How Amplitude Created a Self-Service Data Democracy According to Their Executive VP of Product
Interview series
Stef Olafsdottir

Stef Olafsdottir

August 11, 2020

How Amplitude Created a Self-Service Data Democracy According to Their Executive VP of Product

Imagine the worst cafe rush-hour line you’ve ever had to wade through to get your latte. Now, instead of coffee, imagine that at the end of this hellish line was the data you needed about your products, pipeline, or customer behavior. Each time you need your data fix, you have to get up from your desk, walk down the hall, out of your house or office, and then wait in a queue behind hundreds of other people.

At the end of that line, instead of baristas, sit your data analysts, handing out the data that each individual person asks for. If anyone ends up needing more context, they have to get back in line and wait all over again.

If this sounds like a nightmare, that’s because it is. But it reflects a real constraint for many companies who haven’t democratized their data.

The future isn’t a data dictatorship—it’s self-service data democracy. One where, instead of waiting in line at a cafe for coffee, you can walk up to your own kitted-out setup in your office and get exactly what you need yourself.

Amplitude is the perfect example of this data-like-home-cafe approach to modern business. So I sat down with Justin Bauer, Amplitude executive vice president of product, to get some insight into how the company created a self-service data democracy that bridges the art and science of product development.

“Just like anything, [building a self-service data democracy is about] people, processes, and tools—in that order.” — Amplitude Executive Vice President of Product Justin Bauer

Sounds simple enough. But, below, we’ll take a closer look at what went into building the self-service data democracy that underpins one of the best data products around.

1. Amplitude’s culture is built around a North Star.

A North Star metric is a single bar that everyone on every team at your company can use to measure whether or not your product is living up to the hype. It's the main support pillar that holds up your self-service data democracy.

The idea isn’t new, but it’s one that Amplitude has taken to heart and made their own.

For Bauer and the Amplitude team, their North Star is weekly learning users (WLUs). A WLU is an active Amplitude user who shares a resource they created with at least two other people within a week. Basically, it measures both initial user engagement, and whether or not users find the product worth sharing.

“It really ended up being the perfect metric for us to really reflect on what we were working on,” Bauer said.

But setting their sights on this guiding light wasn’t enough to align with it. Amplitude needed to create a culture and a system of habits that chased WLUs at every level of the company. To do this, Bauer and his team did three things:

  1. Put into words the most basic promise of their product to understand what all other work should be in service to.
  2. Identified what one metric most signifies whether or not their product is useful to users. The performance of this metric was a binary tell of whether or not they’re following through on their promise to customers.
  3. Created a constellation of metrics that surrounded their North Star to build out the net of a self-service data democracy and give each team a way to make progress toward their North Star.

Anyone who has heard of Amplitude knows that their main promise is to make it easier to build good products through data and collaboration. This mission clearly pointed in the direction of the company’s North Star, but the Amplitude team needed to come up with a constellation of metrics that would align work at every level with this guiding goal.

These supporting metrics were in direct service to boosting WLUs and gave each contributor a way to measure their performance and impact, without the need for data analysts or engineers. No more waiting in a long line to find out how products and features were performing.

“We really think of it not just as necessarily a single metric but also a constellation of metrics that all help guide you towards your product strategy,” Bauer said.

By filling their proverbial skies with metrics that all supported the one common goal of driving up WLUs—and, therefore, improving how worthwhile users found the platform—they built a culture that was wholly aligned with their main promise.

As a result, each team’s daily work furthers their progress toward their North Star metric and fits the transparent data-driven product strategy that supports it.

2. They make sure everyone understands the data.

It’s all well and good if your employees have access to the data they—and others—produce, but they can’t make strong business decisions if they don’t understand what that data means. Without informed citizens, your self-service data democracy will shrivel up.

Back before the days of companies like Amplitude, the only people who understood data were the ones who created it, mainly the aforementioned engineers and data analysts. This created our nightmarish cafe scenario from the beginning.

Thankfully, Bauer and his team, like many industry leaders, took the secondhand gold of why data dictatorships suck and did the opposite. They intentionally and consistently educated their teams to ensure that they knew where to go for data, how to read its insights, and how to integrate those insights into practice.

The way that Amplitude began to foster this understanding is the same process for how they maintain the knowledge today:

  • Give each team and contributor a micro metric to focus on that lines up with daily work and supports the overall North Star metric.
  • Educate team members on the tools available to them to pull and read data insights.
  • Encourage team members to look closely at data to answer everyday tasks and experiment with new ideas.
  • Anchor all product decisions to data, and clearly communicate how the work of each team affects higher-level goals.

By making sure everyone from sales to engineering understands the data they’re creating—and knows how to create high-quality data through data governance (as we’ll discuss below)—Bauer and the Amplitude team empowered their democratic citizens to make the best decisions possible.

3. Amplitude empowers engineers to use data themselves.

The truth is, while engineers are an integral part of building a great product, they’re traditionally seen as the last part of the implementation waterfall. This cuts off at the knees the decision-making power and insights of developers, and it’s time that changed.

In a self-service data democracy, Bauer said, engineers need to be empowered to use data themselves. Doing so lets them see the direct impact of their work and helps them make better cases for features they think would make the product stronger.

“Engineers can actually even use metrics to try to fight for the areas they want to invest behind,” Bauer said. “It's a great way to actually democratize who's even making product decisions.”

So, to make both engineers better and the product stronger, Amplitude handed engineers the key to their data democracy. This not only made their engineers happier but also helped them be more engaged in the work they were doing, Bauer said.

“Now they are highly incentivized to actually do the instrumentation because they want to see if the work they're doing is actually helping improve those metrics and helping the business grow.”

But this wasn’t simply a matter of granting engineers access to data tools. The process, like that of educating the company as a whole about data, had to be done intentionally through the following steps:

  • Bring engineers into the strategy and goal setting early on so everyone is on the same page.
  • Educate engineers on how Amplitude measured success so they can measure their work in relation to the North Star metric.
  • Develop a well-thought-out tracking plan to ensure analytics implementation is done consistently from the beginning.
  • Empower engineers to use data tools to see how their product decisions affect performance and experiment with new features.

When Amplitude’s engineers were given a seat at the table and empowered to make the case for new features and improvements through data, the whole organization was able better fulfill its promise to customers.

4. They focus on creating great governance.

Good governance is at the heart of all successful democracies and ensures that people have access to high-quality data and that processes are in place to continue this in the future.

“Leading companies are starting to realize that, in a world of data democracy, governance is important,” Bauer said. “If you invest behind that, then you get the dividends of self-service data democracy ... [But] it actually takes a lot of work to govern data well, to ensure that the people have the right levels of access and what they’re accessing is high quality.”

Rather than restrict contributors and only give them access to the data directly related to their jobs via analysts, Amplitude created a detailed governance strategy that ensured the free flow of consistent, correct information.

But this didn’t happen overnight. Good data takes hard work and mature governance.

“You can’t just get democracy for free.”

Amplitude “paid” for good governance by investing in and enforcing data best practices, standardization documents, and tracking plans. These efforts helped them create a clean, single source of data truth that evolves with the company as time goes on.

“There’s a lot more focus on data quality, and people are starting to really treat it the same way you might treat something like code quality,” Bauer said. “You actually have to have even higher data quality to ensure that people make the right decisions.”

To create governance to guide every team within the company, while granting access to data at every level, Amplitude worked through the following steps:

  • Audited the current state of tracking and governance to understand what was already in place.
  • Created a map of their customer life cycle and the metrics that support it to ensure that their customer life cycle complemented their North Star metric.
  • Assigned a point person who would be responsible for ensuring best practices were followed.
  • Created and enforced a schema for data creation and tracking standardization so each team in every department would know how to create and maintain clean data.
  • Frequently looked at their data governance and revised it so best practices fit the needs of the company at each phase of growth.

This gave everyone in the organization the means to find and work with data well through standardization documentation. This good governance made it easy for citizens of the company’s self-service data democracy to measure success, learn quickly, and pivot with data.

5. Amplitude runs lean and dogfoods their tools.

It shouldn’t be surprising that one of the best tools in Amplitude’s toolbox is, well, Amplitude.

“We’re heavy users of our own products,” Bauer said. “We obviously do most of our analysis within Amplitude and then share those charts through things like notebooks. ... We’ll use JIRA to make sure that we’re actually project-planning around data instrumentation appropriately.”

This approach doesn’t just help Bauer and his team put their product through its paces and integrate feedback back into the platform. It also enables them to run leanly and iterate often, both of which keep the democracy modern and agile.

“We have all the tools to actually ship fast, but then we also need to be making decisions fast and learning fast, and that's where self-service is so important because you can't be waiting for answers to the questions that you need to have,” Bauer said.

By running lean and dogfooding their own product based on their data-backed performance, Bauer and the Amplitude team are able to iterate quickly and make changes to both how they operate and how their product works.

“It’s really important to have the right tools to support the processes and the culture that you’re trying to build,” Bauer said.

But Bauer warns against just adding tools for software’s sake. The tech stacks of self-service data democracies should be carefully assembled and in service to the people and processes that uphold it.

To ensure that the tools you use—whether your own, Amplitude, or others—fit your needs, we recommend working through the following stages before picking new software:

  • Fully build out your strategy and tracking plans so you’re working with well-defined goals and needs.
  • Look at where teams need the most help, where processes can be optimized, and where data governance would benefit from software assistance to meet your people where they’re at.
  • Pick the best tools to complement the exact work of your teams and support your self-service data democracy.

At the end of the day, it’s not about having a really complex stack of the shiniest tools around; it’s about mindfully choosing the tools that will best fit the unique needs of your team and, by proxy, your product and customers.

Avo supports your self-service data democracy.

Companies like Amplitude know that the key to a modern, self-service data democracy is having a single source of data truth that you can trust. Avo makes this—and more—possible, no matter what you’re building.

Avo is a product analytics tool built for the self-service data democracy. It lets you delegate and collaborate easily by helping you create one single source of data truth that you can seamlessly share with your team.

With Avo, you can work with your team collaboratively to ship analytics tracking and get insights faster than ever before, all without compromising your data quality. Get started today.