Data and analytics empower executives to enhance decision-making, manage risks, optimize costs, strategize effectively, ensure compliance, and drive business growth.
In a recent executive leadership panel, Sue Pittacora, Wavicle’s Chief Strategy Officer, and Laura Branch, Wavicle’s Director of Analytics, sat down for a discussion moderated by Marie Winters of the Chicago Finance Exchange for an in-depth discussion about the role of data and analytics in today’s world and the ways to monetize data to drive competitive advantages.
They covered fundamental data and analytics topics; the critical roles of data privacy, governance, and compliance; real-life examples of data and analytics in action; and the future of analytics and generative AI. Dive into this recap to explore how data and analytics empower executives and what’s coming next in the analytics world.
Sue Pittacora: Corporate strategies today increasingly include information as a critical enterprise asset. When you look around, many companies are using data as the key to driving their strategic competitive advantage.
A few big changes have prompted this. Data storage costs have decreased, opening up new possibilities for businesses. Companies can now tap into third-party platforms to gain insights into customer experience and satisfaction. In addition, data collection has become a lucrative industry of its own.
However, we find ourselves in a paradoxical situation: while we have access to vast amounts of data, we often struggle to extract meaningful insights. In the past, obtaining data was a challenge. Today, the challenge lies in protecting and utilizing data effectively.
This is where the role of the Chief Data Officer (CDO) comes into play. Initially focused on safeguarding data, CDOs now have the responsibility to not only protect and govern data but also to derive valuable insights that drive company growth. We stand at a pivotal moment where data is not only a valuable asset but also a catalyst for generating high returns. Executives today must embrace the importance of effectively leveraging and governing data to drive value and maximize return on investment.
Laura Branch: We’re going to be talking a lot today about data, data ecosystems, and analytics, so it’s important for us to set a baseline.
A data ecosystem has many parts. At the heart of these ecosystems, there is a multitude of data sources, ranging from databases, spreadsheets, and externally acquired data to real-time streams from sources such as network monitoring or cybersecurity tools. Those feed into a storage area where you keep and protect the data and manage who has access. You also need the ability to process and govern the data, so you can identify who handled it and assess data quality. From there, you can analyze and visualize the data. Everything from the original data source to the final report is within the data ecosystem.
We think about data analytics in two primary parts: analytics enablement and business intelligence (BI) analytics and insights.
Analytics enablement is crucial for effective analytics. To ensure that data can be properly leveraged for insights, data must be managed in a way that considers data storage, collection, and architecture so that everyone is receiving information from the same sources. Furthermore, platform management plays a key role in terms of infrastructure, user management, and development. Data owners must be able to trust that their data is safe and secure. Lastly, data governance, privacy, and compliance are essential for protecting data and maintaining its quality to ensure that the stakeholders who will make decisions based on analytic results can trust that the information they are receiving is accurate.
Laura: Sometimes when people hear the term “governance,” they think of regulations or governing bodies. However, in the context of data management, governance refers to how data is managed to ensure its integrity and trustworthiness when providing insights.
Concerns about “dirty” data are also addressed under data governance. For instance, when acquiring a company, understanding and cleaning their data as it is combined with the broader organizational data can pose significant challenges. Every company manages data differently, and when some data sources include fields that have been manually input or proprietary systems that do not clearly share what is contained in each data set, finding a clean solution can be a harrowing ordeal. One solution that many organizations are leaning towards lately is a data catalog, which allows you to track the origin and lineage of data, identify trustworthy sources, and manage the quality of data at different levels of granularity.
Remember, analytics can only be as good as the underlying data that feeds into it.
Sue: The rise of AI represents a significant development that companies cannot afford to ignore. It is crucial for businesses to be prepared to embrace AI by establishing a solid data foundation.
Poor data governance has a massive impact both on AI and advanced analytics initiatives and on simple reporting. In my past role as a leader at McDonald’s, I often faced questions like “How many Big Macs were sold last year?”
That seems like it should be a simple question, but when I looked into the data, I discovered that in the top 10 countries alone, there were multiple different point-of-sale systems and ways to process a Big Mac transaction, which could appear as a Big Mac or a Big Mac Extra Value Meal, among others. The many different data streams delivered the data in different formats, which made answering what should be a simple question surprisingly challenging.
For me, this experience highlighted the importance of data governance and reinforced the need to advocate for it.
Laura: When discussing my work with family, I like to use a baseball analogy to explain analytics versus advanced analytics.
Years ago, baseball cards displayed simple stats such as strikeouts and RBIs without much in-depth analysis. These basic stats focus on counts and straightforward trends.
With time, more advanced analytics emerged, enabling us to analyze trends over a season or multiple years or against specific teams. Now, with the use of advanced analytics, we can provide a far more granular examination of players, determining, for example, their likelihood of striking out on a Tuesday against a certain pitcher in an afternoon game.
These mathematical techniques provide insights that were previously unattainable. However, a successful data science project requires a combination of different skills, including individuals with business or domain knowledge who can validate the data, coders who can execute the analysis, and mathematicians who can determine which analytic techniques are suitable for specific datasets.
Cleaning and preparing the data can be time-consuming and costly, but the most crucial factor in successful data analysis lies within the expertise of the individuals involved. They must possess the necessary knowledge and be immune to external pressures or biases that may influence the desired results.
Sue: Strategically, it starts with prioritizing your organization’s data needs. You can’t boil the ocean. Start with the data behind the KPIs that drive your business. Then, peel back more use cases that help you reach your goals.
Sue: I have a great example from my past where, in my previous role at McDonald’s, there were questions about how we would continue to grow. We asked ourselves if we had reached our production capacity. To address this, we installed cameras in kitchens and drive-thrus to assess peak hours.
We discovered that the bottleneck was in the order-taking process. While there was available capacity in the kitchens and on the grills, we couldn’t ask customers to place their orders faster. Through analytics, we determined that drive-thrus worldwide were slowing down by a few seconds on average. This may not seem significant, but even small delays can have a crucial impact.
Working closely with the innovation center, we came up with the concept of side-by-side drive-thrus, which ultimately revealed the potential for approximately a billion dollars in sales worldwide, all thanks to an analytics project.
Laura: What Sue just shared is an amazing story with a massive top-line impact, but it’s not unique. I once worked with a federal agency that was constantly engaged in large multi-year projects. The agency’s director expressed their frustration with issues arising, schedules slipping, and the lack of timely awareness to resolve them. They requested the development of a predictive model to detect delays earlier and prevent wasteful spending.
The agency had the advantage of 30-40 years’ worth of paper records to analyze. We developed models and some patterns emerged. Projects with specific contractors or certain events tended to lead to undesirable outcomes. Although challenges still arise occasionally, the duration of delays and associated costs have significantly reduced thanks to this application of predictive analytics.
Laura: Generative AI recently became popular with ChatGPT. It refers to leveraging AI or large language models (LLMs) built on vast amounts of data to generate results. Generative AI can be utilized for various purposes such as writing code, blogging, and even generating PowerPoint slides. It has the potential to perform tasks previously done by humans, although some of the generated content may not be impeccable or accurate. Automation presents significant opportunities in areas like invoicing, payroll, staffing, coding, marketing, medical image review, and pattern detection.
However, we are not at the point where humans can be completely excluded from the process. While generative AI offers speed and convenience, having domain experts to ensure efficacy is crucial. Similar to the “garbage in, garbage out” principle of big data, AI’s reliability depends on the quality of the data it is fed. Some people may view AI as a technology that can be readily deployed, acting on its output with confidence. However, it is important to note that you may not know the sources or validity of the data generative AI models are using.
In the future, businesses will have customized AI models. This approach allows businesses to combine their trusted data with other reliable sources, enhancing control and trust in the outcomes. This direction highlights the importance of focusing internal efforts in this domain, and it’s something we’ve been working on at Wavicle.
Laura: Years ago, this was true. However, with the availability of open-source AI models, the scenario has changed.
Now, the most expensive aspect of using AI is hiring skilled individuals who possess the necessary expertise. The technology itself is largely open source, resulting in minimal costs. Although data expenses still exist, such as access, storage, and management, those costs are integral for any data project, not just AI.
Although we currently lack extensive data points for comparison, venturing into the realm of AI analysis is the next logical step beyond data science. In terms of expenses, it should be comparable; nonetheless, it requires the expertise of more costly professionals.
Laura: How you store your data is significant, and one opportunity for cost-efficiency can be achieved through the implementation of frugal architecture. It’s possible to reduce expenses by utilizing a unified system, such as AWS or Azure, which allows for smoother operations at lower costs. It’s also important to note that additional licensing fees for other tools can contribute to higher expenses.
Leveraging AI capabilities can also contribute to cost reduction. With a skilled AI practitioner, automation of data management and acceleration of processes can generate significant cost savings.
Sue: We’ve also seen an interesting use case for manufacturers through IMMEX tax incentives from the Mexican government. These incentives allow for import and export tax savings for goods that are manufactured locally in Mexico and exported. Surprisingly, many companies fail to capitalize on these opportunities due to the rigorous data requirements involved.
We helped a global manufacturer execute the data engineering they needed to be able to generate the necessary reports, and they expect to save millions each year, just using their data. With advancements in data extraction from systems, utilizing these incentives has become a much more feasible endeavor.
Sue: First of all, having a strong data analytics partner can help you ramp up quickly and ensure you have the knowledge and skills you need on your team. When seeking a partner, look for a team with expertise in all the relevant areas and with whom you feel comfortable. It is crucial to find a thought partner who can contribute to every aspect of your endeavors, ensuring a comprehensive and long-term partnership.
In addition, it’s important to not just look at what you need now, but also at how to prepare for the future. What AI, data, and technologies would you like to be using down the road? Make sure you’re setting yourself up for long-term success and building a future-proof foundation today to support future needs.
Ready to get started on your analytics journey? Get in contact with Wavicle’s experts.