P&G's DevOps Journey

Okay, our first talk this morning is from Alfredo Colas, SVP of sales and marketing for IT, digital go-to-market at Procter & Gamble, the world's largest consumer product goods company, home to some of the most famous brands such as Pampers, Tide, and Pantene.

I met him last year and was blown away by the work that he and his teams have been doing, elevating the state of the technology practice across the entire P&G enterprise. Over 100 technologists are bringing technology capabilities to almost every geography and brand, reversing the effects of nearly a decade of centralizing those technology capabilities. He will talk about his remit and the programs that he's created.

I am equally excited that Olimpia Nitti, VP and CIO, Global Baby Care, will be sharing their story of how they modernized the consumer-facing sites for Pampers, which is so familiar to expectant parents around the globe, to better achieve their brand goals and so much more.

They have so many lessons that I think will be useful to technology leaders at all stages of the DevOps journey. Here are Alfredo and Olimpia.

Thank you so much for that introduction, and it's a pleasure to be here today with all of you and present P&G's DevOps journey. My name is Alfredo Colas. I'm the Senior Vice President in charge of sales and marketing IT, as well as leading our agile and DevOps transformation. It's a pleasure as well to have Olimpia with me, who is the CIO for Global Baby Care.

I'm sure most of you know the P&G name, but just in case, here you have a few of our brands. I'm sure that all of you have at least one of them in your house, and most likely, you have used it before joining this conference. Our company is 183 years young. We sold over $71 billion last year, and we have 100,000 employees that make sure that our products get sold in almost every single country in the world.

From an organizational perspective, we are divided in three blocks. Sector business units take primarily a product view. That is where Olimpia is based as part of the Baby and Feminine Care SBU. Then we have market organizations, which are looking at the regional and the market footprint, the elements that are mostly driven from a country perspective. And last, we have shared services and corporate functions. That's where I am, embedded as part of corporate IT.

However, our 3,500 IT employees are embedded across all and every one of these elements, and we all work together to maximize the value that is delivered to the company.

As I think of our DevOps journey, it probably started somewhere in the 2010s. I'm sure someone got a copy of the DevOps Handbook, but they really didn't go beyond the title of the book. We did consolidate our development and our operations organizations, but we didn't implement a lot of the other things that are described in the following chapters. That's what I call the DiNO years. We were DevOps in name only.

Towards the end of that decade, we had more and more teams that tried to sneak in agile and DevOps into the company, but they faced increasing pressure from processes, policies, and culture that were not conducive to really these new practices.

Fast-forward to 2020, we had a new CIO appointed, Vittorio Cretella. He helped drive a new digital strategy, and as part of it, agile and DevOps became a pivotal element to deliver value to the company faster.

I guess I had complained enough about the challenges we were facing in my team that I was lucky enough to get the position to lead this transformation across the company. And as I started talking to a lot of people across all the different levels in the organization, I encountered multiple opinions. There were those that said, "Hey, we don't have to change. We already have DevOps." Others said, "Hey, waterfall has worked well for us for the last 20 years. Why change?" But I was really encouraged as I found more and more people that were passionate about challenging the norms, that wanted to do things differently, and that actually many of them were driving the change and making it happen.

This led our key choices. First of all, we wanted agile teams to own their outcome measures. We wanted them to define their journeys. That is not something we were going to define centrally.

At the same time, our Agile and DevOps Solutions teams, and that's the way we decided to call ourselves, were going to own all the processes, all the tools underlying, so that they can really focus on delivering value. We made sure that all those tools are integrated and work together in an easy manner.

We also realized that we had to upskill the organization. We needed people to understand what agile is and isn't, what DevOps is and isn't. That's why I created a reward, and I would pay $100 to any employee that got an external certification on this topic.

We also created some gaming effects, and we created this DevOps challenge that you can see here on the right. We had a survey with over 70 questions that they could go through and see the areas where they were strong and the areas where they had opportunities. This was where they actually opened their eyes to new things that they could be doing and that they hadn't considered before.

Last, we also created maturity points, which was a bit of a competition across the different teams and encouraged them to try to become better, as nobody wants to be the last in a ranking.

Last, we wanted to make sure everyone had skin in the game. While corporate IT would be funding the resources and the cost of the platforms, we wanted to make sure that any variable cost would be paid by the agile teams. This way, we ensure that they will only be bringing tools into the ecosystem that they could see as helpful in achieving their outcomes.

Now, almost a year afterwards, we've made significant progress. We have over 120 teams that have self-declared themselves as agile. We have over 300 external certifications since I announced the new rewards. We created a community of practice with over 900 members that are actively asking questions, helping each other, and advancing the journey overall. We have also created an internal summit similar to the conference we are at here today. We had our last session at the end of April, and we had over 1,000 employees.

Our ecosystem of platforms has significantly improved. We have all these tools integrated, connected, and available for people to use, and we are seeing increasing levels of adoption across the board.

Another change we had to bring early in the journey is to make sure we had our business leaders with us. This couldn't be just an IT transformation. So we created a special training just for them so that they could understand, embrace, and support the change within their organizations.

All of this comes together as we're seeing great evidence of better outcomes. Now we're seeing teams delivering value in weeks instead of months. We see substantial reductions in the cost that it takes to deliver new releases to the market. We also have fewer incidents, and we're seeing people that are happier because they really enjoy the work they are doing instead of having to fight the ecosystem as they had to do before.

With this, let me pass it over to Olimpia, who will showcase one of the best examples in the company, Baby Care and Pampers.com. Over to you, Olimpia.

Thank you, Alfredo. I'm here to talk about the DevOps journey of Pampers.com, one of the digital products managed by my teams. I have chosen to talk about it because I believe it is a great example of how a crisis can become the catalyst of a DevOps transformation. But before I go there, let me provide a bit of background about our business and the consumers we serve.

Baby Care is a business primarily selling diapers and wipes. Our consumers are moms and dads of babies in diapers. Their age ranges typically from 18 to 35. They are millennials and Gen Z. They don't watch TV as much as their parents did, and they consume digital media, mostly on their mobile phones.

They buy diapers for two, three years, then they exit the category unless they have another baby. During those two, three years, they are very engaged on any diapering and parenting content. They actually start during pregnancy, and especially if they are waiting for their first baby, there are a lot of questions to be answered.

This is where our brand site, Pampers.com, plays. It provides parents with content and services that they are looking for. Parents register to access the content and then enter in our CRM program. Our brand site is actually operating in 60 countries and languages. The base is common to all sites, but we have also content and services tailored to specific markets. Our largest operating site is Pampers.com in the US. It has more than a million organic visits per month, and it's a foundational element of our CRM strategy.

Let's now deep dive in the platform supporting this website. It was built in 2014 with 60-plus sites to maintain, and in 2019 started to become a bottleneck to our digital plans. We had a continuous need to integrate based on our marketing plans, and the monolithic architecture was hard to change and maintain. The negative-impact testing was consuming too much capacity and our failure rate was increasing.

The application was running on the cloud since we lifted and shifted most of our applications from on-premise to the cloud a few years back, but it was not optimized to run on cloud. Our legacy CMS was creating significant traffic and costs on Azure SQL. The server-based architecture was difficult to scale up and scale out. Page loading was slow with little improvement possible. Last but not least, agile processes were not properly followed, and our CI/CD tool chain had gaps.

Net, we were slow, prone to errors, and this situation started to impact the team's morale. Every release was painful, and management escalation started to come.

Net, we needed a change, but at the same time, we didn't have extra budget for building a new stack from scratch. We could have asked for it, but I really wanted this challenge to be a catalyst for a mindset change. So the challenge to my team was: is it possible to move from a monolithic to a microservice-based architecture within the existing budget and while still delivering on website objectives?

In particular, this meant, number one, optimize to make sure we always win in organic search since this is our primary source of traffic. Number two, make sure that we kept developing the features needed to increase traffic and conversion. Number three, support new product launches planned in the coming months. Number four, use this opportunity to refresh equity and navigation and make it more effective. Number five, we needed to operate with excellence throughout the change. Needless to say, the request was very challenging.

The team came back with a proposal that we iterated a couple of times on, that basically allowed us to transition from a legacy to the new platform piece by piece at no incremental cost in 18 months. For this period of time, each website contained components of the old platforms and components of the new, and the team was very proud and eager to start.

We defined the target architecture, moving from server-based to client-based JAMstack, and we were able to achieve lower cost of scale and better performance by keeping most of the functionality on the client side and abstracting all the others to third-party APIs.

We fully adopted Agile and DevOps. In each cycle, we had a scale and innovation workstream. The scale workstream replaced the legacy features in target architecture, while innovation workstreams delivered the new feature directly in the target one. We are now at 90% done with the restage.

This is how we actually run Pampers.com today. We have organized the team in a product organization with a product owner and product teams underneath. We have invested in trainings covering Agile, DevOps, security, and cloud. We operate with a sprint every two weeks. We have modernized our tool chain, looking for automation at every step of the process. Last, we measure OKR, and retrospectively, we look at our results and how we get better.

What we have achieved so far: we have seen a much better front-end performance. The graph on the right shows the comparison with other viper websites or e-commerce platforms. We have seen improvements in SEO, which is giving us a higher traffic and better conversion rate to registration. We are faster in adding new features, the operations have been more reliable, and we have increased our productivity.

Here are a few verbatims we have received from our category. They acknowledge the outcome and the journey to get there. Let me read a couple of them.

"Pampers.com restage demonstrates growth mindset at its best. It shows scarcity mentality, thinking out of the box, and impeccable execution. We have now a website that our consumers like to use, and it's driving organic CRM acquisition at unprecedented pace."

This comment came from our Global Pampers Brand senior vice president, and is both acknowledging the outcome as well as the journey that brought us there.

The other one is, "The new Pampers.com site is super fast and with a consumer-centric design. Our organic search results are way better. Thank you, team." And this was coming from the Pampers brand manager in the US.

Now, let me conclude here with, we know that this is not the end of our journey. There are still a lot of opportunities for improvement, and we know that we need to strive every day for efficiency in cost, resilient operations, security, and of course, a website that every consumer wants to see.

I hope you found the Pampers story useful if you are going through a similar path. But what I'm particularly proud of is not only the outcome, because we are now sitting with a technology and a platform that is serving our needs in a much better way. As I said, we have more resilient operations. We are becoming more productive. We can turn around more features. Definitely, we have a state-of-the-art platform that is serving much better our business needs.

But I'm particularly proud about the journey because it would have been easy, in a way, to give the team a check and say, "Well, go and build a new platform." But actually, we decided to go the hard way and show that by keeping the budget flat, we can completely turn around the platform and really the business of Pampers.com. Frankly, this was a good idea because the team was so energized by the challenge that I can tell you, the team morale now is all-time high. So that's really, it's not just the outcome, but the journey to get there.

Alfredo V. Colás: I love this example because it's very real. It's not something small that happens in some corner of the business and nobody has noticed it. Pampers is one of our largest brands in the company. The Pampers website is one of our most important websites. So demonstrating that these practices work and operate in this ecosystem is very critical to eliminate all those people that have doubts or concerns about the ability for Agile to really play and operate in a business-critical operation.

Olimpia Nitti: The advice I would give you is know what you want. In other terms, for us, it was very clear what we were striving for, and so we knew that there were KPIs that were important to us. For Pampers.com, this was the ability to deliver a new feature every two weeks, the ability to have a highly responsive website that was really helping us to achieve our SEO score. So know what your solution platform product, digital product is for, because unless you know what the business outcome is, you will never really optimize for success. Ask yourself, "What am I in for?"

Alfredo V. Colás: I'm really proud of the full transformation that the organization has driven and how they have truly embraced this. This success is not my success or my team's success. It's the hundreds of people that have truly made Agile and DevOps truly mainstream. We are no longer that small niche, that small thing that a few people are doing. Now we are almost in the majority, so that is a piece I'm really proud of. I could have never achieved any of this by myself. It has been the team that has truly taken it even much farther than I ever thought we would be by now.

Olimpia Nitti: I really believe that the challenge is relevant for everybody because what I realize is that to be current and to be affecting the world of today, with the changing dynamics and fast pace, you really have to react very quickly to what the consumers want, what the business leaders are asking. Frankly, with the old way of building technologies and managing platforms, we were just not able to cope with the demand, which was growing and was becoming much more dynamic. So I really believe that unless you are really running a platform and a solution that doesn't need to cope with any change, and I doubt there are that many of them nowadays, I think this journey is actually relevant for you.

Alfredo V. Colás: Yeah, I think this is very relevant for everyone. At the same time, it's not something that has to happen at the same time everywhere. There are certain platforms, there are several areas that are better suited and more ready to move into this journey than others. So we're trying to pace ourselves as we get there. But to those that are the ultimate doubter, the ultimate person that doesn't really want to change, that is their choice. As someone said, change is optional. Survival is not mandatory.

Olimpia Nitti: I would also add that frankly, consumers nowadays don't just want product anymore. They want experiences. So in a way, communicating to our consumers digitally, parenting content, diapering content, is part of what now the consumers, the Gen Z of today, the millennials of today, are expecting from CPG companies. It's an evolution that definitely we will see more and more in the coming years.

Alfredo V. Colás: Yeah, I think that P&G actually has continuously reinvented itself over the last 200 years. We started selling candles. We then moved over to soap. We invented whole new categories like Swiffer or Pampers or even Pringles back in the day. We were the first company that aired a commercial on TV. We invented brand management and so many other areas. So P&G has always been pushing itself to really, truly leverage everything that is available to truly meet the needs of our world consumers. So this is, to me, just another step in this continuous reinvention journey, and I'm sure we will be as successful as we have been going through world wars, through hurricanes, earthquakes, and so many other things around the world.

Thank you very much, Olimpia. Thank you for the great story and especially the amazing results your team has delivered. As Olimpia mentioned, our journey continues. We're very proud of the successes that we have achieved, but we also know we have to push forward.

One of the elements that we're working on is how we bring more agile portfolios, how we scale our agile operations, identify dependencies across the different teams, and put all resources early on to address them heads-on.

Additionally, as every single large company, we have a very strong annual financial cycle. This is not always conducive to really operate in an agile manner. We're working with our finance friends, but definitely we haven't resolved this yet. So if anyone has solved it or has good ideas, please reach out. We're looking forward to all of them.

Last but not least, we want to make sure that agile and DevOps are truly sustainable, and we know that this requires adjusting our rewards structures. How do we make sure we are rewarding teams versus just individual results? Again, this is an area we're working with our friends in HR. We have a lot to learn here, and if you have already resolved it, please come back to us. We're looking forward to any help.

With this, let me bring this to a closure. Thank you very much for listening to our story, and looking forward to continuing the conversation with all of you. Thank you.