Data is powerful. Businesses at the top of their game understand this, and that the key to success lies in harnessing the enormous value that their data yields to drive optimal decision-making. In this regard, advanced analytics approaches incorporating data science, machine learning (ML) and artificial intelligence (AI), are able to transform data into vital operational acumen that empowers decision makers to improve their business.

However, becoming data-driven requires more than buying a technical solution, or hiring a data scientist.

When data can go ‘wrong’

Approximately 90% of the world’s data has been created in the last two years, equating to 2.5 quintillion bytes of data each day. Businesses are keen to apply the latest analytics methods to generate wisdom and value, but many of them make the mistake of trying to glean anecdotal correlations from big data, with limited to no success in translating these to the day to day operations of their business.

As an example, one multinational retailer invested heavily in data science, yet the projects struggled to make an impact and add measurable value because there wasn’t buy-in from the wider business. Another large financial institution had really high performing data scientists using bleeding edge technology, but it was not well connected to the rest of the business and suffered the same fate. In both cases, the anticipated outcomes of the data science initiatives fell woefully short of the expectations to deliver commercial value. And, very often, because the data science methods and commercial execution are not aligned, the blame can be shifted onto the data for not delivering the perceived return on investment.

Experian study: why organisations think they have bad data

How bad is bad data? Well, considering, in the past, it has caused major companies to go bankrupt and started wars; not to mention, according to a recent study from Experian, it’s ruining customer experience, we’d say ‘pretty damn awful’

This is of course not entirely fair. In fact, there is rarely such a thing as ‘bad data’, but rather, a bad understanding of its intended purpose and the underlying processes used to collect it. What’s a lot more common in practice is a failure to connect data with its intended purpose or a failure to apply that data for secondary purposes to deliver value. Data can be ‘bad’ when there is no transparency around how or why it’s being collected. What’s actually lacking is good data governance, which is fundamental for setting the parameters for data management and usage. If no one in the organisation is taking responsibility for its data, inevitably standards will fall and data will not be collected, analysed or used in the way it was intended.

It is understandable therefore that it’s the data that comes across badly when data science and analytics projects fail. But how can it be changed or corrected?

Don’t just change a process – create a culture

We can only make an impact when we ensure that we are using the data to answer the right questions, i.e. those relating to the decisions that drive the business value; and then answer them through developing the appropriate data science solution.

A data-driven culture and governance are rooted in people, process and technology. So often organisations make the mistake of starting with the technology with the best intention to use data to change a business function for better. But without the people or the process to do that, projects inevitably fail. To achieve maturity in an organisational data culture, businesses need to view data as a strategic asset. This requires a plan involving leadership engagement and education across the business, supported by four key pillars:

1. An analytic strategy that has full commitment from the executive team
2. A data science capability that can turn data into wisdom in a repeatable and efficient way
3. A modern technology platform to underpin the data-driven transformation
4. A programme for delivery to ensure these analytic initiatives land in the business to bring about lasting business value

Changing the culture of an organisation to one driven by data is not easy, but there’s no need to go it alone. Data science consultants can help educate your organisation’s business and analytics teams, and support the building of a data culture where analytics initiatives can be identified and delivered.

Top five business analytics intelligence trends for 2019

From explainable AI to natural language humanising data analytics, James Eiloart from Tableau gives his take on the top trends in business analytics intelligence as we head into 2019

Delivering value from data and analytics

Answer the right question

A lot of time and investment can be spent brilliantly answering the wrong question. Identifying the benefits and opportunities that data will bring, for example, increased revenue, better customer experience or efficiency, will help to create a solid foundation to drive the change.

Prove the value

As with any change, it is vital to be able to measure progress and show success. Once executive buy-in and investment has been secured for a business case, each step of the process must be supported with data. This is where data governance plays a key role – identify measurements from the beginning and track them consistently. Not only will the metrics show the overall progress and success of the changes, but they will serve as markers and checkpoints along the way to ensure that the process is effective in practice (and not just in theory).

From insight to impact

This involves transitioning the solution to production, which proves the value to the business and reinforces how it can turn data into operational acumen.

When data is at the heart of an organisation’s strategy, it will enable the fundamental culture-shift that’s needed to realise the potential of the insights that data can generate and to create successful outcomes for the organisation.

Written by Wojtek Kostelecki, data science practice lead, Mango Solutions