A Glimpse Into the Future of Analytics With Kellogg’s New Data Crunch

Justin Honaman: This isn't a new thing, but in the consumer packaged goods industry, data and analytics has been an important part in evolving. I've been involved in the industry for the last 10-20 years, and this is going to continue to be an important topic as we go forward. In the late '90s and early 2000s we were in the reporting and data collection phase, there were ballot scorecards and activity based costing. Then, this whole idea of business intelligence emerged and we started seeing a lot more data, which developed the concept of big data.

When I was at Coca-Cola, we created a committee to figure out what big data would mean for us as a business. Fast forward, we got into the cloud, and now it’s even more real-time with AI and machine learning — these are real now, they’re no longer industry buzzwords. If you choose to make a career in data and analytics, from a practitioner perspective, you’ll likely have a job in this industry for life because it's that important. Some of the biggest challenges in CPG today are around data.

First, most of the major brands have grown through acquisition. In fact, I came to AWS from Georgia Pacific, where there were 9, 10, 11 different back-office systems that didn't talk to one another, connect, easily integrate, or roll up reporting views. But that's not unique to one firm, that's most major consumer products brands that have grown over time. 

In the past, consumer data was buying syndicated data from Nielsen, IRI, or Spence, but now there’s one-to-one data because many brands have launched direct-to-consumer platforms. What do you do with that versus the other data that you have? Let's look at operations. We hear connected factories, digital manufacturing, industrial IOT — think about the data volume those create, and that is happening within a facility, one location let alone across the company. There is a challenge around data, understanding it, and the analytics around it. In our industry it creates a great opportunity for growth.

What's changed over time is the technology, platforms, and capabilities that are much more flexible to enable access to data and insights. A lot of customers focus on things like integrated data, RPA, intelligent automation for processes, BI is still important in reporting, and then advanced analytics, which include AI, ML, etc. There's different go-to-market models. Today, you'll hear how Kellogg's is approaching this: Do you create a COE? Do you put resources in the business? How do you structure that? It certainly doesn't sit in IT. There are many different opportunities and challenges around business models and processes.

Without further ado, Garrett, I'll tell you, why don't you share a bit with us about your background and how you got into the whole data in the analytics space?

Garrett Byrne: Thanks, Justin. It's such an exciting topic and I'm delighted to be here to chat with you. My background, like everyone else, has been pretty varied. I'm based in Ireland, where I'm lucky to be surrounded by a huge number of multinational companies, including Kellogg’s, which has European operations here. That makes Dublin and Ireland an exciting place to work, full of opportunity and bright talent. I've worked across a number of industries, but primarily in financial services and consumer goods, delivering digital transformation programs.

Thinking about the roles and companies where I developed my passion for data analytics, I'd start with AIB. AIB is a pillar bank in Ireland, which was an incredibly challenging, but exciting time in my career. For context, when I joined the financial crisis had hit and the impact on Ireland and Irish banks were particularly difficult. AIB at that point, had just been bailed out by the Irish government and were under pressure to streamline the organization — reduce costs, restructure bad debt, and ultimately, return to profit. The approach to meeting this challenge was to focus investment on fundamentally changing how the bank operated, as well as how it engaged and supported customers — effectively, moving the bank to being digital first.

Ten years ago in Ireland, pillar banks thought the idea of digital first was quite foreign, certainly in the early stages. The first step in this process was to rationalize the branch network and deliver self-service machines, where we started to move the customers from being incredibly reliant on tellers and branch staff, to being able to self-serve. As you can imagine, this was a huge cultural change, for both the branch employees, but also for customers. 

However, with every challenge, there's always opportunities. What else did the transformation bring? It gave us an amount of new, fresh data on customers. The move to self-service and around the clock access to bank services allows us to better understand customer needs, which in turn provides better products and better services.

From this, we continued to invest in delivering new digital products, such as developing one of Ireland's most used mobile phone apps. We introduced Apple and Google Pay to the Irish market and built Ireland's largest payment engine — those are just a few of the things we did. As we put the banking and everyday services into the hands of customers, we continue to gather data and understand customers more. There’s a significant volume of data associated with financial services that the bank holds, including customer demographics, online shopping habits, how much customers spend on fuel or meals out — it's a treasure trove of information.

A good example is that I used to travel quite a bit pre-pandemic. The bank would know that I travel maybe twice a month, I typically fly with Aer Lingus and go to the U.S. They know instantly that I need an annual travel product that covers the U.S., but would also cover both work and private healthcare. Having that information is invaluable. They're able to offer that tailored service to me. They can target it quite well, knowing that I'm in the airport and haven’t paid for a travel product, then they can prompt that engagement. I receive a text, “Would you be interested in a product? Take yes, for more information.”

With all of this rich information, the bank had to significantly invest in data and analytics capabilities to be able to leverage that data. 

In my case, it was travel insurance, but it could be saving for a mortgage for the last number of years and now you're eligible, or whatever it is. Seeing this in action developed my passion for data and analytics.

Guffanti: That's a great background. I love hearing the stories of adjacent industries using data and analytics, and how we can apply them to ours. When you came to Kellogg's, where did the journey start for you and for Kellogg's? How is it evolving?

Byrne: I've been with Kellogg's for about six years. When I joined, the data and analytics muscle was being developed. I've had the opportunity to work on a lot of great projects with great teams, all of which have been focused on maximizing the rich data that we have. One of the first roles I had was working on the Pringles integration, where we integrated the Pringles business into Kellogg’s, following an acquisition. We doubled the amount of data overnight, there was a great amount of harmonization work, and huge opportunities from a data analytics perspective.

I also looked at procurement technology, where I focused on finding ways to get technology to support and enhance the procurement function and within the global data and analytics function, where we're heavily investing on that data and analytics muscle. Looking back, data and analytics has been key within all of the roles that I've had, so when the opportunity came up to help deliver our new analytics platform, I jumped at it.

Honaman: That's a great challenge and opportunity where you can see the impact the team can make. You mentioned “next-generation analytics,” what does that mean at Kellogg's and how do you apply that thinking?

Byrne: That's a great question. We talk about analytics, transformation, digital transformation, and other terms that are across industry and completely another buzzword. Next-generation analytics is no different. It can mean a variety of things depending on who you speak to, but to me it refers to the use of big data, machine learning, AI, and visualization to easily digest large amounts of data, gaining real-time insights and then predict outcomes

Honaman: As you were creating that program, in terms of the structure and approach — you mentioned procurement earlier — was it focused on any specific area or was it more broad-based when you were putting this together?

Byrne: Next-Generation Analytics is enterprise wide. We are focused on pushing Kellogg's toward the data and developing an insight-driven organization. As we're looking at use cases and different opportunities to prioritize, the maturity of data and analytics differs across the organization, but we want to make sure that we prioritize and focus our efforts to maximize the return as quickly as we can.

Honaman: You mentioned your ambitions for this and thinking about what the potential is, what does that look like as you think about this platform?

Byrne: We see the data analytics journey as being a key enabler to delivering business strategy, which is why we're investing in a new cloud-based analytics platform, Cortex. Cortex is part of the wider NGA program and will do a number of things for us, including improve performance, increase capacity to manage big data, decrease total cost of ownership, and ultimately, enhance speed to insight. 

At Kellogg’s, we have a large volume of data and the focus is ensuring that the right data is available to the right people, at the right time. There are many opportunities to further utilize the rich data we have. As an example, the unique consumer data — interactions, transactions, behaviors — is one of the greatest assets of any business.

It's a huge volume of information. When added with second- and third-party data we start to see the value of data. When you harmonize those and marry these data sources is when you see that transformational things can happen and turn those insights into growth opportunities.

Honaman: We don't see a lot of loyalty programs, as you know Garrett, in the industry. But then, 33 million households on The Kellogg's Family Rewards program — that's fascinating. There’s an entire discussion in itself around what could come out of that.

Guffanti: For your fellow consumer goods viewers who are inspired and wondering, can you make it practical or talk about specific cases where you truly have leveraged data and analytics to impact the business?

Byrne: A great recent example of how we combined rich and varied data, then used algorithms, machine learning, and human intuition was the analytics work that was done during COVID-19. Not to bring up the pandemic, we all want to move on from that, but it was transformational for us as people, as consumers, but also the impacts it had for companies. 

When the pandemic first hit, we saw schools, businesses and restaurants closing. We looked at the data on our consumers and married that with the information on buying patterns, the economy, and consumer sentiment. We looked at different pandemic scenarios and generated early hypotheses about how the pandemic would impact consumer and retailer behavior.

Interestingly, and not surprisingly, one of the things we noticed was how people's emphasis on surface cleanliness was changing. For example, a Pringles can or box of cereal — the amount of people walking up and down aisles, picking it up, reading it to look at the ingredients, etc. and then putting it back down — was a huge possibility of cross-contamination. We wanted to understand how that was impacting consumers and may permanently change the relationship with packaging and cleanliness, or hygiene. In response to this, we shrunk the size of our packaging and saw sales increase over the previous period, which is extremely interesting.

That's just one example of the insights generated during the pandemic, but the team kept going 24/7. We looked at production, logistics, our people, statistics, everything we possibly could look at, we did. 

Honaman: Garrett, before today, we spoke quite a bit about personalization and revenue growth management (RGM), which have been a big focus of solutioning at scale. Can you talk a bit about that and how that's played out?

Byrne: These are huge opportunities for us. If we go back to the core thinking, data and analytics is one of the key enablers for us as a company. With technology ever-changing, ever-evolving, the future possibilities are going to be limitless. 

We're going to be able to identify new ingredients and recipes, influence and expand packaging opportunities, and even develop brand influencers. As I mentioned, personalization will also play a pivotal road role in that future of food journey. Personalized recommendations are built on user data, such as interests, past purchases, search history, and much more, which allows us to have a complete understanding of customers and make useful predictions about them.

It's important to understand customer engagement at every point in that journey, to carry powerful analysis and build offerings based on their preferences. As you can imagine, for a company the size of Kellogg's, personalization isn't something that's easy to accomplish. That's where the rich data and analytics and AI capabilities come in. The benefit of AI is that it's going to help create the personalization that our consumers want and crave, but at scale.

Guffanti: We've already discussed many ambitious projects that you're involved in, but you're only one man. As you mentioned, Kellogg's is a huge organization, who are you working with within the organization to partner and accomplish these initiatives? How do you get them to engage with your vision and approach?

Honaman: That's a great question. We are in a lucky position where we are focusing our energies behind the Next-Generation Analytics program and have developed an industry-leading platform. That means we have stakeholders from across the organization knocking on our door, seeing how we can partner together to deliver great solutions. But it's a really good point. Our very close partnership with stakeholders across Kellogg’s is key to the success of this program. 

Guffanti: How do you prioritize that? You have business stakeholders coming to you and saying, "We need to figure this out." We've already talked about a slew of opportunities and challenges, but how do you set the priority? How do you align with the business stakeholders?

Byrne: Prioritization of analytics use cases should never sit within IT. The prioritization needs to be made from a company perspective. As these use cases come in, we help to understand those requirements and translate those into analytic solutions or analytic-specific business problems. 

As we do that, we look at each use case through six lenses:

1. Is it aligned to Kellogg’s strategy? 
2. Is it delivering significant benefits, etc.? 
3. What are the resource requirements on it? 
4. Are we bringing in a partner to help support the build of it? 
5. Are we using internal Kellogg’s resources? What does that look like? 
6. Do we have the capacity to do it? For example, is it funded and are the timelines what we would expect them to be? 

All of those things help those prioritization discussions with the organization.

Honaman: That was something I was going to ask — you’ve got requests coming in from everywhere, how do you prioritize and almost determine value use case potential benefit, which is a big challenge in a big organization. Garrett, in terms of technology, certainly it's accelerated because you are modernizing and that's a big part of why we're working together. Can you talk about that and how that's played out for your team?

Byrne: The Next-Generation Analytics program is going to help enable Kellogg’s to become a data-driven and insights-fueled organization. It does that by modernizing data and analytics capabilities. 

We've structured the program around our enablers — data, technology, process and people. I’ll talk through a few of those to give a bit of color to it.

Starting with data and technology, we need to ensure that we have the right data available for users in Cortex — as pure and simple as that. We need to have a great platform with the right information in it. For us, this involves migrating current data from constrained and expensive on-premises solutions to a new cloud-based platform. There will be huge benefits here. 

We have unlimited storage, faster performance, an advanced suite of tools and technologies — all at a lower cost with a market-leading provider, AWS. Leveraging the breadth and depth of cloud capabilities will help us transform the way we use data. 

In terms of the platform itself, we need to think about how we can make sure that platform is a success. One of the tools we're putting in place to support the platform, and further drive the democratization of data, is a data lock. This is not a unique problem to Kellogg’s, but many data users have trouble finding the right data they need quickly. Not just finding the right data, but understanding how best to use it. 

For example, I'm looking to find data on an ad campaign, but there might be hundreds of files that are similar. The file I think I want may have the right attributes I'm looking for, but I'm not sure how it differs from one of the other files — that’s where the real value of the data catalog comes in. We'll have the data curated to understand the source of data, how frequently it updates, who currently uses it, what Tableau dashboards it feeds, what's the data lineage from source to consumption, and all of that is stored in the catalog.

That puts the data in the hands of the user. It ensures that we're using the best data possible, and helps increase speed inside by reducing the time taken for data discovery. We see a huge unlock with the data catalog. We're continually looking at the technology ecosystem and what new tools and technology are coming to the market to ensure that we can meet the evolving needs of users. We don't simply want to respond to these needs, we want to be predicting, enticing the users with new tools and ways of working. 

In terms of the process enabler, we're heavily focused on the process side to ensure that we've set ourselves up as a function, but also the platform up for success. We've developed new guidelines, new standards, governance, and ways of working to keep the platform safe and secure, and to ensure we maintain agility and speed for end users. 

I’ll talk through some of the governance pieces, starting with the data governance roles, which are central to what I've spoken about in the catalog. We're introducing three new analytic data governance roles across the business, a critical step in driving the full value of the data catalog and helping us understand our data. 

There will be three roles:

1. The data sponsor. Who's going to champion a data-driven culture, find funding for data initiatives, and set that strategic direction. 
2. The data owner. They approve access requests for confidential or protected data, ensure data solutions meet the intended use, and are ultimately accountable for acceptable data quality levels. 
3. The data steward. They are the business point-of-contact for data questions. They are going to be in the catalog, curating the different data objects. They define and monitor the data quality and lead to the resolution of any data risks or issues.

These roles are a huge unlock for us. We’ll roll those out as we migrate the data into Cortex and activate the different use cases. Most importantly, going back to that point around governance enlists a sense of fear in people. The roles and governance will continue to evolve and mature across the organization, and that will help support our ambition to become an insights-driven organization.

Honaman: Garrett, at any industry event everyone likes to talk about the shiny objects — AI, ML, blockchain, but master data management and data governance, doesn't get a lot of attention. However, they’re very important and key to enabling what you're doing.

Byrne: 100%. It is a foundational layer. Just think, freedom without governance is chaos. You need to have a basic level of guardrails, clarity on how to best engage with the platform — all of that is required to maximize the value. There is no point in investing huge amounts of money in a fantastic industry-leading solution, for it not to be used correctly — it just doesn't make sense.

Guffanti: Justin, as you know, at CGT we speak with a wide variety of consumer brands — large, small, all different categories. From your perspective — not only working with Kellogg's, but across the industry — is there a different approach for larger or smaller brands depending on the size of the organization?

Honaman: That's interesting. It's not a matter of approach, it's a matter of function of the organization. We work with the vast majority of major CPG brands globally on their data strategy, analytic strategy, and others. I happen to focus on food and beverage, which is where I’ve spent my time AWS. Let’s put Kellogg's over here for just a moment. Think about the SMB brands, the digital natives, those entering the food and beverage space as new entrant startups.

It's fascinating, they are all starting on cloud, that's where they've launched. They don't have the data integration challenges, and they're able to move faster in terms of being able to grow and leverage learnings from the business — as they're getting online and e-commerce and eventually getting into retail stores. That's where you start seeing the complexity and the challenge, especially with the new growers.