How to Create a Culture of Analytics in Your Organization

PinIt

Weaving analytics into every aspect of an organization isn’t a simple process. But when you do, the rewards are worth the time and effort.

The biggest obstacles to creating data-based businesses aren’t technical; they’re cultural. By getting buy in from everyone within the organization, your biggest obstacle is gone. But how do you get everyone aligned, engaged with the process, and excited about changes that put analytics at the center of your business?

Data-driven culture starts at the top

There are many things in life that aren’t measurable. Should you marry this person? Should you have a baby? What’s for dinner? There’s no easy algorithm for the hard questions in life. But in an organization, almost everything is quantifiable with data and metrics.

Therefore, top business executives need to ensure all decisions are data-driven. Having a culture of data-based decisions rather than ad-hoc, random choices, not only inspires trust in the data, but also provides needed justification.

This is a simple psychology concept called modeling: simply model the behavior you want to see in others. Others will observe and imitate. This is the only time the trickle-down theory works.

See also: A Call for End-to-End Data Analytics Supply Chains

Don’t think AI is all about profitability

Used correctly, artificial intelligence (AI) can increase revenue and decrease overheads. While this will make a business more money, short-term thinking like this results in small changes that don’t provide long-lasting gains. Adding a one-off analysis of customer information and finding incorrect email addresses fixes a process for one quarter but doesn’t make significant changes or add to business value in the long term.

Successful AI-driven companies view the purpose of AI as transformational. It’s an end-to-end process, with automation and AI applied to every step of day-to-day processes. It should be viewed as an opportunity to establish a continuous loop of changes, feedback, and improvement.   Doing so will help increase profitability, but this should not be the ‘why’ of AI projects.

Don’t rain on your data scientist’s parade

Organizations often have an ‘IT department’ and ‘everyone else.’ Within that IT department, there are data scientists who hold the key to business success. The best data scientists have plenty of domain knowledge and an implicit understanding of business goals.

If you don’t give data scientists the opportunity to interact with staff and analytics, your analytics won’t be as strong as they could be. To eliminate boundaries between the business, bring the experts into meetings or cycle them through other roles in the business. This will allow them to see the daily reality of the business.

By integrating the data scientists within the broader organization, they can identify gaps in processes and gain deeper insight into the business that executives may not always catch. This not only better informs the data science team but helps to get the staff on board with changes or decisions. When people are consulted about the processes they are involved in, this helps to refine and improve them, and then staff are more inclined to accept the changes to the systems as they designed them.

Giving more trust in the data and the data scientists will allow employees to clearly see the positive changes. In turn, this will make them more trusting of the analysts and will ignite a positive feedback loop, where they can suggest more places for improvements.

This process also inspires psychological ownership, where employees feel invested in the organization and its outcomes. This increases motivation and loyalty and helps to keep your people involved in refining and improving business processes.

Make sure everyone can access the data they need and empower them to use it

The democratization of analytics is one of the biggest challenges in organizations. Despite attempts to take data out of silos and make it available to all, there are still parts of businesses locked away from others. Without information, analysts can’t analyze, accountants can’t account, and marketing can’t quantify the outcomes of their campaigns.

Rather than allowing everyone access and arranging huge overhauls that never happen, organizations can open small areas of the business at a time. A marketing department’s needs will differ from accounting, so enabling employees to have easy access to the specific data they require will set the business up for success.  

Business leaders should then train their teams to use the tools just before they need to so they retain the information and are equipped to work with their data effectively. When we think about integrating data scientists more with the rest of the business, consider transitioning them from “workers” to “coaches” who can share best practices for data with various lines of business.

Quantify everything

Every organization wants to be better, faster, and stronger, but telling teams to simply do things better is not enough. Managers need teams to provide explicit, quantitative levels of uncertainty and outcomes. This does three things.

First, it challenges the data: is it reliable? Next, it gives analysts a far deeper understanding of their models if analysts have to evaluate uncertainty. Finally, it pushes businesses out past the shark nets by understanding uncertainty. Statistically rigorous experiments, controlled trials, and crossing your fingers is a good strategy for testing before making wholesale changes.

Keep proofs of concept simple and sturdy

Only 87% of data science projects make it into production. Analytics offers a lot of blue-sky-type promising ideas and fewer practical ones. However, it’s not always clear which are the actionable projects and which are not until a proof of concept is put together. And then, even if it IS actionable, the team at the top can put the kibosh on something if it’s too expensive or will take eons to implement.

Staff see the recommendations, they create projects, they make a proof of concept, and then it’s scuppered. It’s so disheartening, especially if the data supports it. Not taking action when there was a compelling opportunity to do so can negatively affect psychological safety, foster mistrust in employers, and stop that feedback loop that is so vital to establishing a data-driven culture.

Instead, the focus should be on creating proofs of concept that are viable. Start with a simple project, and it can be added to later, making that existing system more elaborate.

Use automated analytics to make employee’s lives easier

Implementing automated analytics will help drive business forward in a more meaningful way. By bringing more automation to analytics, data scientists are removing boring and repetitive tasks, the need for rework, and freeing themselves for the creative, complex tasks.

The organization must not only make the data and decisions understood to a broad audience but also explain how people can leverage data for value in their work. When analytics become useful in a day-to-day context, people will embrace them. When a staff member can stop doing the absolutely boring data entry task they hate, they’ll love AI more.

Allow flexibility rather than consistency (just in the short term)

No matter how hard an organization tries, systems evolve. Departments start using tools in different ways, different programming languages, and different data organization systems. While consistency across the business is the goal, spending countless hours training and enforcing policies that will ultimately change will leave teams feeling frustrated.

Short-term goals can lead to long-term one day, and the systems will start to coalesce with time. It’s important to maintain flexibility and adapt with the technology and processes as they evolve.

Explainable analytics

In analytics, as with most problems in life, it’s rare to have one correct answer. There are usually a few, each with different positives and negatives. What alternatives were considered? What were the trade-offs? Why was one approach chosen and others discarded? Employees are more likely to trust and use a model or system when they understand how it reached a decision.

Explaining choices made helps people to understand why choices were made and help to get buy-in. It also, if discussed early enough in the process, gives people a chance to offer explanations, alternatives, or solutions. Get that feedback loop working with open and honest contributions.

It can also go some way to explaining why some predictions fail or models break. COVID-19 changed everything about customer purchasing patterns and the wider economy. Having clearly visible, explainable AI means models are easier to fix, or recommendations can be over-ridden because a model is operating on faulty assumptions.

This also helps to quell the long-standing employees that have an attitude of ‘I know better than this.’ The data can be used to verify (or challenge) their experiences and get them on board with changes. Data + real-life knowledge = success.

How do you know when analytics are part of the culture?

Here are seven signs that indicate when the use of AI and data for decision-making is a habit apparent across every level of the business:

  • Analytics is represented at the C-suite level
  • A quality AI strategy that shows how many percent of processes are automated
  • The cost of analytics per use case drops
  • AI-driven suggestions are deployed faster
  • AI system changes are identified faster
  • There is a wider range of readily accessible resources to use with AI
  • AI isn’t tactical; it’s transformational

Don’t treat AI like it’s just the latest trendy thing a business has to talk about in meetings; make it a real, actionable strategy in every part of the organization. When analytics is simply the way of life in an organization, you know it’s been embedded and is part of the culture.

David Sweenor

About David Sweenor

David Sweenor is senior director of product marketing at Alteryx. In his current role, David is responsible for go-to-market strategy for the data science and machine learning portfolio. He shares his expertise on ways to effectively implement data strategies for businesses to stay competitive in the quickly changing landscape to transform data into actionable decisions. David has close to 20 years of experience spanning the analytics spectrum including product marketing, industry marketing, product development, strategy, BI & analytics, semiconductor yield characterization, enterprise data warehousing, IT program management and competitive intelligence. He currently leads global product marketing initiatives for advanced analytics.

Leave a Reply

Your email address will not be published.