Data science is big business. It’s about leveraging a company’s data to optimise operations or profitability. It’s putting the big into big data and providing insight into operations that only years before seemed like something that was out of science fiction. Promises of predictive technologies that can see into the future have companies wanting more but when do you decide that your organisation is ready to take the plunge?
If you’re not into data science in some form at the moment, hiring a data scientist company or staff member could be a costly mistake. If you don’t have data warehousing, data engineering and/or industrial IoT process in place to support the data science function then you could spend all you like on data science but it won’t get you anywhere.
So how do you decide when is the right time to make an investment into data science? Below are a few indicators to tell if you get a good return on investment
Indicator 1: You’re collecting a huge amounts of data, and are ready to scale
So your probably already aware that a data scientist probably won’t be your first hire. But when your organisation has grown to a size that produces complex problems without an obvious solution, a data scientist may be your answer.
If you hire one without the infrastructure in place, i.e data warehousing, big data sets etc. then your likely to be wasting your money. You wouldn’t hire a formula one team to repair your Volkswagen, would you?
Let’s use an online self-help portal as an example. In it’s early stages you organisation is likely to be encouraging your customers to use the platform to move away from expensive face-to-face transactions to low cost digital transactions. Only once the user is onboard and using the digital services can you potentially use data science technologies to provide an even richer experience.
Imagine the above example where a user is using these digital services and are provided a recommendation engine that asks “have you also?”, and lists additional services. Of course through the power of data science we know we they haven’t :).
These features probably wouldn’t offer a significant lift to the satisfaction rating of transactions without a sizeable list of potential transactions.
Indicator 2: You already have a BI team in place
Business intelligence lays the foundation for data science. While there is some overlap between the two, data science goes beyond business intelligence on multiple levels. Data science provides programming and data management knowledge and adds value to BI by providing statistical interpretation as well as building predictive models.
If your looking for access to regular historical reports, then BI is your answer, not data science.
If on the other hand, your looking to build predictive models that forecast revenue, usage, resource allocations, life cycle costs and any other parameter that’s of value, based on actual data and not a standard increase, then you need data science.
Indicator 3: Your IT department is telling you that you need data science
It’s not uncommon for IT to make choices about software that affect the way an organisation or process will operate, like an ERP or CRM. However in many cases, the IT department is choosing this software based on their best guess, which results in lost opportunity and sometimes increased costs for the operation.
For example, someone in an IT role may choose an app based on technical specification, unaware that although the app had the best feature set, it doesn’t fit with your process.
A data scientist will ask, whats the task at hand? How many times do you y? How does it happen? What are the productivity gains? In light of this information data science will help the organisation decide on the effectiveness that app will have on the overall productivity armed with a complete set of stats to back it up.
While this level of optimisation might not matter to your business at first, there will be a day when it has a major impact. At that point
Ultimately it’s up to the decision makers within an organisation that determines the best time to hire a data scientist. But considering the investment that’s required, you should have your data collection and management processes sorted before you go looking for someone to help you with data science.
If you jump ahead, you will never capture the value offered by data science.