Alan Jacobson, Chief Data and Analytics Officer, Alteryx, outlines an evolution he says sets world-class companies apart from the rest.
Global organisations are taking a range of approaches to scale their use of analytics. This isn’t to say that every approach results in similar successful outcomes and these organisations should be aware of the optimal approach to enabling analytics.
The most advanced data analytics team found in most organisations is the centralised business intelligence (BI) team. This isn’t necessarily inferior to having a specialist data science team – but the world’s most successful BI teams do embrace data science principles. This isn’t something that we see in all ‘classic BI teams’.
As analytics best practices continue to achieve cut-through with BI practitioners, competitors that haven’t adapted risk getting left behind. The charter and organisation of typical BI need to be set up correctly for analytics to address increasingly complicated challenges and drive transformational change across the whole business
Classic BI – no longer fit for purpose?
BI’s primary focus is descriptive analytics – summarising what has happened and providing visualisation of data to establish trends and patterns. Visualisation is foundational in data analytics. The problem is that BI teams aren’t working in a set-up that’s aligned with world-class best practices. It’s often the case that BI teams are following an IT project model. They build specific reports and visualisations designed to be consumed by requested business departments. That’s if they’re consumed at all. It’s often the case that such teams are mainly judged on how quickly they can produce data visualisation and how ‘nice’ this output looks.
The BI team that follows best practice data science principles has totally different aims. It’s set up and empowered to explore data to uncover new insights and even change a business outcome.
The world’s best BI teams have also evolved past following an IT project model. In practice, this means reporting to senior business leads rather than central IT teams and being emboldened with the authority to influence broader business strategy or transformation. The modus operandi of these teams is getting under the skin of the business and driving real change and Return on Investment (ROI). That’s a stark contrast from ‘traditional BI’ which produces backward-looking work that’s siloed and disconnected from an organisation’s core strategic objectives.
Becoming world-leading necessitates centralisation
How can organisations put themselves on the course of global analytics leaders and steer away from ‘traditional’ BI and its pitfalls? They should consider centralising data functions with a simple chain of command that feeds directly into the C-Suite. This aligns data science with the business’s strategic direction, offering several advantages.
- Solving multi-domain problems
A compelling argument for centralising data science is the cross-functional nature of many analytical challenges. For example, an organisation might be trying to understand why its product is experiencing quality issues. The solution could involve exploring climatic conditions causing product failure, identifying plant processes or considering customer demographic data. These are not isolated problems confined to a single department. The solution spans multiple domains.
A centralised data science function is ideally positioned to tackle such complex problems. It can draw insights from various domains as an integrated team to create holistic solutions without different parts of the organisation working at odds with each other. In contrast, where data scientists report to individual departments there’s a big risk of duplicating efforts and developing siloed solutions that miss the bigger picture.
- Fostering talent with clear career paths
Organisations shouldn’t neglect the fact that data scientists need clear career paths. The most important asset of any data science domain is the people. If data scientists work in small, isolated teams within specific departments career development opportunities can be limited as they’re not exposed to a broader range of problems. For example, a data scientist in a three-person marketing analytics team has fewer opportunities and less interaction with the overall business than a member of a 25-person corporate data science team reporting to the C-suite.
Centralising the data science team enables a more robust career path and fosters a culture of continuous learning and professional development. Data scientists can collaborate across domains, learn from each other and build a diverse skill set that enhances their ability to tackle complex problems. Moreover, it’s easier to provide consistent training, mentorship and development opportunities where data science is centralised, ensuring that teams are fully equipped with the best tools and techniques. In this model the BI team has a foundation to upscale and with a strong data science leader will move well past BI.
- Analytics as a connector
A centralised data science function acts as a valuable connector across different parts of the business.
Let’s take an example. Two departments approach the data science team with seemingly conflicting requests. The supply chain team wants to minimize shipment costs and asks for an analytic that will identify opportunities to find new suppliers near existing manufacturing facilities. The purchasing team, separately, approaches the data science team to reduce the cost of each part. To do this, they want to identify where they have multiple suppliers and move to a model with a single global supplier that offers larger volumes and reduces costs. These competing philosophies will each optimise a piece of the business, but in reality, the outcome should be a single optimised approach for the business.
Instead of developing competing solutions, a centralised data science team can balance competing objectives and deliver an optimal solution that’s aligned with overall strategy.
- Applying analytics across domains
The best breakthroughs in analytics come not from new algorithms, but from applying existing methods to innovate use cases across domains. A centralised data science team, with its broad view of the organisation’s challenges, is more likely to spot these opportunities and adapt solutions from one domain to another. For example, an algorithm that proves successful in optimising marketing campaigns could be adapted to improve inventory management or streamline production processes elsewhere in the business.
- Champions for analytics maturity
Finally, a centralised data science function is best positioned to drive the overall analytic maturity of the organisation. This function can standardise governance, best practices and drive the change management processes, ensuring that data-driven decision-making becomes core to company culture.
Getting on the right side of change
The shift from classic BI to a centralised data science function is not just a structural change; it is a crucial strategy for companies looking to stay ahead of competition that’s increasingly data-driven.
By centralising data science and enforcing a charter for BI to independently solve key problems of the organisation, companies can tackle complex, cross-functional problems more effectively, foster talent development, create inter-departmental synergies and drive a culture of continuous improvement and innovation. This evolution sets world-class companies apart from the rest and there’s no reason your company can’t unlock the same opportunities.
