The idea of the single rockstar data scientist who can do everything is now debunked. If every person who finished university with a relevant course could come in and instantly know what to do, communicate complex ideas, and improve business results, we wouldn't have the data skills gap that we currently see. There are, of course, a few of these kinds of people around, but they are the ones who sit at the top of the tree, the ones who earn the 6 figure salaries for companies, then 7 figure salaries when they realize they can start out on their own.
This means that companies are now forced to look at creating data teams to fully implement their data plans effectively. This causes some confusion and many companies get it wrong, stacking it with people who either can't do specific roles or make the team so lopsided towards technical skills that the other side simply can't function effectively.
It makes the creation of these teams difficult, so we took a look at the 5 skillsets that you must have in your data team.
This one seems obvious and to be honest I can't imagine that many people would be stupid enough to try and create a data team without people who knew how to perform technical actions. People on the team need to be able to not only use the technology that's already available, but they also need to be able to learn about new technologies quickly and keep abreast of the latest developments in the area.
A data team must therefore have somebody who is genuinely interested and clued up on the technical aspects today and those that are going to be big tomorrow. Big data is all about having the ability to stay one step ahead of the competition, which means collecting information in new ways, analyzing it more quickly, and using the analysis for business results faster than your competitors. The industry is one of the fastest moving in the world and technology that was seen as cutting edge only 2 years ago is now essentially obsolete.
It is, however, not simply about identifying what's new, but also about what's relevant. Companies today could invest millions of dollars on a quantum computer because 'that's the next big thing' but, despite the technology being available and the power of it being fairly well documented, it would ultimately be a waste of money for the moment. The technical knowledge needs to go well beyond being able to use algorithms, they need to be technical advisors too.
Data is not something that you can instantly look at and say 'I know exactly what to do'. It is why we feed it through algorithms, visualize millions of data points, and spend hundreds of thousands making it understandable and actionable. However, understanding what to look for within the data that's already there is only half of the battle, it is also essential to look ahead at what you will be looking at in 3/6/12 months and put plans in place ahead of time to make that possible.
You can only get decent analyses from a decent data set, so there needs to be the skill within the team to identify what needs to be collected, how to collect it, and how this cold be used in the future. It is this kind of forward thinking and foresight that puts teams ahead of their competitors more than having somebody who understands how to write slightly better algorithms, because those who have the data ahead of time are the ones who will ultimately see the most benefit from it.
As much as senior executives would love to be data experts, the reality is that, while most can understand bar graphs and basic visualizations, anything deeper becomes too time consuming, especially when it comes to needing to make quick decisions. This means that a key part of any data team is having the ability to communicate complex ideas and analysis as simply as possible.
This requires a skillset entirely different from technical knowledge and vision for the future, it requires the ability to take incredibly complex ideas and analysis and condense them down to a level where people with little-to-no data education can understand and act on it. There are reports from larger companies that they have even taken journalists from business publications to do this role, because they understand the ways to communicate with high-level executives.
Ultimately it doesn't matter how much analysis a team does, how groundbreaking it is, or how much potential the findings have, if it is not communicated effectively. Data teams are always under a certain amount of scrutiny from above, especially in the early stages of their existence, so the ability to communicate effectively from it is essential to any kind of long term success and support.
Analyzing data is what a data team's purpose is, so this seems a little obvious, but the ability to analyze needs to go well beyond looking at what's in the data alone. Instead the team needs to have a real impact on the business, so this analysis not only needs to be of what's in front of them, but also how it could impact everything around them.
This requires the ability the look at the overall business or individual departments and recognize how a specific analysis could impact them. For instance, how could it be used by sales to improve their performance? How could this data promote savings in the supply chain?
It is about more than this though, it is about looking two, three, or four steps removed from the data itself and recognizing what other changes this could cause throughout the wider company. For instance, there could be a change to the marketing strategy based on data, but how will that change impact the sales department, supply chain, and suppliers? Looking two or three steps beyond the obvious is what sets a great idea apart from a poor one.
Gartner has predicted that 60% of all data projects will fail in 2017, and the blame for these failures will naturally fall on the data team which is ultimately in charge of them. Unfortunately projects fail, often despite the effort and time being put into them, but a failed project does not mean the end of the team or that people within it should hang their heads. If this happens the chances of failure in the next project only increases, so it is essential that people in the team have the self confidence and belief to continue doing what they belief to be right.
Within the first stages of development of data projects, there is often little to show externally, and even when there is, those who were promised the world at the start can often be underwhelmed by the gradual results. This all means that data teams need to be brave and resilient in the face of hostility from above. Without this it is easy to roll over or leave without allowing the full impact of data to take effect.