by Ana Maldonado, Data Science Practice Lead at Altius
Most organisations are on board with the fact that data is one of their most valuable assets. All over the globe, companies aspire to be data-driven. In fact, increasing competition and market pressures will soon mean that all organisations will need to innovate and be data-driven up to a point.
The problem is, there’s a massive gap between wanting and needing, and actually achieving. All organisations have the potential to be a data or AI-driven business, but there are a few pre-requisites that need to be addressed first. I’m talking about your data, your strategy, your company culture and your team readiness.
Your data strategy
The first step to a successful data project is having a robust data strategy – one that aligns your future vision for the use of digital data, with your business goals. If you don’t have a clear data strategy yet, that’s not a problem. Altius can help you. Together we can develop a detailed roadmap for a transformation focused on capturing and utilising digital data to improve business intelligence and empower your organisation to drive growth.
Then we will need to investigate and prioritise which data-driven projects your organisation will focus on. This is achieved through sessions that involve different stakeholders from across your company. Together we will look at your data sources, data availability, as well as the quality of your data. We’ll also look at the business challenges that need to be solved and clearly define your objectives. We will then help you to rank your data projects according to feasibility (data availability and data readiness) and impact (a concrete measure of value).
At the end of this process you will have a list of innovative (and prioritised) data science use-cases. These are projects that we can get off the ground and will deliver value reasonably quickly. We will also guide you on how to engineer scalable algorithms and deploy them for production use. This gives you good leverage when you need to justify your Data Science/Machine Learning (ML) investments to the business.
Is your data ready?
Key to the success of any data science project is not only understanding where your data is, but how available it is. This involves an assessment of how your data is being collected, stored, integrated and employed. Many people think that having everything in the same place is enough. That doesn’t necessarily make the data ready for modelling.
Typical things we assess are:
- ‘Completeness’ – will the volume of missing values affect the data’s usefulness?
- ‘Structure and distribution’ – does it represent reality and cover the modelling space we would like to describe?
- And finally, ‘historical availability’ – how far back does the data go, and do you have a duty to erase any data after a fixed amount of time?
It is important to note that the “getting your data ready” step is incredibly time consuming. Many of our clients are surprised to learn that a good chunk (70-80%) of a data scientist’s job involves finding, getting permissions, accessing, extracting, completing, curating, structuring and cleaning the data in preparation for modelling it.
If your data isn’t ready, then I’m afraid there’s work to be done. If you are serious about being a data-driven company in the long-term, this is something you will have to take ownership of and remedy. But be reassured, the investment in time pays dividends in the long-run.
Is your company culture ready?
It doesn’t matter how fired up and excited you are about the innovations you’re about to launch, if your team aren’t fully onboard. It’s a common problem. We often come across business owners and business unit leaders that love the idea of being data-driven, until they realise it requires them to share their data with other units or departments. The power of data multiplies when aggregated and integrated. Successful data science projects require full transparency, and all too often this is quite an alien concept across a business.
Preparing for dealing with this situation is something we can also help with. During the strategy and use-case brainstorming sessions, we will assess the level of transparency and openness that your business and people are receptive to. We will flag potential issues and help manage expectations. If your people or business units are not aligned, then we will try to identify common goals that they are happy working towards together.
Do you have the right team in place?
We help many organisations get started on their data driven journey or to continue on a growth path. Naturally a lot of businesses aspire to taking this role in-house eventually, but they underestimate the costs and effort required to do so. If your long-term strategy is to be a fully data driven or AI business, simply employing a couple of data scientists is not going to cut it.
Applying Machine Learning (ML) or data science in different areas of the business (for example in finance, HR or marketing) requires a diverse team of data professionals, including data platform, data scientists or ML engineers. They will need to work together to create minimal models (MVP) that “live” on your infrastructure. They will need a productionising and implementation mentality and they will need to work closely and iteratively with the problem and data owners to create tangible value.
The costs involved in creating this environment can be considerable. But if you don’t get it right, you’ll end up with Proof of Concepts that lead nowhere. The reward of doing it right are extremely high, not only from a business perspective, but also on the human side. When people work together on a solution design, they learn from each other and adopt the solutions more willingly, which multiplies its business value.
Being an innovative data-driven company is achievable for most businesses, but there’s a lot of work to be done in preparation. For the most part, the effort will primarily focus on ensuring that you have a consistent data strategy and a winning use case with a clear value. Your data sources need to be accessible and your data primed and ready to be modelled. Preparing your business culture and team environment to be open and receptive to data centralisation and sharing is also a key factor in becoming a successful data-driven company.
We are passionate about helping clients to achieve their data goals and that involves properly preparing you for the journey ahead. We have worked on data initiatives with organisations across the UK, Europe and USA, to help them build compelling information strategies and road maps that deliver real, enduring value. If you’d like to talk to us, then please get in touch.