Host Intro (Gene Kim)

Thank you, Gus Paul.

The next speaker is Chanade Hemming, who is head of data products at Virgin Media O2, the entity that is the result of the merger of two of the largest UK companies, with over 18,000 employees and 46 million UK customers.

When their new chief data officer was brought on in October 2020, his challenge to the organization was to find new ways to be disruptive, and Chanade jumped at the chance to contribute.

She will be co-presenting with Richard He, head of data services, on how their team of now over 130 people are creating data platforms and capabilities to help solve important business problems across the Virgin Media O2 enterprise.

They will be talking about specific data technology choices they've made, patterns they've chosen to invest in, as well as the philosophies and the engagement models they use to help elevate the use of data in the organization.

One of my favorite parts of this talk is the fabulous testimonials they've gotten on the work they are doing.

They were recently awarded the Google Cloud DevOps Award, which resulted in a pretty special congratulations message to the team.

So here is Chanade and Richard.

Chanade Hemming

Thank you so much, Gene.

We're absolutely delighted to be here representing Virgin Media O2 and the journey that we've been on in the last 12 months.

So let me tell you a little bit about Virgin Media O2 for those that don't know.

So we were born in 2021 combining the UK's largest mobile network with a broadband network offering the fastest speeds.

Now, just because I mentioned 2021, don't think that's simple.

So Virgin Media goes back to 2006 through a merger of many legacy companies, and O2 goes back, right back to 1985 through the merger of many different companies coming together.

We're actually one of the UK's largest businesses now, and we've got 46 million customers and around 18,000 employees across the UK.

And within our group that we're going to focus on today, we've got hundreds of technologists working with us and making the dream come true.

Richard He

Thanks a lot, Chanade.

I'm Richard, and I look after the data platform and product engineering team in Virgin Media O2.

So I've been here since May 2021, where we started this amazing journey.

Just a little bit more about myself.

I spent maybe 15, 16 years, at least as a software engineer in the industry.

Started with on-prem services and then worked with different cloud providers.

But in the last five years, I've been specifically focusing on helping large organizations or tech companies to transition into the Google Cloud space, and that's been quite an amazing journey.

And over to you, Chanade.

Chanade Hemming

Thank you, Richard.

Yeah, so quite similar to Richard, but I joined this team a couple of months earlier.

So I joined in March 2021, following maternity leave.

So for me, for the last, I'd say four or five years, I've wanted to move into data.

I've been at Virgin Media for five and a half years, and working in product management with software engineers for over a decade now.

When I came back from mat leave, I was looking for other things to do, and then an opportunity came up to basically set up a product management function within advanced analytics and data science, working in the Google Cloud platform, and it couldn't have come at a better time.

So within advanced analytics and data science, we've got about 130 people now, and we're growing rapidly.

Within this group, we've got people that have been with Virgin Media for some years, many years, and others that have come from external companies, so hiring talent from all across the UK.

I guess we're very much remote.

We go into the office for purposeful time.

What that has done is tapped into other people that probably wouldn't have applied for particular roles before.

The journey that we've been on has been mega exciting, and it kicked off really May last year.

But taking a step back to that, just before I went on maternity leave in 2019, we had a chief digital officer join us, and that was the first time we had that representation on the C-suite.

So when I came back from mat leave, I saw a massive shift in the behaviors of the leadership team and the commitment, and the journey that we're on, basically, and that's why we're here.

So I guess the business problem we're trying to solve is we're bringing a legacy company with very complex systems in today's world into the cloud.

And why are we doing that?

Customers' expectations are changing faster than ever.

You and I know that.

Speed at which ideas can be executed is faster than ever.

Competition is in many places, and retaining and hiring the best talent is really, really hard.

So there's just a couple of experiences that I wanted to share where we use these as examples to, I guess inspire and encourage this kind of mindset where technology, speed, and convenience can help us win, and solve the jobs that we want to do seamlessly.

If I think about this example, like Blockbuster, Blockbuster massively could have seen what was going to happen in the future if they leveraged the data that they had available to them and actually took that seriously.

So rather than using the data of sales and be like, "Yeah, we're doing great.

People are still renting videos from us." They didn't look at other trends and opportunities to predict what was going to happen.

And you only have to look at the streaming industry now.

It's absolutely on fire.

And not just Netflix there.

You've obviously got Disney, Amazon, and many others playing in that space.

Another example there would be the music industry.

So going from right back to the cassette and flipping to side B, right through to the likes of Spotify now.

We've got things at our fingertips.

From a data perspective, I absolutely admire Spotify and what they're doing, using many different data products to build this wonderful product.

And thinking about data products, like the recommendations.

So if you do like the recommendations they give you, amazing.

There's a little thing that they do there where if you skip the track before 30 seconds, they kind of learn.

So that goes back into the machine learning model, and they just don't show Chanade something like that again.

And then another example we kind of share often with people around the company to get them thinking about the technology and how it can do wonderful things is banking.

And I think one thing that I've massively seen a shift in, you've obviously gone from a physical bank through to telephone banking through to digital-only banking.

The data and the insights that you can get now through the likes of Starling and Monzo is incredible.

And you start to see legacy banks like NatWest coming through with those features.

Not quite as fast, but they do eventually get there.

So that's a bit of a summary of some of the brands and the experience that we talk about when we're trying to show having a commitment to technology and data, where it really can pivot you to in the future.

So onto a little bit about our team.

So we built a winning team, as Gene said.

Delighted that we received the Google DevOps Award for 2021, having been so early in our journey.

We're around 130 people, as I mentioned.

And I would kind of say we're like the rebellion.

We're a bit big to be the rebellion from the Unicorn Project, but really we are challenging the way that we're working today.

We're doing things differently, and we're showing people around the organization what can be done when you decouple yourself from that legacy platform that Richard is going to talk about and do things in the cloud.

I think it's been a phenomenal journey, and the feedback's been awesome.

And hopefully we can give you a little bit of insight into that and the journey that we've been on today.

So I mentioned briefly about data products.

What we're doing is we're unlocking value through the creation of data products.

And a data product really is a product that facilitates an end goal through the use of data, quoting DJ Patil there.

If you don't follow him, he's awesome.

Give him a follow.

But basically what we're doing is we're using all of the data from around the organization to unlock experiences.

It could be recommendations to a customer, it could be predicting churn.

There's many different things that it's doing, but we're using data in real-time in production, whereas historically, companies like ours and many other companies are still doing this, are just using data for reporting to look back at what happened and not to predict what could happen.

So that's the journey we're going to take you on and some of the examples we're going to show you.

So over to you, Richard.

Richard He

Thanks a lot, Chanade.

I think it's important to talk about some of these technology choices that we're making.

You can see on the screen there's quite a lot of these blue boxes is actually Google Cloud-focused services.

The most important reason for Google Cloud that we have a primary focus on is because it's data-focused.

So we've got experience in different areas using on-premise services, other cloud services, but none of these other cloud services has a focus on data.

The other reason that is really important is Google Cloud is actually behind the scenes and more open source-focused.

So as you can see, some of these product services behind the scenes, they're not actually just Google for Google Cloud's sake.

They are actually based on open source services that...

Sorry, open source product, or open source frameworks that you can use to deploy to other cloud as well.

But Google is actually the cloud provider that's actually making it run really well.

Especially, the last point, but not least, is a lot of the Google services, they are fully managed or even serverless.

So those product and services that Google offers that really give us the ability to focus on not just the technology, because technology is just a tool, right?

The most important thing is how quickly you can actually drive your business forward by delivering value.

And that's the key thing we're focusing on.

So just to dive into some of these details.

So first of all, in order to get started in a cloud journey, you obviously need to collect data, right?

So Cloud Pub/Sub is a message queue system that you can see as like a Kafka or something kind of equivalent, but it's fully managed.

Again, you don't have to manage any services or servers.

It's planet-scale.

You just send the messages into there, and you start sending it to downstream system, then collecting them.

So it is really important to focus on the event first approach.

That's why I mentioned Pub/Sub in here specifically.

As an organization, you can collect data in batches, you can have ELT jobs getting this data into BigQuery, for example.

But what I want to emphasize, that have a message queue service in your organization to start collecting a lot of the data based on events, right?

Something happened in the different kind of systems.

It really gives you the capability to not just...

It's not just about reporting analytics, it's also about building downstream subscription services based on those messages to be event-driven.

To be using those events to drive, for example, sending communications to customers or changing some of the statuses, updating customers' details.

So you can do a lot of things in real time based on the message system, which gives you a really powerful way to integrate with many teams working together.

Obviously, after your data gets collected, we need to process it, right?

In many cases, you need to even convert the data format from one to another, that you need a fully managed service to be able to process those data in very large volume.

This is where Cloud Dataflow comes in.

But again, as I mentioned, the open source first approach.

Dataflow is actually just a wrapper of the Apache Beam framework.

You can run the Apache Beam anywhere on different cloud, on-prem, but Google gives you a really kind of serverless way that you can basically chuck your workload into there, and you don't think about it.

You always have to just give it the CPUs, the RAM, how do we want it to scale, and then it does the job for you.

Right.

It's really important to focus on just the delivery in terms of deploying your services.

But the open source core, that gives you control, and then because that is not something you just have to run on Google Cloud as well.

But where it comes down to data in the next steps is also about modeling the data, cleaning the data, making sure you can actually use the data that has good quality, has good meaning, then to unlock other business values.

So this is where I'll probably say the most important thing in the entire Google Cloud ecosystem is Cloud BigQuery.

So BigQuery is a fully managed, I would say, just analytical database, but it's typically a choice of where you can put all of the data in there permanently.

What that means is you have all of your data in a central place.

This helps many processes, including the GDPR processes.

DBT is the open source tool, again, and that you can seamlessly integrate with BigQuery to do your data modeling.

What's important about DBT is it gives you a way to actually visualize and model your data based on a lineage diagram.

Each box in a lineage diagram is like just a simple SQL query.

So it takes not much time at all, typically like a week or two, for your team to learn these things, and then they can start working on production system in no time.

And this is extremely powerful.

And that you can actually enable this, especially as a data platform engineering team, to build the tooling for the individual analytics team to be self-sufficient and building those data models themselves, which are extremely powerful to introduce this more like a self-service culture in your organization.

In terms of unlocking more values, many organizations kind of stops at the reporting, maybe a bit of batch data processing, but it's really important to also unlock the value by building product and services.

This is where Cloud Run and Cloud Datastore comes in.

In a nutshell, it's basically Cloud Run gives you a serverless way to build APIs, exposed to other services.

I'll talk about why building APIs is very important later on in this, the cloud transition process.

And also Datastore is basically a document database where you can store the data in there for very fast concurrent lookups for your Cloud Run services.

You obviously can't do that with BigQuery's analytical database, but there's nothing stopping you from storing the data in a good, shaped, modeled way and ship that into Datastore as the API backend.

And in addition to that, you can also use some other services, such as Vertex AI, to build training pipelines.

And then after the training pipelines is built, now you can actually reinforce your learnings by putting this back into Cloud Run APIs with the scores in Datastore to actually do predictions.

So it's extremely powerful ecosystem when you connect some of these things together.

Also, let's not forget about reporting.

Looker is one of those services that really allows you to have the teams themselves, like analytic teams, maybe product teams, and also the sales teams.

They can use this thing called a semantic layer, which they don't have to understand SQL, they don't have to learn anything about SQL, but they get the same power to be able to build, to explore data, and build dashboard themselves.

And again, if you think about you can enable your entire organization to be self-serving and self-sufficient and building their own dashboard.

And then that is going to shift the transition into a data-driven decision organization.

This is the really serious stuff.

The few other things I want to talk about in the kind of supporting stack is the Cloud Composer is a service.

Many of you probably run ETL jobs in other areas, or ELT jobs in other areas.

I typically call this orchestration workload.

So it gives you something called a concept of a DAG.

DAG stands for directed acyclic graph, which means a dependency graph with no circles.

This is really important.

If you design your pipelines, your orchestration pipelines, you have circles, that means things goes backwards.

You don't know where things are coming from and where things goes to anymore, and then it kind of gets really mixed up.

So it's really important to have a service that can run, managed by a cloud provider, then you don't have to worry about the infrastructure.

But again, behind this is an open source framework called Apache Airflow, and that is actually the core foundation of the directed graph.

Google just gives you a wrapper to run it, but they do it pretty well.

So it saves my team, our team, a lot of the time.

So we typically spend almost like, compared to some of the other areas I've worked where we didn't have a service like this, it saves like 95% of the time.

I'm not kidding.

It's like 95% of the time in terms of cluster management, all of this infrastructure side, you just don't have to worry about these things anymore.

Also, let's not forget about the sensitive data.

In a large organization, especially in enterprises, this area is extremely important.

And this is yet another reason why I say Google is a data cloud.

There's no equivalent services, as far as I know, that does a good job as Cloud DLP, which stands for data loss prevention.

And what it can do is even with actually a recent offering called Automatic DLP, that you can basically plug it into your organization.

It can actually use the machine learning algorithms Google provided to give you the detection of what the category of sensitive data are.

What's more important is you can then tag those data, what's called PI policy tags, into your permanent storage.

In this case, it's BigQuery.

Then you can layer your sensitivity of data based on the tags and give different user groups different privileged access based on the tags.

So this is extremely powerful on unlocking.

This is not only about protecting the data so nobody can use it, this is about protecting the data that the right people can get access to it in a way that is safe and is efficient, is quick.

This is really important, I think, in organization, that not only you get the protection, but you get the speed as well.

Last, not the least, let's not forget about monitor and alerting.

So this is where we keep everything together.

The tooling is very standard.

You've got monitoring stack, you've got logs, you've got metrics.

These can be imported into, again, the BigQuery storage engine, where we can actually analyze all of this stuff.

But at the same time, when there are issues with the systems, we have multiple Slack channels, which is categorized as such as UAT alerts, production alerts, warnings, and also security.

There's different kind of stuff happens in your data products, and then you want to be able to monitor these in real time, then give you alerts if something isn't quite right, like something is going wrong.

But why did I put Looker in there as well?

Looker is not just for analytical reporting.

It's also very, very powerful for operational insights.

So let's say if you have an API you deployed, and then there's different kind of errors happening.

So you can use Looker to map into those API logs that's been imported into BigQuery.

They give you a representation layer, or you can explore those things, or with dashboard, even setting up alerts based on the analytics metrics, based on rolling windows of what is actually happening with your data based on aggregated insight.

And this is extremely powerful to rather than spending two, three hours, which is what typically happens in the old world of developers trying to crunch through the log messages to see what's gone wrong, you can transition yourself into this data-driven debugging in the development teams.

So instead of taking two hours, it will take few minutes for you to get straight down to the problem, then you can fix that, which is really important.

So the journey moving to the cloud is not simple at all, especially in a large organization where you may have even systems been there for 15, 20 years, and that is not a joke.

And many of these systems are actually priority one operational system, transactional system, that dealing with people's orders.

Dealing with when the next truck is going to be sent, or with engineers to fix people's issues at their homes.

So you cannot just replace all of these overnight or move all of these things into the cloud.

And it is really important to collaborate and working together with other teams in the organization, and to then come up with good ways, which I will talk about it in a little bit, and in order to transition organizations by bridging the two worlds together.

So as you can see, you've got, in this picture, you've got the legacy part, which is the thing I was representing earlier on the on-prem.

And then you have the new part, which is on the cloud.

So in many scenarios that you cannot, as I mentioned, just to get rid of everything.

But what you can do, and it's very important that you have this capability now, is to build your product and services with better data quality, with better scaling capabilities in the cloud.

So once you've got that, you can adopt this thing, maybe a little bit buzzword, it's called microservice-based architecture.

But simply just put it in a way that you can have APIs exposed, but actually all your product and services is built in a scalable architecture with modern ways of working serverless in the cloud.

Then you expose those APIs to your legacy systems where it is struggling to handle those workload.

So it is, in my opinion, never a good idea to try and to do a lot of this lift and shift without really thinking about what is the important thing about delivering value.

So many areas, especially on the on-prem, the reason this system exists for 15, 20 years is not necessarily mean they all just need to be get rid of tomorrow.

It's because the agility, the ability to develop and deploy is very difficult.

But by utilizing what you have in the cloud with all of this modern tooling and native cloud tooling, you can use the APIs to actually connect the two worlds together.

That while working with the teams and have really good domain knowledge in those areas, which extremely important, and then you can actually start solving these really critical problems in your on-prem or legacy environment by using cloud technologies.

So over time, this might take month or even years, but you continuously delivering value.

And over time, you start transitioning more and more of these things and lift bit by bit into the cloud to deliver the most valuable stuff.

So it is really important to think about how do we work with other teams to bridge this together, to move on with the same vision how do we transition into the cloud over time.

Chanade Hemming

Thanks, Richard.

I think that's such an amazing explanation there, I think of that.

I think just to bring that to life a little bit, if I think back to the times pre Google Cloud in this company, as a product manager, it was almost like you were a little bit in the dark and there were many handshakes with other people.

Whereas now we can be having alerts come into Slack.

We can pop into a dashboard, and as a product manager, we can instantly see what the problem is, and within five minutes we know the issue and somebody can be working on it.

And I guess it's that notion of focus, flow, and joy.

We can see the change, we can make the improvement, we can ship it without having to go to other people and fill the forms in.

So I just wanted to cover a little bit on how we work, so we've taken a leap into various areas in the company, and there's probably three areas that I'll probably talk about in where we work.

So where data hasn't been used to drive decisions is an area where we've jumped in to help those areas of the company start to use data better.

Where data is being used but it's historic data, like how did we do, not predicting what is the next thing that could happen.

So we've been shifting to a world where we've got real-time decisions in production.

That could be from everything from pricing or recommendations.

And then other areas where you think we've grown over time, there's a lot of incumbent third parties that we work with, and whilst we didn't have this Google Cloud platform in the company or this amazing talent, they did things that we can now do.

So we're starting to look at some of those areas and see where can we actually reduce cost and bring that work in-house and excite our people with these brilliant problems to solve.

Some of the things in terms of how we work, it adapts for the work.

So here's some of the things that we do as a team, and you'll all be doing these things anyway.

But in one instance, we could be working on a data product that's got that build, measure, learn lean startup approach to it where we're constantly iterating and learning.

And in another sense, we could be working on a project.

So we could work with an area of the company, let's say, I don't know, if we're working with the network area of the company to deal with all the fiber and everything that goes under the streets, into people's houses to give them Wi-Fi.

We could be working with those people on a project for two months, for example.

That project could be around how can we predict X, Y, and Z and put tools in the hands of others to be able to get to things faster.

Or it could be to spin up a proof of concept.

So if you think about us, we're all around the UK.

What we're trying to do is get the company to use data and then test what we're building in regions.

So we could choose to work in Birmingham, we could choose to work in London, and we could break that up into areas and start to run some of the data products we're building and test them out in the wild, if you will, because that's where you're going to learn the best.

And then we could scale up.

So I won't go through all this stuff, but I think the really, really massive thing for me is collaborating with the experts.

So on Richard's side, it's more with the people that are deep down in the on-prem systems and the data.

For my team, I guess it's about being translators between the technical teams, but also the company.

So spending time with those commercial managers, spending time with those people in the operations that are on the frontline talking to customers, and getting out in a truck from time to time and visiting our customers.

We work in cross-functional teams, and we have particular problem areas that we focus on or domains of the business.

So we typically have a data product manager assigned to a domain with a cross-functional team of data science, data engineering, analysts, et cetera, that focus on a common problem.

And they could be working on a product that's basically always on, or they could also take mini little projects in as well.

We speak often, and we plan frequently.

We get together quite a lot, whether it be remote or in person.

And we kind of in the last 12 months, trying to bring that product thinking into advanced analytics and data science and getting people thinking about the North Star, where do we want to get to, and then how do we take baby steps to get there.

Because often I think if you don't set that North Star, you end up veering off to the right and not necessarily going in the right direction.

In terms of our outcomes, so I guess what have we achieved?

We've achieved quite a lot in a short time, which is awesome.

It can be tiring at times.

We've definitely got less handshakes.

There's less handing off to other teams, and it's the first time in the company where I've seen the delight of a data scientist when they've had an idea for something, they've coded something, and then one of the engineers is like, it's been deployed within hours.

It's just amazing to see.

There's less forms to fill in.

There are still forms, but there are less forms to fill in in certain areas where you're working with the business.

And we're learning faster than before and gaining trust, and I think that's a really, really big one for us.

At the beginning of the journey, there was a lot of work me and my team had to do around hearts and minds, and because we'd not done some of this stuff within the estate before, it was like, yeah, okay, you've done that in another company, but we're not that company.

So we had to kind of start small, release things, get trust from people, and now we've got people queuing up at the door to do things with us, which is awesome but scary as well.

There's stuff we've been doing around our product recommendations, which has been really interesting.

And kind of driving towards that data democratization around the organization, so getting everybody in every corner to feel comfortable with data, but to get them to feel comfortable with data-driven decisions.

If you think about pricing, for example, that's quite a sensitive area, and if you're starting to have real-time decisions happening, you've got to get the humans to kind of step away a little bit.

So you've got to try and get them to trust you and what you're doing, and you can do that by proving the quality of your data and the results that you're achieving.

I wish I could tell you more.

The company views it as a valuable competitive advantage, which is a fantastic sign that we're creating value for the people who matter.

Always open for conversations outside of the conference and stuff, so do get in touch if you want to chat about anything in particular or you're experiencing any of the problems.

So I guess talking about the value they create and the impact that we're having, here's a little note from our CEO, Lutz.

He's an absolute legend.

And for the team to see this across, I think it went across our workplace.

It was literally everywhere you could look it was.

And having this outward communication, even in LinkedIn, we don't really post much there, but the company was shouting about this stuff, which is great for our hiring.

But Lutz sent a message around to us, and he was like, "Congratulations, team.

This is a shining example of high-performing team play and a huge achievement for all of us at Virgin Media O2.

For me, this is proof that we are becoming a game changer in the industry using innovative technology and new tools to support our millions of customers." We're on a mission to upgrade the UK and become a digital-first company.

So this recognition for Virgin Media O2, and specifically the teams involved, is incredible.

This went down great.

We're relatively flat, and we've got a really open leadership team, but I don't care what anybody says, I guess if the CEO comes and congratulates you as a team, that's a massive win.

So that was awesome, and we've had tons and tons of other feedback from around the company where the work that we're doing is having a really great impact, whether it's making someone's life easier, giving them things better, faster, or they're learning more.

It's been pretty incredible.

So I'm going to hand over to Richard, Richard's going to talk to you about some of the obstacles that still remain on this journey, because we are only at the very beginning, I would say.

Richard He

Yeah.

Thanks a lot, Chanade.

I can't agree more.

So it's not like things have already been done, all of this fancy stuff is out of the way, and we are moving to the cloud journeys, all finished.

It's not like that.

As Chanade, you mentioned, we are at the beginning of this journey.

We're embracing more adoptions and trying to get a lot more teams on board to working on this together.

And one of the key things I want to mention here is very difficult but it's so rewarding at the same time, is to scale the whole company.

So obviously some of the teams involved in here are more pioneers and trying, failing a bit more at the very beginning.

But it's very important to scale the whole company to do this, and that is a really difficult process.

Because there's lots of new ways of working, it's not just about the technology we choose, it's also about the way we work.

And I think one of the really toughest things, I think many would agree, is hiring.

It is not easy to hire at all.

And what is even harder is retain the people you hire.

So that is, I think, pretty much challenge I hear from many different friends and companies in different areas, that it is really difficult to do.

But I think at the same time, it's also important for us to think, okay, we must be doing something right.

When people joining us.

But it's more important to work with the people you hire as permanent employees in the company to grow them.

So when working with people in the same team, it's not all about just get something delivered and then one followed by the next one.

It might feel like you keep hitting deadlines, but it's not very efficient way to grow because people need skills, right?

They work on things, but they also the time to come down and do some trainings and to take a break, to have a hackathon together, to bring people together, to be innovative.

Try something sometimes when you don't have time, maybe not have enough time to solve this problem with normal work time.

But then you bring each other together in a hackathon, in an innovative day, where you bring so much more ideas together to solve problems that you probably wouldn't even imagine you would solve on a daily basis.

So I think that's really the key part, and it is also part of the job is to influence a much wider group of people to change together, to work together.

I think, as we mentioned, we are at the very beginning of this journey, so there are quite a lot of difficult obstacles to overcome.

But I can only say we're very confident and we enjoy it.

Chanade Hemming

I think, Richard, on that one, I think it feels like it means so much more when you're in a big company, and it's hard, and you deliver, it's big cheering moments.

I just want you to end on, this is one of my favorite quotes of 2021, I think I came across this one, and it's from Simon Sinek.

And this is the ethos that we stand by.

When we're starting to look at building something, we say to ourselves, "What's the fastest, simplest thing we could do with the highest probability of success?

Build that." And we enforcing it into the teams rather than trying to boil the ocean.

How do you get to market the fastest with the simplest thing?

And it might not be the most amazing thing you've ever built in your life, but you're getting there and you're learning.

And like Elon Musk says, "Just get me the data and let me learn." So I think for us, we're working within the current business model, but we're also looking at other ways we pivot on that business model.

And I think some really great examples we try and inspire people around the company with that is you've got Tesla with the cars, but now there's batteries to power the house.

You've got Netflix replaced the DVD, and now they're making award-winning content.

And then finally, Amazon came online to be a bookstore, and now look at them.

They've got grocery stores and everything, and Prime delivery.

So all these companies have used technology and data to pivot themselves, and they've gone to market with something that's got a great product market fit, and then they've evolved it over time.

So I think when you're operating as a tech company, you benefit from speed to market, pivoting quickly, trying new things, and your company has the ability to go past the unimagined vision.

That's what I wanted to leave you with there today.

And yeah, just thank you so much for listening.

We'll be on Slack.

We're also on other channels you can hit us up on, like LinkedIn and Twitter.

But yeah, we'd love to carry on the conversations.

Richard He

Yeah.

Thank you very much.

I think it was amazing having this opportunity to share our journey and share our experiences.

Thank you.

Host Outro (Gene Kim)

Thank you, Chanade and Richard.

So shortly after they recorded their session, they asked their CEO, Lutz Schuler, if he'd be willing to record a video that we could play for this conference.

And here is the result.

CEO Message (Lutz Schuler)

Congratulations, team.

Winning the Google DevOps Award, such a proof for high-performing team play here in Virgin Media O2.

We're all so proud about it.

And for me, that is a proof that we are really transforming to a digital-first company using state-of-the-art technology, different ways of working, and doing so much great stuff for our customers.

We are on the way to upgrade the UK.

We are changing quickly, and winning this award makes me so proud and so confident we are on the right path.

See you soon.