Effective data analytics can be a competitive differentiator today, giving companies keen insights into customer preferences, product development and usage trends, and market gyrations that others don’t see.
To get the most out of their analytics efforts, businesses need to assemble a highly talented team to make sense of all the data that’s coming in from multiple sources, and to translate the analyses into real value for the organization.
What does it take to create a crackerjack analytics team? Here are some key best practices provided by experts.
Just the right mix of expertise
Broadly speaking, a high-performing analytics team needs to have three basic skills: technical data skills to empower the team, analytical skills to drive the work itself, and business skills to ensure the right work is being done and that it’s driving business value, says Dan Magestro, senior manager of advanced analytics at consulting firm West Monroe.
“Very few people have this full set of skills,” Magestro says. “In fact, what I just described is truly a unicorn. But a team of people with these three elements can be even more effective.”
The technical data skills can be provided by those who understand how to organize the data. “These are likely the more traditional IT employees, and including a couple of them on an analytics team can be a powerful ingredient to team success,” Magestro says.
Data scientists bring the validity or “science” to an analytics team, Magestro says. “The hard skills of a data scientist are very important,” he says. “We’ve also found that solid problem-solving and critical-thinking skills are incredibly important, often more than deep experience on a certain platform.”
And a dedicated business expert on the team can be essential to making sure the information analyzed is relevant to the business, Magestro says. That person would effectively communicate the insights back to the larger organization. “The business expert bridges the frequent communication gaps between the analytics work and business need,” he says.
“At least one expert needs to understand REST [REpresentational State Transfer] and how to effectively retrieve data over REST,” Bowers says. “At least one expert needs to understand relational databases and how to get data using ETL [Extract, Transform and Load] tools and file exports. The team needs at least one expert for each type of database,” including SQL, NoSQL document, and NoSQL wide column.
Put a strong leader in charge
Perhaps most essential is making sure the team is led by someone who not only understands the importance of analytics and how it works, but who has a keen grasp of the needs and goals of the organization.
“It takes leadership that can effectively integrate, orchestrate, and broker all of these requirements and teams working together to produce a top-notch analytics team capable of initially producing, then scaling accordingly to meet today’s analytics requirements,” says James Burke, director of IT sourcing and digital advisory services at consulting firm ISG.
“Entrepreneurial business and technology leadership at upper and mid-levels that is empowered [through funding] to ‘fail fast’ is a requirement,” Burke says.
When Sangeeta Edwin, director of business intelligence at Rockwell Automation, was named to head the analytics team at the company, her first step was to clearly identify the team’s goals and objectives, and to align stakeholders.
“To build my analytics team, I worked with our executive leadership team to align strategies and define the team scope, goals and timelines,” Edwin says. “If alignment doesn’t happen up front, analytics teams are prone to volatility. At Rockwell Automation, data analytics occurs at all levels; both in our own production from the plant floor to the enterprise level, and across product development from hardware to software and services. I needed alignment companywide.”
The leader of the analytics team should have a strong grasp of the company culture, and incorporate that into the team. “For instance, is your company fast-moving, or do stakeholders need to understand the implications of a decision before moving forward?” Edwin says. “You should assemble a team around you that fits with the wider company climate and characteristics.”
Many companies require analytics teams to break down technical language into business terms, Edwin says. “However, at Rockwell Automation, most of my colleagues have an engineering background, so the technical details motivate them and get them excited about analytics,” she says.
Get data storage access right
A world-class analytics team needs secure, reliable access to resources such as data hubs, data lakes, and data warehouses, Bowers says.
The data hub “loads data as-is for unbiased and unfiltered analysis,” Bowers says. “It indexes data on ingest so that queries return in sub-seconds across terabytes of data for rapid data discovery, analysis, and data wrangling.”
The hub simultaneously indexes data in multiple ways for search, hierarchical queries, flat queries, graph queries, and semantic analysis. “It can track data lineage, ensure data governance, enforce security, transform data, cleanse data, and filter shared data,” Bowers says.
Data lakes, such as those provided by Hadoop, are batch-oriented and enable data analysts to run jobs to discover data. “The turnaround time is not ideal for data discovery, but batch jobs can process data using any algorithm, including machine learning,” Bowers says. “It is great for bulk transforms of data, so it can be loaded into a data warehouse.”
And a data warehouse is ideal for taking results from the data hub and data lake and delivering information to business users to answer questions in pre-defined contexts, Bower says.
Break down data silos and connect data to business value
It’s somewhat of a cliché, but to get the most out of the analytics team, organizations need to knock down walls between departments and eliminate data “silos” that prevent groups from sharing valuable information with each other, and with the data analytics team.
“The key to a successful analytics team is knocking down barriers within an organization to create a data-driven culture,” says Matt Hogan, senior director of engineering, analytics and reporting at McGraw-Hill Education, a provider of educational services.
“Many of today’s organizations exist in silos with their data, processes, and reporting,” Hogan says. The analytics team at McGraw-Hill Education is divided into three main roles, “which has opened up incredible doors within our organization,” he says.
One set of roles includes data scientists, who essentially serve as the research and development function at the company. They develop new models and visualizations that feed into McGraw-Hill’s product pipelines. “They are the backbone developing new processes, uncovering new insights, and getting deep into our infrastructure to keep us scalable and efficient,” Hogan says.
Another set includes data engineers, who help drive business value out of data. “All of their insights and reporting typically start with one business-focused question, where they are conducting analysis to get that question answered,” Hogan says. “They then make connections as to where those insights can be leveraged in different parts of the organization, or when certain questions and answers start to create patterns. It’s imperative that our data engineers operate in an agile fashion to extract data, build visualizations, and pivot as needed at any point.”
The third role includes front-end engineers, whose primary goal is to deliver value to the company’s products based on insights that exist within the data. “They make the connections between the analytics we uncover and what is best for the product,” Hogan says.
It’s also important to communicate with business teams regularly and from the beginning, Edwin says. “Everyone around you needs to buy into an analytics strategy,” she says. “In the end, they’re the ones that are going to have to use it. The best way to involve others is to understand how they are motivated, and then present the value of analytics in those terms. Ask yourself: ‘Why would that business team give up their legacy systems for a modern analytics platform? What value will they get from changing?’ Once they understand the value, they’ll be motivated to change platforms.”
Keep team members motivated
It’s a good practice to keep team members engaged and excited by developing challenging goals, Edwin says. “Show them that innovation, training, continuous improvement, and evolution are part of the fabric of your team.”
At Rockwell Automation, the analytics team began with data reporting analytics. “Then, with some stretch goals, we dove into machine data analytics. We then created an entire data hub and an IoT [internet of things] platform. Each time my team creates something, they must innovate and evolve. Don’t let your team sink into a comfort zone.”
One of the best ways to keep the team motivated is to deliver value to the organization. “Even the best teams will fail if they aren’t delivering analytics solutions that work,” Edwin says. “Proving the value of your analytics solutions, even incrementally, will help motivate the larger workforce to stay in your corner. After all, your work should make everyone’s jobs more efficient and improve revenue streams. Prove you can make that happen.”
Once the team delivers solutions, it’s time to market its achievements. “You must invest time in showcasing your team’s successes,” Edwin says. “My team continuously presents our accomplishments and key learnings. This has helped us grow our network and excite our colleagues at Rockwell Automation. Marketing our work brings pride to the team and the larger company, inspiring future successes.”
Expand the team outside the organization
“A top-notch analytics team, at this stage of market maturity and talent/expertise availability, is a combination of the best of a several organizations,” ISG’s Burke says.
“This is because very few organizations, if any, have the breadth, depth or scale of talent, data and technology resources to achieve this on their own,” Burke says. “It takes a ‘value chain’ of technologies, teams, and organizations to achieve effective analytics, or ‘effortless information’ for an organization.”
This includes the support teams for the various platforms or services ingesting the data, hosting the software, and crunching the numbers; and the developers programming the software and algorithms needed for the desired analytics.
“Roles that are part of this value chain of analytics team capability include platform suppliers, platform support specialists, agile/DevOps teams, analytics, database and data science specialists, business process and product/information owners, and marketing [and] end-user experience specialists,” Burke says.
Related data analytics articles: