Becoming a data-driven organization requires some form of data leadership. This conjures up the usual visions of chief data officers and elite data scientists driving the business to data nirvana. But true data leadership comes from the ability to make data leaders out of everyone in the organization.
That was the general takeaway from a panel session during a recent online corporate event hosted by Tableau, in which three data experts from different industries shared their experience with building up data leadership in their respective companies.
Fiona Gordon, global director of Business Intelligence Strategy at JLL, is a very strong believer that everyone in the organization not only can be a data leader, but should be.
“Oftentimes we get analysts who say it needs to be my VP or my director who’s making these decisions and defining the vision and the strategy behind it, but what I find is that sometimes the best ideas come from the bottom up,” she says. “It’s really important that analysts have the ability to challenge the status quo, feel like they have a voice and talk to people about how data can change things for an organization.”
Tristan Tan, VP of data analytics at Zuellig Pharma, agrees, adding that being a data leader isn’t about having a fancy title with the word ‘data’ in it. “To lead an organization with data, you don’t need that big title. You don’t need a position of seniority.
You just need a good use case that you’re solving with a good piece of data, and that you’re bringing to your team or to a management team to influence.
That doesn’t mean that organizations are wasting their money hiring chief data officers or similar titles – it means the role of a senior-level data leader is to facilitate the ability of everyone to leverage data by supporting and enabling potential data champions on the ground, supply sufficient training and tools, and finding ways to increase data velocity so that everyone can move from source data to output as quickly as possible.
“So when we think data leadership, it’s the guys on the ground championing and driving it and making the use cases, while the ones at the top are playing the support, facilitation and enablement role,” Tan says.
Governance and privacy
The other key role of a senior-level data leader is to ensure that data is properly governed, says Sarajit Jha, Chief of Business, Transformation And Digital Solutions at Tata Steel.
“Data can burn you,” he warns.
“When you start, there’s no data, and nobody cares about it. Then you start creating some systems and processes, some use cases, and you have a bit more data, and then there’s always this trouble phase where there’s too much data – everybody wants to report and then there’s 100 different reports for a single topic area and it gets messy.”
Jha strongly advises enterprises that are just starting their data journey to think carefully from Day 1 about things like agile governance, better process management and organizational structure. “If I could do what we did over again. I would definitely think more carefully about things like that.”
The other crucial task for data leaders is putting a framework in place to deal with privacy and confidentiality issues, he adds. “As data leaders, we need to play a role in talking more about it, considering the different options – is it a draconian framework or is there a better way to think about privacy that also then enables good analytics?”
Skills for today and tomorrow
Naturally, upskilling is an essential component of transforming your workforce into data leaders. JLL’s Fiona Gordon notes that her company has been investing heavily in training programs to increase data literacy and skills, but adds the focus isn’t just on technical skills – it’s also about communication skills.
“We help people level up from rookie to rock stars on platforms like Tableau and Alteryx, but as they progress through the journey, we introduce sessions where they have to actually present back to their peers on an area of expertise, or something that they’ve learned, that they think will help people who work on this journey get a hand up along the way,” Gordon says. “That teaches them presentation skills, and we help them to build their confidence. We also get them to work on the community sites for Tableau and Alteryx where they solve problems for people.”
Tan of Zuellig Pharma agrees that it’s not enough to know the technology – you also have to communicate your findings effectively. That includes understanding your audience so that you can tailor your explanations accordingly.
“You need to know who your audience is – some people just don’t respond well to data science concepts and they just need that very simple chart,” he says. “So we need to be aware and conscious of who the audience is, whether it’s an internal stakeholder, your boss or a client, and make sure that we can match expectations in terms of what analysis is required and what will be acceptable and understandable from them.”
Meanwhile, Tan adds, a related issue for senior data leaders is managing attrition. Put simply, he says, many data analysts get bored and move on to a new company that presents new challenges and opportunities.
“You have to think about how do you retain talent, and how do you keep them upskilled, because a lot of the analysts in this space are looking for new challenges and looking to learn new methodologies, new languages, etc,” Tan explains. “These are the things that drive their mind, so how do you rotate sufficiently within your organization to keep things fresh for your people? It’s one of the biggest things that you will need to think about as a data leader – how do you manage your internal people?”
Jha of Tata Steel notes that there are always new skills to learn, especially as organizations move on from data analytics to predictive analytics, which requires four specific sets of skills:
- Visualization – knowing what happened today, then drilling down to find out why it happened.
- Scenario – knowing what things could happen and what are the payoffs of each of each
- Prediction – calculating what is likely to happen
- Stochastic optimization – making predictions even when you don’t know what’s going to happen.
“Within that entire cycle – visualization, scenario planning, prediction and stochastic optimization – it’s a hard journey and you need to keep at it, but definitely that raises the data maturity and the data operating model for the entire company,” Jha says. “[But] it’s a long journey – I see it spanning over three to seven years.”
You must remember this …
Naturally, each enterprise has to figure out its own strategy to become fully data-driven. But each panellist offered their own all-purpose recommendation for senior data leaders regardless of what stage they’re at with their data strategy.
For Gordon of JLL, a key ingredient to success is to leverage the data communities out there, and build your own.
“There’s a great Tableau community and a great Alteryx community, and there’s many other meetups that go on around the place,” she says. “You can also jump onto Twitter – there’s a very active Tableau community on Twitter, and there are ambassadors you could follow just reach out and say, ‘Hey would you help me with this,’ and without a doubt you’ll get a response. Engaging with other people with great experience and networking has really helped me to understand and navigate through some challenges so that I’ve got some ideas about how to approach it.”
Tata Steel’s Jha advises up-and-coming data leaders to keep one eye on the future as data analytics becomes increasingly prediction-driven. That also means keeping tabs on new technologies in the pipeline and the possible opportunities they present.
“If you’re really an organization that now wants to lead, you really need to start thinking about how Wi-Fi 6 and 5G are here, quantum computing is going to be there – so how can I become an anticipatory organization? What can I do with blockchain to make sure that data is permissioned, trusted and secured? How can I start using the platforms that I built, and how can I onboard other people onto it and start monetizing my data?”
That said, looking to the future shouldn’t come at the expense of the present, or even the past. Tan of Zuellig Pharma says that data-driven organizations should always be focused on output and use cases. “That’s always going to be a starting point – you have to focus on a product or use case that’s going to help somebody, and you should never lose that. Even if you’re 10, 15 or 20 years down the journey, that should still be your focus.”
Moreover, he adds, data leaders should think about how to maintain the value of a product in the long run, rather than abandoning it for the next sexy thing.
“We have to be disciplined in terms of thinking about what happens with that first use case that you did – does it get decommissioned, does it get extended, what is it required to maintain and sustain it as a resource?” he says. “Everybody loves the new, shiny thing, but then it eventually tapers off and doesn’t actually have lasting value. So thinking about the idea of lasting value of the things that you do is so important.”