The data science strategy – part II


The eagerness of how data science can help businesses ‘achieve’ still seems be high on the agenda amongst business and IT leaders that are looking for ways to leverage their data. In fact, many are of the view that ‘if this process can help solve our business challenges and create new opportunities, let’s definitely hire someone knowledgeable’. Sounds good, right?

Following my previous blog post on the data science strategy – part I, where, you may be feeling motivated and inclined to advertise a vacancy online for a data scientist candidate – looking for ‘the one’ – I would still caution against being gung-ho in this approach. Of course, there is value in the role of the data scientist. However, in most cases to truly achieve the desired business outcomes will require a team – a team with the technical capabilities and the communication skills to translate the data in a way that c-suite and frontline staff can understand.

In our experience, this is where many businesses and organisations get it wrong. They look for a candidate – a data scientist – with an expectation that this candidate (alone) will be all things data related to the business. Unfortunately, this often occurs even if a company is given the option of creating an A-team – that is a team which understands the key business drivers of the organisation and has the communication skills to articulate the data findings to the business – many still opt for a single data scientist. An individual can not replace the capacity of a team.

So, what then are the final steps your IT team can take, to finish setting up and implement a data science strategy?

The A-team should expand the impact of data science

We know that business are wanting to improve decision making at all levels, from the c-suite to the frontline level – which often stems from wanting to eliminate emotional decisions and build a stronger analytics culture. While this may not be suitable for every department, the organisation can still explore technologies that can align and provide departments with the freedom of Business Intelligence (BI), analytics and data discovery. In this way, your A-team can still take the lead in terms of growing data and then fitting it into advanced data science and data interaction for nontechnical users.

Give access to all the data

This means just that; allowing the data science team access to all the data. This will allow the team to extract new insights in real time. However, never underestimate that new insights may be different for each organisation (as per the data goals that have been outlined with the team). In fact, the team will often tell you that having all the data (semi- and unstructured data) allows them to give a solid analysis, especially given that sometimes there is rapidly growing data which may have been marginally analysed, if at all. The problem is, in many organisations data owners are reluctant to share their data for the greater good.

Improve on data governance

If an organisation wants to minimise the associated risks of possible data thefts, hacking, online and geolocation tracking reports that we read about daily on the news – then data governance is a must. We know of far too many brands who have suffered great losses – particularly through reputation damage – because  of security breaches – all of which they never saw coming. So, it is then imperative that the data science team, while working on the data of the organisation, ensures that data is stored, accessed and used in a manner that aligns well with the data protection regulations of the company.

David Stodder says that, “data science teams, along with business leadership, must be cognisant of the right balance between what they can achieve through advanced analysis of consumer data and what is tolerable—and ethical—from the public’s perspective” – and I share the same sentiments. Data governance policies are a strong safeguard in protecting sensitive data through a data science process, as only anonymising the data may not be sufficient.

Of course, I understand that this kind of data science strategy may not apply in all businesses or organisations across sectors and industries, but I do know that data science is a team sport and in most cases cannot be undertaken alone.

Concluding remarks

Much like the whole is greater than the sum of its parts. So, if you want to ensure your business is not lagging behind, implementing this type of data science strategy will assist in stronger continuous growth – as you will now have a team that actively engages with product managers, product designers, and engineers to devise appropriate measurements and metrics. In addition, while they focus on their strong points, they will be able to back each other up too. Sounds good, right?


Leave a Reply