If your company is or intends to be data driven; you must develop a data strategy that aligns with your overall business objectives and achieve senior buy-in to gain traction.
We need to incorporate our business strategy. That is, the key business questions that we need to answer that align with the overarching business strategy of the business. We have a section which talks about business strategy in a wider context.
A core part of the overall business strategy for example may be to cut costs by X% over the next 5 years. Data, in many businesses can help us to do that by increasing efficiences through automation and it can also help us to target investment within the business more accurately, so we only spend in those areas that absolutely require it.
Next, we need to consider adoption and continued investment. So many data initiatives begin and fail to achieve continued investment and they don’t get adopted by the business. Why? Because they focus on the longer-term machine learning deliverables, which are very clever but often don’t provide the perceived business value and definitely don’t deliver a quick turnaround to demonstrate capability.
By delivering a few quick-win use cases, you can demonstrate the value of your data initiatives & drive adoption. It’s important that the use cases you deliver cater for either the masses or for some very senior figures in the business that can help drive the initiative forward and use their seniority to influence others.
Now we need to consider what data we need to achieve our goals.
Where is it?; How will we get it?; How much of it is there?; Is it streaming or batch? How will we process it?
Without answering these questions, we will be unable to: plan our infrastructure properly; assess whether we have the in-house skills to deliver the data initiatives; understand the level of investment required; plan timelines for delivery of insight and much more. This is a critical part of the data strategy definition.
Information governance is all about managing data to keep it safely tucked away in our systems and away from unauthorised individuals. To do that, we need to put systems, processes and policies in place to keep our customer data secure and to protect the reputation of the business.
We have seen a lot of data breaches in recent times, even big companies with departments dedicated to the task of protecting customer data are falling victim to large scale data breaches – damaging customer trust in the brand and incurring very large fines. Talk Talk; British Airways; and Yahoo are some examples of large breaches in recent times.
We are never going to stop data breaches, but we can work to reduce the frequency and limit their impact by implementing appropriate governance processes and policies and by driving a data culture within the business – holding everyone accountable for the safety of your customer data. Click here to read the ebook on this topic.
We must also consider the technology and infrastructure that will be required to deliver on this vision. The type of data you have & the use cases it will be used for will help drive these decisions. Never plan your infrastructure requirements, before considering your longer term vision & roadmap. Click here to read the ebook on big data architecture.
Part of this is also considering your end users. If you are expecting the CEO to be consuming your data, they probably won’t want to see a massive CSV. Rather, you should consider investment in visualisation tools like Tableau or Qlikview to give the CEO a more interactive and intuitive way to analyse the data.
Finally, we need to consider skills & capability. Do we have the skills in-house to deliver the initiatives, if not, how can we develop those skills? Code dumps can be a very helpful resource for data engineers, like those found on this website: Python, Pyspark, Machine Learning and Django
There are plenty of courses on websites like Udemy, Coursera and Pluralsight that can help to upskill your team too.