greenfield data warehouse projects.
Nathi Dube is data engineer at PBT Group. He is a data engineering consultant with local and international experience spanning telco and broadcast media industries, and large-scale greenfield data warehouse projects.Dube holds a BSc degree with computer science and mathematics majors from the University of the Witwatersrand. He also has scientific computing, Python, machine learning and statistical analysis certificates from WorldQuant University and is an AWS certified cloud practitioner.