📊 Enabling Business with Data at Cloudflare
Learn how the Business Intelligence team enables different teams at Cloudflare by providing actionable data insights.
Hi, everyone. Thanks for joining us today. I am Uday. I'm a Principal Engineer at Cloudflare in the Data Analytics team.
And joining me is Sirisha. She is our Head of Data and Analytics team at Cloudflare.
Today, we'll be walking through about how we are bringing data engineering, data analytics, and data science together to provide actionable insights for various business teams at Cloudflare.
So without further ado, let's get started.
Sirisha, thanks for joining me on this talk today. To begin with, I'd like to hear about what your role is and how you got started at Cloudflare.
Yeah, sure. Thank you, Uday, for inviting me to this talk. And thank you everyone in advance for listening.
I've been with Cloudflare for over two and a half years now.
In this two plus years at Cloudflare, I've seen the company growing in pretty much every aspect, whether it's our employees, number of employees we have, different office locations, customers, products, revenue, and infrastructure, and so on and so forth.
My job is to really help the teams that are driving this phenomenal growth and really be able to help them continue to scale this growth with efficiency and automation and efficiency.
And being able to do that using data and insights is really what my role here in Cloudflare.
As part of that, I get to partner with almost every business function and team in the company, really partner with them on their roadmap, their business goals, and also taking a step back to figure out how different business goals come together and how they can be actually mapped to the company's overall business goals, and then translating them into data needs so we can come up with a platform, tooling, and tech strategy, and capabilities that can enable these teams to really be able to drive this growth at scale using data and insights.
So in a nutshell, what I do is connect different business teams to different data and LTE capabilities and, again, get to work with a lot of different teams at Cloudflare.
The journey has been nothing but exciting and challenging so far, really supporting the entire business, evangelizing the data and metrics, culture, and mindset with every team in the company.
And I know, Uday, you've been part of this amazing journey. And you are one of the founding members of this team.
And you have seen this journey from early on.
So I would love to hear your thoughts as well on this. Yeah, of course.
Yeah. Today, I work with different products and go-to-market teams to provide end-to-end visibility on the data from a customer standpoint as well as from a product standpoint.
However, I do remember when I joined a couple of years ago, around March 2019, we were just building our platform from scratch and getting all the data from different data points in the company.
And our primary goal was to provide financial KPIs for going public.
And for the context, we went public in September 2019.
So in this short span of just six months, we not only had our platform up and running, but we also connected the dots between different data sets across the company.
In fact, we even had to time travel for what happened 10 years ago for KPIs.
And using all that, we were able to stitch all that information in a consistent way so that we could report the KPIs while we were building the team and growing the team over time.
Speaking of which, can you talk a little bit more about our overall org strategy and how our org structure is today?
Sure, of course. So today, within the data and analytics function or the BI team, we have three different teams.
And we have a team of engineers, really data engineers, machine learning engineers, and full-time engineers, helping to build this unified data platform that unlocks not only analytics capabilities, but also programmatic and machine learning capabilities for every business team in the company.
So we have engineers focusing on platform tooling and tech stack, making sure all of the data is in one place, making sure all of that data is easily accessible using self-service tools and data products.
So that's one team. And then we have a team of analysts and data scientists.
And the analysts really focus on delivering key insights and working with every team to define KPIs, health dashboards, so they can provide visibility into the health of the company and success of the company.
So really reporting, dashboarding, key insights, and also identifying key drivers and trends that really help our company and business grow in a strategic manner.
So that's what the analysts are focusing on. And we have a team of data scientists that really focus on building machine learning models, generating predictive signals, again, to help every business team achieve this growth, whether it's customer growth or revenue growth, data and insights in a very automated manner.
So going back, just to summarize, we have a team of engineers, analysts, and data scientists really working together to deliver data and analytics as a service to every team in the company.
And that is how I think about whenever we build the teams.
Business stakeholders really don't have to have the burden or the ownership of figuring out whether they need data engineers, whether they need analysts for solving their business problems with data, or the data scientists.
The idea that they come to us, they partner with us, we help refine the problem statement with data, and we figure out what we need in terms of platform and analytic capabilities to deliver business value from this data.
Yeah, I totally agree.
I think these three are actually the core pillars for any data-driven organization.
And the best and optimal environment for each of these roles to be successful is actually to work together on a common business initiative so that you can share the learnings as well as the subject matter expertise on the data with each other.
Speaking of business initiatives, I know you mentioned about we work with different functions at Cloudflare.
Can you talk a little bit more about how we engage with different business units at Cloudflare?
Yes, sure. So I'll talk a little bit about how we engage with different right, whether it's or not.
So imagine we have teams that, engineering teams that work on building these awesome products.
We have a really good IT team that helps manage systems that are posting a lot of our data, whether it's Salesforce or Zendesk, or a lot of different systems that post our data today.
So the idea is really to take the data from these producing teams and be able to connect them to the teams that are actually consuming this data.
So we work with pretty much every team in that sense, right?
So the teams that we work with range from a spectrum of very technical to very non-technical.
And the idea is that we have the producing teams producing data, we have consumer teams consuming data, and then we take the insights from these consumer teams and then have a feedback loop into our data platform again.
So we sort of have this data platform that is connecting different teams with data in a way.
So in terms of the actual business teams, as you know, Cloudflare is a SaaS company.
It's providing all of its products and services using the SaaS model.
And for any SaaS company, there are three things that they really focus on, customer acquisition, retention, and expansion at a very high level.
So we work with our go-to-market teams, like product teams, sales and marketing teams, to really help them understand how our customers are interacting with our products, what other products could they benefit from.
So really customer engagement, health metrics, and providing customer insights to retention and expansion at scale.
So that's really focusing on the top line metrics, right?
Looking at the revenue and the customer growth. Then as we know, Cloudflare is also a public company, so we want to drive that growth.
But we want to do that by focusing on a healthy gross margin and profit.
So we work with finance in delivering KPIs and building sophisticated forecasting models, providing insight into cost analytics, profitability, and gross margin as well.
So if you think about it, it's really a very holistic interaction with every team in the company and bringing the data together in a way that we can derive value to drive the growth with a healthy margin and profit in the long run.
That is the one thing that I really love about my job because I get to interact with so many different functions at Cloudflare.
And before I joined Cloudflare, I did not know much about the company, but as I started with Cloudflare, I started using Cloudflare products on my own.
I started my own website and just trying to reverse engineer that how these things connect on our data layer.
So speaking to different functions, it actually helps me in designing the solution, which are more reusable and are less redundant for our folks.
With talking about the work, can you elaborate a little bit more about what work we have accomplished so far as a team in the past two years?
Yes, sure. So I would really categorize our work into three different areas.
So I'll start off with business enablement. So the first thing that we've enabled the entire company with is to provide a single source of truth data layer and making sure that data is accessible using self-service tools and data products.
This really helps enable all teams at Cloudflare to use our data.
And to be able to do any standard analytics or ad hoc analytics or exploratory analytics on their own.
So really focusing and emphasizing on self -service and then making sure there is a single standard source of truth data layer.
So that's the first enablement. Enablement, you touched upon this, which is to really work with the finance team to help establish and deliver our public facing external KPIs that really show the health and growth and success of our company to the external investor community.
We also continue to work with them in building risk assessment and risk analysis simulation models.
We also continue to work with them in building more and more sophisticated forecasting models, especially when it comes to gross margin and revenue and customer growth.
So those are some of the key things that we work with finance on.
Then we have go-to -market teams.
So sales team is one of the go-to-market teams that we work with. Given that sales team primarily interacts with customers, we enable them with a key data product that gives them a holistic view of our customers journey and engagement and activities with Cloudflare, which products are they using, how much are they using, how else we could really help them grow with Cloudflare.
So that's a key enablement tool that sales has today to drive retention and expansion of scale.
Then coming to the marketing team, we work very closely with them in building customer journeys, customer segmentation, also help them measure the efficiencies of different campaigns and different marketing channels.
We are also working very closely with them on different attribution models as well.
When you think about the infrastructure team, which is primarily responsible for managing our edge goalers or data centers, we enable them with key insights into resource or capacity planning, resource utilization, and demand forecasting.
So I can go on and on, but in a nutshell, there is a lot of enablement projects that have happened in the last two years.
There is still a lot more we can and we have to accomplish, but I'm really proud and pleased with the progress we have made so far supporting the entire business.
So that's the business enablement. All of that business enablement is only possible if you have the right tooling, tech stack, and platform strategy in place.
We have a great data platform that has data from several internal and external sources, all of them in place, really enabling that single source of truth layer.
But also the power of the platform is that it supports many different use cases and access patterns.
It's almost like in a capability, once you leverage it for different teams in the company.
So today, it supports analytic use cases, programmatic use cases, and machine learning use cases.
So we have a great solid foundation that can really be enhanced into multi-cloud strategy or hybrid strategy.
And our tech stack is really a combination of open source, in -house, and vendor tools.
And again, all of that is really determined or chosen based on the business enablement in mind.
So I know you have a segment coming up that's going deep dive into our platform and tech strategy, so I don't want to steal the thunder here.
But the last but not the least is really building a diverse team of data engineers, machine learning engineer, full stack engineer, analysts, and data scientists.
I'm really proud of the diversity that we have on the team so far.
So that's the third accomplishment or the award that the team has done together in the last two years.
Absolutely. I think regardless of the complexity of the work that the team has handled, what I really like is that we are delivering the insights and the outcome of our analysis in a much more concise way in a self -service manner.
And self-service is definitely the key for broader adoption and data democratization in a company.
So I really enjoy that.
And we are investing a lot in building some self-service tools for the users who are not SQL savvy and wouldn't like to write a query to get to a data.
And it also helps in minimizing the business logic because it's just you have a jigsaw puzzle and our team's responsibility is to put the puzzle together so that everyone can look at the picture.
Speaking of that, I actually like to understand we have versatile use cases and we have a lot of business knowledge that constitute to the domain knowledge as well as understanding which data sets fit together to derive what insight.
What makes this team unique or different than any other data platform team?
Good question. I think there are two things that make this team unique.
So the first thing is definitely the opportunity that we have to really partner with every team at CloudFront.
I think it gives a unique insight and perspective.
And a lot of data products, any strategy that you put together is more relevant and I think is more successful when you actually have that unique perspective, right?
Very holistic, comprehensive perspective. Imagine if you were to build a platform or if you were to build a data product that just cares to one team, you will be ending up with a lot of different things in a very siloed manner.
So I think to me, I think that helped us to be very nimble and very lean and successful as we started rolling out this platform and tools and tech and data products and data sets, insights and predictive signals.
So that I think helped us to get to a theme of build it once, leverage it many times for different needs and different use cases and access patterns.
So that's one thing. And the second piece is really, like I said, the way our team is fractured, we are really focusing on delivering end-to -end capabilities, right?
As a service to every internal stakeholder and internal team within Cloudflare.
And by that, what I mean is, so imagine you have a company -level KPI.
What you're doing is not only defining them and delivering them in a standard way, but also really establishing a framework that can translate those high -level KPIs into meaningful goals for the company and then providing insights to really drive that growth.
So you're really connecting all the pieces together like you just mentioned, right?
So I can talk about, hey, how is our company doing in terms of KPIs?
And then what are some of the key drivers and insights, right, that are driving the growth and success of this company to where could our next customer or growth could come from?
So you see how it's connected with defining a KPI to understanding the key drivers to building some predictive signals into where our revenue could come from, our growth could come from, our customer could come from, right, in future.
So that is how I see it. So that, I think, gives us, again, the leverage to build a data platform, but also use it to do that end-to-end analytic capability or build that capability for different teams in the company in a very, very standard and simple and, I think, flexible manner.
Yeah. So we talked about what the work we have done so far, how the team looks like.
I'm actually curious how things look like for 2021 and beyond, and what are some of the challenges that you see?
Yes, yeah. Now, 2021 is definitely a challenging and interesting year for this team.
And the reason I say that is maybe taking a step back, right?
So I define the data maturity in three different phases, right?
So there is a data-informed phase where the company has data. It can learn it and may use it.
And then there is a data-enabled phase, right?
And the data-enabled phase is really about you have the KPIs and the metrics defined at the top level and at every functional level.
You also have the key insights and drivers, and you may also have some predictive signals, right?
But the automation is not fully in place.
It's not systematic yet. And I can give you an example of that, right?
So if you have a propensity model, for example, generally speaking, that says, oh, this customer is likely to benefit from one of the products we have with Cloudflare.
Now, you have built a signal. The signal comes out of this platform.
And then now you have go-to -market teams that can act on the signals and run with it.
Today, that action still happens, but it may be manual, right? And then the next phase really is the data -driven phase, fully automated as much as possible.
And the decision-making is automated. It's fully system-driven. That's a huge piece, right?
And it's going to be an incremental capability that we have to build over the next couple of quarters and years to come.
But one of the things we want to do this year is to really build that end-to-end automation by making sure that we have business systems, business processes in place that can take the signals from the data platform, and we can really optimize our business and growth, right?
That scale that we are talking about in a much more automated and in a much more efficient manner.
This is where I think we have to have, if you think about the platform throwing the ball, there has got to be someone catching it, and then they may do something with that, and they're going to get the signal back.
So there is a lot of that going in that's going to improve the value that we will get out of the data over time, right?
So that is where we are going to focus a lot of our time on in terms of making sure that there is that end-to-end automation capabilities for the company.
So that's the first one.
The second one is tech, again, platform tooling and tech stack, right?
We have an amazing foundational platform set up. We have a few self -service tools.
We also have data products. But again, to really support the business with that end-to-end automation, we also have to invest a lot more in maturing our platform capabilities, right?
And I know you have a session or a segment coming up that's going to talk about what we are going to do in 2021 and beyond, right?
So that's one. And the last one, again, our business is growing. Our data needs are growing.
So we need to also grow our team to really be able to build these platform capabilities that can actually drive this business enablement further and further.
So we are continuing to hire data engineers, data analysts, and data scientists to support this growth and evolving data needs in 2021 and beyond.
Yeah, I actually want to reiterate the fact that the iterative approach of building the solution here at Cloudflare, I think it is in the DNA of the company.
Cloudflare ships out so many products, features in each quarter.
And on the analytics side, sometimes it's hard to keep up with that pace of development.
So I can totally resonate with that.
And I think that is one quality that we look for talent in our team as well, because we are not always ready to boil the ocean all at once.
We deliver things in an incremental fashion and add more capabilities with a vision that where we are going.
So speaking of hiring, I know we have some open positions for data scientists, data engineers, and data analysts on our team.
I would love to hear from you, what are some of the key attributes that you look for a candidate when we interview them for a role in our team?
So some of the things I look for is someone that's very diverse in their thinking, right?
Someone that's very open-minded, very flexible.
And the reason I say all of those things, and obviously curiosity is one of the things that we look for in every candidate, right?
That's interested in exploring an opportunity at Cloudflare.
But going back to flexibility and curiosity, open-mindedness, and being diverse, just because we are a highly cross-functional team, right?
We work with several teams. We are building these capabilities to drive the growth for the company.
We work with a lot of different business teams.
We work with the engineering and IT teams as well to understand where the data is, how it's structured, et cetera.
So we talk to a lot of different teams.
Again, they're in a range from very technical to very non -technical, which means whoever is part of this team is going to have to have those interactions.
So for me, it's important that they are able to communicate based on the audience in that meeting, whether it's technical or non-technical, and to also being able to work with business teams, because most of the data projects are not going to start with a very crisp problem statement, right?
You're going to start with a vague question, and you're going to refine that with data, which means you have to be curious.
You have to understand the business problem by partnering with your business teams and stakeholders, right?
So being able to partner with different teams, be able to communicate and understand and empathize all the different business problems that we have or challenges or opportunities that we have, and solve them with data and technology by working with your team of engineers and analysts and scientists.
So to me, you can see how much cross -functional interaction and collaboration we have.
And that's why it's important to have all those traits for someone that wants to be part of this data analytics or BIT.
Yeah. And I think, as you mentioned about the ambiguity part, that is definitely one big trait that we look for, because there is no perfect world where you will get precise requirements and hard building, and you will never change them again.
So you have to think in terms of what are you building right now, how quickly I can make an adjustment to it in order to meet the next.
And with that said, I do want to add one point that the culture of our team is very collaborative.
So even though we have people from different backgrounds and diverse skills, everyone is available to help each other in order to achieve a common goal.
So that's one thing that I like about the team.
And anyone who joins you in our team feels the same way, which is a good thing.
But I think this has been a really great segment.
I really enjoyed talking to you, Sirisha. Thank you so much for your time. And to all our viewers, thanks for watching, but to learn more about how data and analytics team works at Cloudflare.
Thank you. Yep. Thank you. Thank you for having me.
And thank you everyone for listening as well. And I hope you have a good rest of the day and rest of the week as well.
Thanks, everyone. Bye-bye. Bye-bye.