Patryk Hyla

Okay. Hello everybody. It's a pleasure to be here. My name is Patryk Hyla, and I am coming to you directly from Poland, Warsaw, where I am head of IT, data, and analytics. And today I have a pleasure to present you our super migration project — our wonderful journey from an old on-premise environment to the new, brand-new, completely brand-new cloud-based environment. So I will share with you best practices, our ways of working, how we made it, how it was possible for us to achieve such a huge success in our Danone organization.

You might know Danone organization from food products you can find in your neighborhood stores. Our three main categories is Specialized Nutrition, Plant-based and Dairy products, and Water and Beverages. And our fantastic mission is to bring health through food to as many people as possible. And indeed, to fulfill such a huge mission, the key role here is playing data and decision — thanks to our doctors and nurses, they can reach out every single day even to 7,000 patients, including the oncological ones. So as you can see, my department, data and analytics, is playing a key role to be part of this wonderful mission.

Also, our business would like to be a data-driven organization. They would like to — instead of going outside of organization and, for example, reaching out to vendors to conduct exercises like mix marketing modeling, pricing optimization, and so on — they are very ambitious, and we would like to do it on our own, in-house. Yeah, so that's really the mission we have. And the common ground for this mission, for the solutions we would like to implement, is of course single source of truth. So another ambition really is to make sure that all of the data we have maintained can be kept in one single reliable source of the truth. Basically, that's the name of our migration program: "One Source." So that's the code name of our program.

And believe me, it was very difficult to be part of this mission of delivering such fancy solutions like advanced analytics and so on when you have KPIs like this — yeah, please take a look — it was horrible, believe me. 77%. I will go one by one to describe you how bad the situation was on our on-prem environment. 77% — our environment availability. So you can imagine: you are a sales-analyst guy, so you can work maybe starting on Monday till Thursday afternoon, and Friday you have day off, eh — because it's like 100% sure that our environment will be not available for you, so you can go home, basically.

Another KPI: 16 hours of data processing. So another very tangible example, from sales department: we have a solution called Sales Cube where they can take a look on most important KPIs, especially for our senior management. So we can imagine — the data processing starts about 4:00 PM, it lasts the whole night, whole morning. If you are lucky enough, data can be available for you maybe somewhere between 11:00 AM, maybe 1:00 PM. And as I said, if you are lucky enough, because it might happen that our environment is totally not available. So then it means that maybe you have a window of four hours where you can really analyze and play with your data. So that's the business perspective. It was really horrific, horrible situation from their perspective.

But what was our perspective — data and analytics, IT and data, so my team? Basically 50% of tasks was incidents. So you can imagine the level of frustration in my team. Every morning after opening the laptop, the very first thing you can see is like 10, maybe 20 new incidents. You are a firefighter — you are focused on really solving ongoing incidents just to make sure that at least something can work for our business stakeholders. So that is your reality. And as a consequence of those three previous KPIs, we had like an average four months of delays for each project we've done for our business stakeholders. So that was our reality two years ago.

But there was a light in the tunnel: global decision to migrate to the cloud. And all of these beautiful, wonderful technologies can be available for you — Snowflake, Databricks, and so on. But the question is: if top technologies can do it for you, successful migration? Of course not. The top technologies don't guarantee any success. Even taking a look on this McKinsey report — I took for me the most important two KPIs: 38% of companies experienced delays in their migration programs, and 75% exceeded planned budget. So this is clear, that migration from old environment to the new one is something very difficult.

And for us — this is our local context, and I will stop here for a moment — the challenge is huge, because me and my team, we are located in IT and data hub in Warsaw, and we are supporting three separate business units: NORBAL, Poland, and Central Eastern Europe. So we have three completely different business divisions. Of course, we are Danone family, Danone company, but business looks different in these separate markets. In total, we are talking about 12 countries and 14 years of heritage with 140 solutions. And what does it mean? It means that we really had to decide, analyze what we have on our on-prem environment and choose what should be migrated to the cloud. And very often the situation was very difficult, because we are talking about solutions where documentation is lacking, where the people who are responsible for these particular solutions, they are not anymore working in Danone organization. So it was very difficult. On top of that, additional thing — from technological perspective. For example, in Poland, we were using Power BI, while for Central Eastern Europe countries it was QlikView. So we were completely not harmonized in terms of technologies, in terms of every let's say layer of our business intelligence landscape.

Another challenge: time. For this mission we had only one and a half years. So you can imagine — a lot of solutions, a lot of heritage, a lot of black boxes, and only one and a half years to migrate everything.

And that's why I'm really proud of what we did, because 25 people from my team for 18 months working on 63 sub-projects delivered fantastic results. You might remember this slide with the KPIs I presented you seconds ago. So that's the real changes we achieved. Environment right now is available almost all the time. So we have a huge improvement here. Regarding the use case with Sales Cube where we were talking about 16 hours of data processing of the ETL — we reduced it by 75%. Right now, the longest data process lasts more or less four hours. So from business perspective, from our business-partner perspective, technical improvement is huge. Also, from my team perspective, the change is visible, because we are finally focused not only on being firefighters or solving just simple tickets, incidents, being all the time under very stressful situation. Right now we are finally focused on delivering value to our business, on building really the solutions I presented you at the very beginning — with these fantastic solutions that can bring value to our business stakeholders. And last but not least, customer satisfaction: 96%. Guys, believe me or not — but I have a benchmark with other similar teams like mine, so data and analytics in our FMCG industry, and this result is massive. We are measuring this as a Net Promoter Score. So 96% is something we really can be proud of.

And everything with an impressive cost-to-value ratio, because our migration was cheaper — the cost was lower, in general about 45–50% comparing to other teams in Danone Europe. And this number — 2.5 million USD business value — this is our contribution by project we already delivered on the new platform. And I can share with you our go-live happened December last year. So as you can see, just after a few months of new platform we're able to unlock huge potential. I can assure you that this number will be significantly higher. The 2.5 is what I can already confirm, but at the end the projection is to have significantly more, thanks to new platform, thanks to unlocking this potential.

So how we made it? So finally we are going to the tangible aspects of this presentation. That's the magical formula of successful migration: we have a RAD — Rapid Application Development — plus Quality Assurance, multiplied by fantastic Team, and Business Partnering based on Value. In details.

We came back to old, good eighties for RAD, Rapid Application Development. Please raise your hand — who knows what RAD is? Okay. Yeah, so I'm surprised, because we didn't know at all. We discovered that this methodology can be our growl at the very beginning of our migration program. So RAD — we can say Agile methodology, Agile model — introduced by IBM. It's funny, 20 years before the Agile Manifesto we had a very nice methodology, how to conduct project in Agile way.

And I can share with you three most important aspects we gathered from this RAD model to our program. First of all is dynamic resource allocation between the sub-projects. Second is reusing components. And third is quick rapid feedback loops with our business stakeholders.

And just to give you tangible examples, starting with this dynamic resource allocation — you remember, 63 sub-projects, 25 people in a total. It was not possible to conduct, to organize, 63 independent project teams. Because come on, math is not lying. So we had to find a way how to really play with resources, how to allocate people smartly in between the projects.

Second aspect — this components reusability. Just to give you a tangible example, let's imagine Sales Cube in Sweden. It's almost every time the same story. We are taking data from SAP, building some data flows, data warehouse, and at the end beautiful dashboards in Power BI. And the request is very similar to what we have for Sweden in Czech Republic, Romania, Poland, and so on. So instead of building exactly the same components in four separate projects, thanks to RAD we were able to reuse these components — to take templates, to take already developed code, and to reuse it with minor adaptations for other projects, for Poland, Romania, and so on. So as a result, we had huge time and cost reduction.

And last but not least, this quick rapid feedback loops with our business stakeholders. It was also very important here. Because I believe that's the uniqueness of migration projects: that you are very often not developing something from scratch — you are using what already is existing. So the UAT was designed to really compare the numbers, the results from analytical tools on on-prem platform versus what's already being recreated on the cloud. So instead of waiting till the very end of each project, we split each sub-project into smaller parts where we were really able to discuss it quickly with our business stakeholders, check the values and so on, just to speed up the whole process of development.

So that's about Rapid Application Development. Really the change of using this methodology for us, it was really like to have a growl. As I said, it's significant time reduction, lowering cost. But to make it happen, you need to make sure that your team is strong enough to handle this methodology.

So the team is key here. And that was my strategical decision at the very beginning of migration program: that I would like to build in-house capabilities. Instead of going outside of organization and looking for vendor who is able to migrate everything, I wanted to build these capabilities inside, just to make sure that I can continue developments, I can continue building new solutions on this new platform. And what is, I believe, obvious — we had to make sure that we are able to maintain this solution, this platform.

But the biggest problem, I believe, at the very beginning was related to lack of cloud competencies. So I decided to gather experts from the market and incorporated them into our team. So this competency transfer was very successful, because we gathered people like masters in Databricks, Snowflake, and so on. So in general, we upskilled not only my team but entire organization thanks to this upskilling program, where we really used the knowledge of these new people in our organization. And technical mentoring also — believe me or not, but the average in my team is 25 years old. That's the average age of people in my team. So it's a very young team. So really, it was very important for us just to make sure that somehow we can upskill these people as well. Regular mentoring — it was not so good, because of this luxury, let's say, of having just 18 months of time. So we decided really to be available for them, especially the masters in this technology, and to mentor them on the job — on-the-job as technical mentoring.

So that was the key elements related to the team and how we developed our data and analytics team thanks to this migration program. And this fantastic side effect happened — because when I'm looking on this KPI, I believe the change is massive. 44% — that was the turnover rate two years ago, just when we started the whole migration program. Right now we are talking about 8% this year. I believe that the biggest part in this is related to change of ways of working — because as I said, right now we are not anymore focused on solving incidents, we are really able to work with our business stakeholders, business partners, on delivering the value. It's not anymore work under continuous stress where everybody's screaming "why my data is not available." So this change, I believe, helped a lot, and that's the fantastic side effect: we changed the turnover rate to 8%.

And last part of our equation: Business Partnering based on business value. So I really wanted to change the image, the perception, of our business stakeholders around my team — from being temporary assistance, recipients of requests, to really build this image of being trusted advisors and real business partners.

So how we made it? It was like really inspiring our stakeholders, always showing value, being proactive, people who are really serving them — as a business, people who are really making money in our organization. And I know it might sound like very simple stuff, but even Jung said simple things are always the most difficult one.

And I would like to stop on one aspect here. This — meetings onsite. Hope you remember our headquarter is located in Warsaw, so it's not possible to be all the time onsite with all of our stakeholders in Sweden, in Czech Republic, Slovakia, Romania, and so on. So my decision was — and that was my request to my people, really — if we have crucial moments, milestones, of our migration, please go to these people. Stay with them, spend some time together just to understand what is the pain, what is the frustration coming from, what are the real business needs. And show our human face — that we are not robots, we are not classical IT somewhere in the office, locked in the server room. We are real human beings and we can help them, we can be real partners. So I believe that's a very important aspect of changing the whole image of my team.

And believe me or not, what happened — right now when we are receiving calls from our partners, especially senior stakeholders, it's not anymore message like "Patryk, or data and analytics team, why again our data are not being refreshed? Why our KPIs again showing wrong data?" It's not anymore like that. After receiving calls, rather we are hearing messages like, "Patryk, we've heard that your team implemented fantastic solution for data analytics in NORBAL — how to make it available for everybody in Europe?" So the change is massive, and I'm really proud of it.

Our project was noticed even by major players in Poland, in Central Europe, and we had a chance to present our ways of working, our successes, our story in front of the biggest one — especially during one conference, opening of the data center, Microsoft Data Center in Poland. So for me personally, it's like a quality stamp that we know how to do our job, and it's being done in a very good manner.

And it wouldn't be possible without fantastic team. And that's the biggest lessons learned. For me, that's the thing I'm taking for the next initiative, for the next project together with me — that without this fantastic engagement, these fantastic people, such a success is not possible. You can have unlimited time, you can have millions of euro, everything basically — but without this real passion, real feeling that we are playing as a one team, we would like to win together, it wouldn't be possible, because really the migration was challenging. It was a lot of moments of frustration, of difficult discussion with business stakeholders, but we made it, and we grew up as one fantastic team during this migration program.

And here is the problem that still remains, and I'm really curious how you are dealing with such things. Even if we can say that our migration was a huge success — how you are talking, how you are convincing your, especially senior, stakeholders to invest even more? Because on one hand we can say, come on, we are on the cloud, we have fantastic results — but come on, we are IT guys. We know that we shouldn't even stop here. We should invest sometimes more, to move forward. And how you are guys convincing your senior stakeholders to really be on your side and to spend money to really move your technology, your business intelligence landscape forward, basically?

So that's all from my side. Big thank you. Thank you guys for listening. Thank you guys for having a chance to be today in front of you, basically. Thank you.