COSS Community 🌱

Cover image for OCS 2020 Breakout: Gavin Johnson, Rudder Stack
Joseph (JJ) Jacks for COSS Community

Posted on

OCS 2020 Breakout: Gavin Johnson, Rudder Stack

Gavin Johnson is a programmer-turned-product marketer with experience in IT, tech consulting, and product marketing. At RudderStack, he leads Product Marketing and is helping build the company's open source brand, focusing heavily on increasing both community engagement and product awareness. Before joining RudderStack, Gavin attended business school at the University of Southern California and earned his bachelor's in Computer Science from Oregon State University. He’s worked for Deloitte Digital as a Senior Consultant and Manager, designing and delivering custom applications for clients, at AT&T as a Lead Product Marketing Manager for their digital video products, and at New Relic as a Principal Product Marketer for everything open source as well as their Digital Customer Experience solution. An Oregon native, Gavin now lives in Los Angeles. He is happily married, has a purple belt in Brazilian Jiu Jitsu, and is a photographer, homebrewer, and former marathoner.

Relevant Links
LinkedIn - Twitter

Gavin Johnson, Product Marketer at RudderStack on the role of product marketing and function inside of technology and software companies.

JJ introduces Gavin and the topic of product marketing — 00:00

Can you give an introduce to yourself and where you’re working today? — 0:34

Gavin describes computer science background and early programming career, before graduating to SysAdmin and work implementing ERP systems.

Went to business school and joined Deloitte, implementing large tech projects with Fortune 500 companies.

Gavin describes growing disinterest in the type of work and the lifestyle of consultant, so started looking for new opportunities. Came across a Product Marketing role with AT&T working with Entertainment group. Joined the company.

Liked the role and liked the team, but didn’t so much like working for AT&T. So joined New Relic and spent two years there, before moving to Rudderstack as Product Marketing lead.

“I’ve been doing product marketing since I left Deloitte, almost 4 years now.”

Tell us about Rudderstack. What’s the story behind the company? — 2:35

Gavin gives high-level summary of customer data platform (CDP). Describes the noteworthy approach from Segment.

Rudderstack’s founder was looking to build a CDP, but needed more flexibility and cheaper option than Segment, so he started building a new approach.

What makes Segment and CDPs cost so much is that they store your data, keeping your data in their data warehouse. So they have to charge high amounts to maintain that and get a margin from it.

Makes you pay more for data, and limits what you can do with it. Can’t integrate as well.

Rudderstack founder decided to create a CDP, built on the data warehouse.

Rudderstack is a CDP that works like Segment and mParticle, sends it to destinations and performs transformations, but stores all data in your data warehouse, so you don’t pay Rudderstack for that, you pay your warehouse for that, making the CDP service much less expensive.

Don’t rely on Rudderstack to secure your data, because they don’t store it or persist it. This is better for most companies - one less attack vector.

And it’s more flexible. You can use more BI analytics and ML tools on it.

“We’re the warehouse-first CDP. We work just like you’d expect a modern CDP to work, but we store your data in your data warehouse.”

How is Rudder building out it’s open-source technology in the open? Is it clear what Rudder’s roadmap is? How does Rudder take advantage of the open-source movement? — 5:26

Rudder’s founder made strategic decision to make the core functionality, the server, as an open-source project (called RudderServer).

Has a roadmap with improving clarity.

Elements of software that aren’t open sourced are limited to elements which aren’t core functionality, such as the control plane and the UI.

Rudderserver is built API first, so you can use the API before using the UI.

The functional piece of software, and all the integrations, are open source and available. That was done for specific reason:

  1. It’s easier to get people to use open-source tools.
  2. Founder has strong belief that you shouldn’t be forced to buy expensive things that shouldn’t be expensive (e.g. piece of software, hosted SaaS that are expensive for no good reason).

"We think people should be able to build a modern CDP on their own, without relying on a company like Segment or us to host it for them, but we also provide a SaaS hosted service with a really nice control plane that makes it really easy to point and click your way into having a big full functioning CDP."

When you articulated SaaS companies that license technology as a service and charge you for the data, how is that value proposition so unappealing or sub-optimal, in light of the value proposition you just described, that customers shouldn’t have to pay for the technology and they should have the freedom to run it where they want. But you’re also offering a managed service. How do you think those types of dynamics? — 8:50

Before jumping into business model, Gavin addresses the technical side of what a CDP should do — in particular, breaking down data silos. Most providers do this well. Only issue is, the CDP is another data silo. “You’re literally building a data silo to break other data silo. It just doesn’t make sense.”

Gavin also details the security problems of having sensitive, customer data floating around in someone else’s warehouse.

From a business model perspective, problem of paying for data storage instead of just paying for event processing, is that it’s incredibly expensive. Gavin shares several examples of hearing the problems with Segment’s costs, and details the problem as the pricing model, which is based on monthly total users. If you have a lot of visitors, but not a lot of monetization, you just can’t afford Segment.

"With Rudder’s approach, we charge based on the number of events processed."

"We have a free tier, easy and simple to sign-up for, and you can send 500,000 events per month. Our next step up is 25 million. We build these huge tiers, and we don’t want people to fall off a cliff. At 25 million events per month, it’s $750 per month. That’s cheap. That’s incredibly expensive, especially compared to something like Segment."

"500,000 events per month is an average of about 10,000 monthly active users. At 25 million, you’re looking at a ton of users for $750 / month. You still have to pay for storage, but it’s significantly less expensive then just going with Segment, and you get the same functionality."

"That’s what’s drawing people into Rudder. People are realizing it offers the same thing as Segment and mParticle, but costs them a tenth. Significantly, significantly less expensive with Rudder’s approach -- in addition to other benefits as well."

Gavin describes the importance of being able to “roll your own” with the open-source project. The place where Rudder can add the most value is by making Rudder easier to use, so the enterprise version is very easy, very click-centered, so you don’t have to do much programming to use it.

"Reduce the toil by taking maintenance of infrastructure off your plate, and make the product easier to use, and charge based on use -- which, for us, is the number of events that come through. Doesn’t matter how many users you have, it’s all based on events processed. More sensible approach for most businesses and how Rudder can drive the most value."

Digging into product marketing, how do you see the PMM role at a software company? It’s a fascinating question in COSS context, as an open-source project is very different from a commercial product, which are two very different domains? How have you seen those differences? — 15:21

Gavin describes experience with B2B and B2C product marketing, in media and tech.

Traditionally, product marketing is the go-to-market side of product management.

A lot of time, product managers don’t take product to market (for several reasons). So product marketing role popped up and became popular in the past few years.

Product marketing, the primary responsibility, is to write your positioning and messaging. It’s storytelling. How do you tell your company’s story, or product’s story, in a way that makes sense to the people who need to hear it? How do you tell the story about how this helps devs do their job?

A lot of storytelling, a lot of messaging.

There’s also go-to-market strategy, execution, content creation, sales enablement and building sales materials to scale sales teams effectively and tell story well, analyst relations, some level of press relations and media relations, and a lot of other small things.

The big thing overall is the storytelling. Everything behind it, falls out of the messaging and what you write and the story you’re trying to tell.

We take the technical piece, and put it into words that people who aren’t as technical can understand, and find the value and differentiation.

This takes lots of different forms. At AT&T, Gavin’s work was focused on research and product direction. At New Relic, role was more traditional, focused on building positioning and messaging, talking to people, blog content, all normal PMM things. Towards end, worked on open-source at New Relic. Needed a way to tell the story and build the reputation.

With open core, it’s not just the story, it’s how we build the reputation.

Help define the channels, help define the personas needed to talk to, and finding the tactics to talking to people the right way.

That’s the difference about OSS and open-core projects. (1) How do we talk about our product? (2) How do we talk about our company as an open-source company as well? (3) How do we go to the right places and find the right places to talk to people about this?

How do you think about persona definition, and clarity, and that influencing how you grown the open-source community?— 21:48

“When we think end-user personas for Rudderstack, we’re specifically targeting devs and data engineers. … People that are instrumenting the application.”

Founder built this project, because as a data engineer, he couldn’t develop a CDP well.

Gavin speaks about similarities in how open-source analytics projects and open-source products have sprung up. Considers Confluent as an example. Apache Kafka spun out of LinkedIn.

Someone at a large company develops a key piece of technology, which became open-source, which becomes its own flash point, and a company is built around that project.

“There’s a problem to be fixed, go and fix that problem.”

“If it’s not a competitive advantage, you should open source it.”

That’s why you’re seeing amazing companies spun out of a technology, which aren’t a competitive advantage. “You don’t always know what’s going to grab hold. What you do know is that you can’t do anything other than make something useful for me, and if other people find it valuable, a business pops up.”

Engineers and data scientists are entering in a similar way, forming organizations in a similar way, and having similar results.

On the topic of ‘everything outside of your competitive advantage should be open source’, how do you articulate competitive advantage as a product marketer, given that open-source is so fundamental to your company? — 29:15

Gavin talks about core and integrations being open-source, as a given, which makes things easier.

Normal functionality for Rudder will be open source.

When thinking about competitive advantage for technology, you have to weigh what the bigger advantages are. Walks through counterfactual of keeping Rudderstack closed source and selling as an enterprise vendor.

Gavin details implication of adoption, and benefits for adoption, through open source. Allows people to come in and use it and try it.

When thinking about decisions now, starts with looking at personas. If it’s a specific enterprise request, perhaps a technology around that isn’t open-source, because it’s a very valuable thing that Rudderstack can do. In case of SSO being closed source, it doesn’t give any value at individual dev level, but adds value at enterprise level.

Trying to keep value-adds for individual dev level open-source, and value-adds for enterprise level closed-source.

Approach is based on where the value lands.

Outside of technology, can open-source in other areas (e.g. research, anything you can put on GitHub). Ruddesrtack is open-sourcing blog and doc site, which is relatively simple and doesn’t do company any good to have that material closed source.

If a user wanted to contribute a good thought piece on the blog, open-sourcing that enables it. Same thing for doc.

Good way to encourage people to come in and contribute, because it’s not every going to be a competitive advantage.

“What can we do to not only encourage people to use it, but come in and actually engage? More things you can open-source, more opportunities for people to come in and engage.”

I think that’s so brilliant. It forces clarity on what your true competitive advantage actually is. — 34:12

Gavin agrees and elaborates, in context of Rudderstack.

“Our role is gathering your data, unifying it, and connecting it to your downstream systems, and letting you be smart about it, so you can do transformations. … That’s where we’re providing value. And how you make your data warehouse better for your customer data. … That’s where we provide value. … We’re the connector, we’re the hub, at the center of that customer data universe, that connects everything and make sure it all works well. That’s what Segment played well, but I think we’re going to do it a little bit better, because we’re open and transparent and allow you to apply more tools to it.”

Share your questions and comments below!

Top comments (0)