The success or failure of a data project relies on the organisation's ability to move from raw data to an analytics-ready state.
By Kevin van der Merwe, Sales Director at iOCO Qlik.
In this second article on data literacy and its importance, I will look at the data lifecycle and then how to integrate the right levels of data literacy into your organisation.
Previously, I discussed why data is so important in business and thus why it is so imperative that every employee can 'speak data'. Before we get to the question of how to make the organisation data literate, we need firstly to recognise that data itself has a distinct life cycle, which needs to be carefully managed.
The value of data is only unlocked once an improved decision is made due to the use of the data and that decision brings an improved action. But before we can get to insight, we need analytics, and before we can get to analytics, we need analytics-ready data, and before we can get there, we need raw data.
The process to go from raw data to insight and action is called the data value chain and there are different people involved in that process.
Data comes from a variety of sources, among them internal systems (such as the enterprise resource planning, customer relationship management and HR applications) as well as external sources. The latter could include the internet of things, social media and so on.
All of this data will come in a multiplicity of forms, structured and unstructured. Raw data needs to be managed and integrated to make it available for analytics.
The value of data is only unlocked once an improved decision is made due to the use of the data and that decision brings an improved action.
Raw data is consolidated into data lakes where it can be categorised; from there, it can be transferred into the data warehouse and marts. The data then must be integrated and then analysed.
The final step is to provide users with various tools to help them discover what data is present, and then request reports or create dashboards relating to data that will be useful to them.
Each one of these steps must be undertaken by people with the requisite skills. The kinds of job titles we are looking at include data scientist, ETL (extract, transform, load) specialists, data architects, database administrators, data engineers, business intelligence professionals and data scientists.
However, please do not lose sight of the important point made in the first article: ultimately, the end-users of the data insights generated will be the business decision-makers and operational staff who would typically be the least data literate.
In other words, the success or failure of the whole data project relies on the organisation's ability to move from raw data to an analytics-ready state and for the business staff to translate that analytics data into action-ready insight.
Obstacles to creating a data-literate organisation
Before considering how to improve the organisation's data literacy, one needs to be aware of common barriers. These are:
Resistance from the workforce, including the C-suite. Change is always hard, and most employees are used to relying on some element of gut feel to make decisions. The case for basing decisions on evidence needs to be carefully made. This effort must span the whole organisation − Qlik research showed that only 32% of the C-suite is data literate. Data champions need to be identified across the organisation, and the chief data officer or chief analytics officer needs to play the champion's role at the top.
The need for governance. As more and more data becomes available, employees are starting to use a growing number of datasets to help them make better decisions. This democratisation of data has the effect of driving decision-making further down the organisational hierarchy but, like the similar problem of shadow IT, it can end up creating a chaotic environment. The organisation's leadership must provide governance along with self-service capability for business users. At a practical level this could be achieved through using a uniform platform on which to deploy new datasets and applications.
Overcome employee insecurity. Increasing use of data can be seen as a threat to job security, in the same way that automation and robotics are. Most employees, even young ones, are also not certain they are data-proficient. Company training in this important area is crucial.
Break down organisational silos. Organisations must take care to ensure the data-literate or specialist employees they employ are not isolated in IT or business intelligence teams. One approach is to establish forums in which data leaders or champions can share knowledge and answer questions from the entire employee body.
How to create a data-literate organisation
The process for changing the organisational culture to one in which data plays a central role involves four basic steps:
Communicate the power of data. This process needs to take place across the whole organisation and should be consistent. Providing practical examples is recommended. For example, showcase how a customer insight led to a new business opportunity, or how an individual was able to gain approval of a new idea by backing it up with facts and figures. It would also help if leadership reported regularly on how the use of data improved important metrics.
Assess progress. Organisations need to begin by understanding the current status of their workforce's data literacy, and then track progress. A granular understanding of how each part of the organisation is progressing on the journey towards data literacy will also help in customising training programmes.
Establish training programmes for each type of data user. As noted in the first article, there are four main kinds of data user. Training needs to be appropriate to how an individual will use data. Qlik research showed 66% of respondents believe they have received adequate data training − one-third of employees need to be upskilled.
Repeat. As already noted, the volume of data currently being generated is unprecedented. The skills to deal with this volume of material, as well as the emergence of new types of data, need to be refreshed constantly − data literacy is truly an ongoing journey. Here's a top tip: assess new joiners; if skills are lacking − get them onto a foundation programme immediately as part of the onboarding process.
Creating a data-literate workforce can seem like a mammoth task but, as the old adage goes, “the only way to eat an elephant is one bite at a time”. As I have argued, data literacy is essential to the sustainability of all organisations, so it is vital this process begins at once.