How Paddle Makes Sense of Its Data Landscape

This post originally appeared on the Paddle company blog.

Data . . . everybody wants lots of it, but often they’re not quite sure about the why. Being “data-driven,” leveraging “big data,” and measuring everyone and everything is fundamental for most businesses, but how this data is used and how it directly impacts the business can often be unclear.

At Paddle, we’re a business that creates a lot of data: every day, we help hundreds of software businesses all over the world sell their products. We gather data on every part of the purchase journey – creating tons of data points, metrics, attributes, and events that feed our business.

Making sense of this data so that it’s easy to access and proportional to our needs can be a real challenge, and has been a core focus since I joined Paddle to set up its Data and Operations practice in early 2017.

I’ll cover how we went about implementing the right tools to access and manipulate data, then some of the main data projects and how they’ve been useful to our vendors and our team.

Making Sense Of Our Data Landscape

I joined Paddle with a background in Management Consulting and a degree in English Literature – so I wouldn’t describe myself as your classic ‘data guy.’ I’ve found, however, that the real advantage of my background has been to try to constantly approach our data in the context of our business, aiming to deliver ‘data’ that is accessible, actionable, and without barriers to entry.

We had 2 main requirements:

  • One of our core values is our transparency: we share the same dashboard with our board and with our team, so it’s important to us that our core business metrics were made available to everyone, rather than hidden in a spreadsheet.
  • We needed a business intelligence/data visualization tool that would make our data come to life, and that could be easily managed, manipulated, and scaled, irrespective of the changes to our back end infrastructure.

We landed on Sisense for Cloud Data Teams – a beautifully simple data visualization tool that takes queries and scripts in SQL, Python, and R and converts them into dynamic charts and dashboards. With the ability to plug in multiple databases and enrich the insights through cross-database querying, it has proved to be a great tool for our scaling data function.

The team pretty quickly built out a selection of dashboards, covering our core business metrics (net revenues, checkout conversion, refunds, chargebacks, etc.) – and it wasn’t long before we began to be inundated by requests from the business for more in-depth analysis and additional metrics.

Building the Right Dashboards

“We have a lot of business metrics that we report at many different levels. Using Sisense for Cloud Data Teams, we report on all of these in close to real time, enabling our teams to see performance data in a transparent and clear way.”

Gerry McHugh
Gerry McHugh, Operations Manager at Paddle

To drive the right kind of adoption amongst our teams and avoid the dreaded case of beautiful dashboards that no one looks at, we’ve tried hard to ensure that there is a specific ‘Use Case’ for each of our dashboards.

For example, for our Customer Success team, we built a comprehensive overview of each of our vendors to provide immediate insights that help them build better relationships and constantly bring value. Their ability to start a call by saying “here are 3 specific ways you can increase your revenue” is extremely valuable!

For our Product function, we’ve built dashboards and reports to measure the success of new features (most recently the rollout of Apple Pay) or the results of A/B tests.

For our Risk Management team, we’re creating a suite of reports to measure the impact on our business of risky customer transactions and our efforts to manage them. All of this is only scratching the surface!

Like many start-ups, we also rely heavily on third-party tools (CRM systems, helpdesk solutions, surveys, marketing automation, ATS, payment processors, etc.) and as our teams have scaled, access to this data for reporting and trend spotting has become essential.

We’ve been working hard to pipe this data into a central set of databases, enabling us to measure our key business metrics and detailed team-based KPIs within a single system at the click of a button.

I started this blog post talking about the dangers of collecting data for data’s sake – and how it’s a dangerous pitfall for any data-driven organization to fall into.

In the beginning, we didn’t have much in place – so we’ve had to lay a lot of foundations to ensure we have a working data ‘infrastructure’, but it’s always been with particular objectives in mind.

I’d like to go over a few of the cool ‘so whats’ that having access to this kind of data has enabled us to achieve.

Building a Great Product: Subscription Analytics

Analytics platforms dedicated to subscription businesses can be expensive, time-consuming to configure and slow to accumulate sufficient data to drive actionable insights. We wanted to see if we could leverage our existing data to achieve the same thing for our vendors using our extensive historic data.

We built a prototype, covering all of the core metrics that any self-respecting SaaS business should be interested in including:

  • Monthly Recurring Revenue (MRR)
  • Average Revenue Per Account (ARPA)
  • Churn
  • Customer Lifetime Value (LTV)
  • Cohorts and movements

We built this kind of reporting for several key reasons – mainly to quickly test potential platform improvements without any development and find additional ways to add value to our vendors’ businesses.

In showcasing this data with our existing vendors, we can quickly see what is important to them and tweak them until they’re perfect, then hand them off to our Engineering team to integrate directly into our Dashboard. This allows us to stay very agile.

Making Our Internal Reporting Transparent and Effortless: Our Global View Dashboard

We have a lot of business metrics, that we report at many different levels. Using Sisense for Cloud Data Teams, we report on all of these in close to real-time, enabling our teams to see performance data in a transparent and clear way.

These metrics roll up into our ‘Global View’ – the top-level metrics that we report to our Exec team and our Board (and share across the whole company), and that indicate the health of the business at the highest level.

Up until quite recently, compiling each of these metrics was manual and time-consuming, whilst the output was a Google Sheet – not the most dynamic way to present where we are as a business!

We’ve worked to automate as much of this as possible, reducing this reporting burden to a fraction of the time it took previously, and creating a fantastic, comprehensive and dynamic snapshot of our business’ performance in real-time.

As a result, we’ve managed to move the effort away from ‘reporting’ and towards ‘analysis’ – giving our teams across the business more time to engage with our metrics by spending less time calculating them.

Managing Risk and Keeping Our Customers Safe

Using data to manage risk, detect fraud and protect our customers is something we work really hard on. Risk profile changes over time, and strong data infrastructure, tooling, and thought is key to taking a proactive rather than reactive approach to risk.

On one hand, we use dashboards to track and share health metrics, which helps us set goals and measure the impact of our risk initiatives.

There is always a trade-off between customer experience and risk: to give two extreme examples, we could, for example, ask everyone to come into our office with their passport (therefore eliminating risk but seriously decreasing revenue), or at the other end have no risk checks at all (therefore increasing conversion but enabling all kinds of potential fraud). Measuring this trade-off is not trivial and our dashboards are a massive help.

On the other hand, we dive into the data on a project by project basis, to better understand what is happening and formulate ideas on how we could improve. Ad hoc dashboards are essential to share this across teams and facilitate discussions, as opposed to cryptic spreadsheets hidden away.

What’s Next?

The three examples above are just some of the data-related projects that the Data and Operations team at Paddle are involved in right now. The function has grown with the needs of the business, and now that we’ve got decent foundations in place we’re able to spend more time focussing on outcomes, rather than dashboard building.

There’s plenty more to do. We have a lot still to do on the data infrastructure side including thinking about data warehousing, how to achieve real-time data flows in an efficient manner, and of course, factoring in the impact of GDPR on our business.

On the analysis side, we’re working to be more and more sophisticated with our metrics – moving away from retrospective reporting and thinking about how to use our data more and more predictively to drive our business forward. Definitely exciting times ahead!