Playbook
GTM Playbooks
How GitGuardian Leverages Data and GitHub to Fuel Its Growth
Read more
This post by François Dufour is part of our series on Product-Led Growth and Marketing Playbooks. There, we share insights and advice from leaders who have built successful PLG businesses marketing technical products to technical audiences.
Confluent is a poster child of having leveraged an incredibly successful open-source project - Kafka - to build a business so successful on top of it that it went public in 2021, just 7 years after its founding in 2014 (Kafka was originally developed at LinkedIn, and was subsequently open sourced in early 2011).
Marie Gassée, who previously led the self-serve business at Box, joined Confluent in June 2019 when Confluent’s CMO, GC Lionetti hired her as VP of Growth.
Marie shared with me parts of their Growth playbook which can be applied to many companies selling an enterprise license and/or a cloud service that complements a successful open-source software.
As VP of Growth, Marie was responsible for three core functions:
In terms of goals, pipeline value was the north star metric that Marie was driving with the Sales team. They also tracked PLG metrics such as signups, downloads, and conversion rates throughout the funnel.
This evolved as Confluent's pay-as-you-go motion became more central, which wasn't the case when Marie first arrived. In her second year, PLG efforts became far more strategic for the Confluent Cloud strategy and pay-as-you-go revenue, as well as related adoption metrics, became key metrics.
Marie’s team ran a primarily Developer-led bottom-up motion that started with Kafka.
They considered the adoption of Kafka to be their total addressable market. As Kafka’s reach expands, so does the total addressable market.
“Even though we were focused on the commercialization of the open source, we also knew how important it was to increase our total addressable market by continuing to evangelize Kafka,” said Marie.
Continually fueling the expansion and adoption of Kafka was one of their top growth levers. “We wanted to be the go-to source for any learning or questions regarding Kafka– whether that be via docs, quickstart guides, demos, or webinars,” said Marie.
The DevRel team led by Tim Berglund was focused on growing that Kafka adoption. That created some healthy tensions between Marketing and DevRel (e.g. “don’t capture developer emails pls!”), which ensured they weren't alienating the developer audience.
“Once someone is engaging with Kafka, their company becomes a far better prospect for Confluent, and we learned not to mess with that order of operations.”
Marie and her team didn't just focus on maximizing the inbound funnel, which was built on the back of Confluent's expert Kafka-related content. The commercialization motion included outbound strategies that brought in the SDR and Sales teams to reach out to various personas within target companies and essentially say: “Hi, I see that you’re using Kafka, and here's how I can be helpful with that.”
The actual number of organizations using Kafka had grown to tens or hundreds of thousands and was continuing to grow, but identifying open-source users and their companies was far from obvious.
To get a better sense of who these organizations were, Confluent built a system to find what they called “Kafka signals.” The system would look at a variety of different inputs that might indicate a company was using Kafka. For example, they might inspect job sites to find companies that were looking for engineers with Kafka experience. On LinkedIn, they could see if developers at a given company had Kafka skills listed in their profile.
They would also get a sense of which companies were likely to be using or considering adopting Kafka based on the engagement with Confluent's plethora of Kafka-related educational online content.
“We basically aggregated all these inputs to gain a sense of the organizations who were using Kafka so that we could focus our outbounding or targeted Marketing efforts to accounts that were more likely to convert.”
Based on the information aggregated with this system, the team had smarter intelligence about who to target via Marketing tactics and Sales outreach.
Confluent built out a paid Marketing program to reach different personas who could be good good entry points into an account. In particular, they used ads on LinkedIn, Facebook, and Google to target technical users, who were likely to be using or interested in Kafka.
“These ads are a fairly unintrusive way of saying “we’re here– here’s a resource you might find useful”. That strategy worked well for this technical audience, but mainly because the content we were promoting was actually valuable and relevant to them.”
They felt strongly that they could be more performant and efficient by managing their paid Marketing program in-house, instead of going through an agency. Working directly with the likes of LinkedIn and Google eliminated a costly middleman. They also invested a lot in data cleanliness and visibility, partnering closely with data engineers and analytics to measure and track everything. This meant that over time, they were able to track all Marketing touchpoints, including any engagement with an ad, and see how that contributed to pipeline generation in the long term. Without this investment in the data infrastructure, the team would not have been able to actually tell whether the paid Marketing spend was having a positive ROI.
Despite outbound SDRs seeming potentially counterintuitive in a developer-first motion, the team made SDR outbounding work. They focused on targeting four categories of personas with dedicated messaging that worked for each. “The product marketing team did a really good job of establishing these four categories – Developers, Architects, DevOps, and Technical Executives – we would then use this framework to identify which contacts within an organization to reach out to and how to ensure we would be adding value for them,” said Marie.
One of the juggling acts that open source companies with a commercial offering face is how much to promote the open source technology vs. their commercial offering. You want to grow your open source reach and usage - especially when it's essentially your TAM - and you typically don’t want to come on too strong with the commercial offering.
The Confluent team, like many other companies, solved this by offering a strong free version of their cloud offering and encouraging users to start there. When users started experimenting with Confluent, they essentially gained access to Kafka with additional functionality, which Marketing and Sales highlighted consistently in their messaging.
“We certainly would always encourage Kafka usage and highlight the open source’s many advantages, but we also were very clear about what would be gained from doing so with all the Confluent bells and whistles.”
At one point, the team attempted to skip the Kafka adoption stage and target companies immediately with Confluent's value proposition. But this strategy of selling Confluent to organizations that were not already using Kafka “clearly failed.”
This failure helped the team really internalize that the Kafka user base was Confluent’s target audience. The team course-corrected and re-invested back in Kafka education, bringing conversion rates right back up and, eventually, contributing to building Confluent into the events streaming juggernaut we know today.
“As great as it sounded to say that we would massively expand our target audience by going beyond the Kafka ecosystem, it just didn’t work in practice. The numbers don’t lie, and I’m glad we were tracking closely enough to identify this and fix it,”
Thanks Marie for sharing your insights once again! To learn more about how other open-source leaders built successful commercial businesses on top of their free software, you can also read playbooks from Mulesoft and Strapi.
.png)
