10 lessons from working at Atlassian


published: (updated: July 27, 2019)
by
16 min well spent

After three successful years of building SEO, doubling organic traffic, and running cross-functional teams, I left Atlassian by the end of 2018.

It was hard to turn my back on the $19b brand that hired so many smart colleagues and built such great products. But I took a lot away from those years. Following are 10 lessons I learned during my time at the makers of Jira, Trello, Bitbucket, Confluence, and Co.

What is Atlassian?

Honestly, I didn’t even know the brand “Atlassian” before I joined. Sure, I’ve been using Jira for over 10 years. But I didn’t know anything about its history, unique business model, Land and Expand, or how its products are marketed.

Atlassian was founded by Scott Farquhar and Mike Cannon-Brookes, who created Jira in their dorm-room back in Sydney. The name “Jira” (pronounced “dsheera”) is derived from “Gojira”, “Godzilla” in Japanese, as a counter-brand to BugZilla. Through numerous acquisitions, Atlassian added products like Bitbucket or Trello to the portfolio. Four years after its IPO, Atlassian is worth almost $20b, has almost 3,000 employees world-wide, and still leads the deep-end of project management applications.

All products come with strong, built-in viral loops: Jira is a collaborative issue tracker, Bitbucket is a code repository management platform, Confluence is a knowledge management CMS, Trello is a lightweight organizational app. Every product could be used by a single person, but it doesn’t make much sense. At the same time, Atlassian’s true power comes to shine when several products are combined.


Lesson 1: A great product is like using paper to ignite a fire

If you ever made a fire you know it’s much easier to start when using a newspaper to ignite it. The same principle applies to a great product. It ignites growth.

Atlassian’s products have a much stronger brand than the company. Jira, Atlassian’s cash cow, was released 17 years ago and has been the dominant project management and bug tracking solution ever since. It represents the idea to be the most professional solution to project management you can get and increase the productivity of a company. This brand-driving notion, coupled with built-in viral loops, forms the epitome on which we built organic Growth.

My view is that branding is the process of attaching an idea to some object, or to a service or organization.

from the foreword by Rob Walker of Debbie Millman’s “Brand Thinking”

Building a brand is the result of a deep understanding of the customer, creating a narrative around your product, your business model, and the competitive landscape. But it’s also a lagging indicator for Product/Market-Fit and user-acquisition.

When you have a strong brand to work with you reap many benefits:

  • Backlinks = easier to rank in organic search and more referral traffic
  • Brand recognition = more people click on your snippets in organic search, open your newsletters, and read your blog
  • Trust = users are more willing to try your product and more likely to take in-product friction
  • Network effects: building growth on a strong brand strengthens the brand in return
  • CAC = companies have lower customer acquisition costs in comparison to weaker brands because users are more likely to click on ads from a brand they know (see screenshot below)
  • Hiring = it’s easier to hire people for a company with a strong brand
Source: https://www.marketingsherpa.com/article/chart/how-likely-consumers-personalized-ads

Lesson 2: All hail the bottom-up SaaS model!

The bottom-up SaaS model, also called “Land and Expand”, is a strategy that aims at the product initially being adopted by a single employee of a company (land) and then spreading through other departments over time (expand). It’s the opposite of a top-down sales model, in which reps sell to C-level managers of a company. That’s why it’s also called bottom-up SaaS model.

The business model takes a couple of ingredients to work. First, the product must be inherently collaborative, such as Slack, Zoom, Jira, or Dropbox. It’s inherent that one or more core features of the product are based on interaction with others, whether through chat/commenting, file-sharing, or project management. Second, the product is only adopted when paired with a freemium model, or at least “free trial”. Otherwise, friction is too high for easy spread throughout the company because users don’t sign up.

Atlassian clearly has these components in place. Not just for Jira, but also Trello, BitBucket, Confluence, StatusPage, OpsGenie, etc.

Atlassian’s product portfolio

Freemium also increases efficiency of organic and paid user-acquisition because a certain awareness is driven by Word of Mouth. When users have heard of a product from their colleagues and then encounter an ad or search result from related to the product, they’re very likely to convert.

Now, Land and Expand don’t only work from single user to whole company adoption, it also works with cross-product adoption. That means users jump on Jira, it spreads through the company, and then they look at other products of the Atlassian stack. Their likelihood to adopt another product of the same company is higher because they complement each other. Those cross-adoption effects are network effects and they’re visible in product-adoption and traffic-quality metrics. That’s what the bottom-up SaaS model enables: more efficient acquisition. Atlassian is able to spend ~30% on R+D because it saves on sales department cost. In fact, there are no outbound salespeople at the company. There are only Technical Account Managers (TAMs) because some companies want to speak to a person.


Lesson 3: Work culture ≠ company culture

“Culture” is a buzzword thrown around way too often without clearly defining what it means. One of the best definitions I heard is from Ben Horowitz: “Culture is the collective behavior of a company.” (Source)

Work culture is the way people work. Company culture is the way people act at work. Why are founders so obsessed with culture? Because it shapes how people act. As soon as a company grows beyond 10 people, you can’t follow everyone around anymore to tell them how things should be handled. That’s were culture guides people. It’s a crucial factor, especially in knowledge work companies.

There are big differences in culture between manual labor and knowledge work:

  • The sheer number of hours in the office doesn’t reflect productivity.
  • It’s much more about finding the right questions because answers are easily available.
  • Interaction and collaboration are crucial to be successful.

Atlassian has a well-shaped work culture with lots of autonomy and flexibility. I think that’s crucial for knowledge work. Atlassians live the company values so strongly that you regularly hear them in conversations. People really embraced them, they’re not just phrases. One of our values, for example, is “open company, no bullshit” and it really gives people a door to walk through transparency.


Lesson 4: Agile SEO is more structured than you think

Agile methodology is a very effective model for software engineering that retires the classic waterfall approach. It’s a very pragmatic approach to anything that’s “eaten by software”, also marketing. However, how to implement it is often still unclear. One reason we were able to drive results at Atlassian is the practicality “agile” allows.

Recurring meetings

At Atlassian, we had a structure of two weekly meetings: the “Weekly SEO Meeting” and a “Weekly Business Review”. The former is an hour of talking about some personal stuff, current tasks of each team member, a recap of last week, and what’s done that week. The latter is an hour with the whole performance marketing team to discuss problems, progress, and planning of each team. Both meetings overlap a bit in scope but provide enough visibility to dismantle daily standups, sprint kick-offs, and retros.

I discussed many details on the Growth tl;dr podcast with Scott Tousley and Kieran Flanagan (Hubspot).

OKRs

Part of an agile work-style is goals. At Atlassian, we used the OKR (objectives and key results) framework invented by Peter Drucker and Andy Grove. We basically set annual goals and quarterly milestones we wanted to achieve. Quarterly milestone results and what has to be done to get them needs to be precisely defined in advance. We had big, hairy, audacious goals (BHAG) for the year, which were defined by top management. Then, we set our goals strategically to hit the BHAGs bottom-up.

The beauty of OKRs is that they are visible throughout the whole company. You know pretty much exactly what everyone is working on. Therefore, performance should be constantly measured and discussed at the end of each quarter and the year.

The retrospective aspect of OKRs is just as important as their prospective quality. They help planning because you have an idea of what goals each department is striving for. At the same time, you constantly evolve by discussing how the last quarter went, i.e. what could be done better and what went well.

Integrated Marketing

Integrated marketing is a recurring process that aligns all marketing channels to a campaign, which creates synergies through consistency. Each quarter, we’d have 3-5 integrated marketing campaigns that we’d scope and execute from an SEO perspective. The product managers would approach us with their goal and idea, and we’d talk to other channel managers to see how we could best approach this. The benefits are enough time to plan execution thoroughly, visibility into what others are doing, and concerted efforts.

Integrated campaigns are a very reactive approach to performance marketing. To be successful, you need both, a mix of reactivity and proactivity projects. We tried to run a 50/50 approach at Atlassian: half of the time was dedicated to supporting integrated marketing campaigns, the other half for our OKRs.


Lesson 5: Document and then document some mo’

Documentation is big at Atlassian and I’ve learned to love it. It comes in different forms:

  • Internal blog articles
  • Dashboards
  • Planning/strategy
  • Reports
  • Success stories

Most of these formats live(d) on Confluence, a knowledge management CMS. Atlassian’s hate it, but in truth, it’s an internal wiki. It allows you to like and comment pages from colleagues, see what’s trending, and communicate outside of Email or Slack (or Stride…).

Many companies fail to historically document decisions and information, which undermines the value of someone new coming in and quickly getting an understanding of how things are and why is high. Good documentation, on the other hand, is a great way to communicate and collaborate on anything related to information, creating transparency and alignment. Being able to look back and understand what’s done is also valuable when you experiment a lot and need to keep a track record. There are clear limits to documentation, but most companies undervalue it.


Lesson 6: Growing a product from scratch and then selling it

HipChat was a long-standing incumbent on the messenger market, but then Slack came along. In response, Atlassian decided to give HipChat a huge overhaul, which gave birth to “Stride”. The product was promising but couldn’t keep up. Eventually, in a strategic retreat from the messenger market Stride was sold to Slack in return of a substantial amount of equity.

We, the SEO team, were involved in the Stride project from the get-go. I evaluated different names for “the new HipChat”, helped define the Go-To-Market strategy, and, eventually, its demolition. The challenge was to grow to a certain number of sign-ups within a year. Though we almost hit the target, it wasn’t enough in the end.

Stride was doomed to succeed with a really low budget, which made us bank substantially on SEO, Social, and Email. Unfortunately, these channels take time to scale. We faced all sorts of challenges, such as the wrong decision to put the blog on medium.com instead of our own site (in a subdirectory). ”Stride” was a competitive brand name: there was “Stride Health”, “Stride University”, and another software product called “Stride”. We were able to rank #1 for our own brand but it took us a good bit of work.

We heavily evangelized Stride through content, public speaking gigs, and our existing Email newsletters to leverage the brand equity and customers we already had. Product features like DnD (Do not Disturb) and “actions” (To Dos) gave us an edge over other messengers.

Our SEO strategy revolved around landing pages for generic queries like “team chat” and an app directory. The latter is something lots of good SaaS solutions use to scale SEO, such as Zapier or Slack. For link building, we went all natural: guest blogging, public speaking, PR buzz, partners. The results we saw from these tactics were really good. As you see, there’s not much magic. Just lots of hard work, good execution, and doing the basics really well.


Lesson 7: What it’s like to acquire another company

When we came back to work after the holidays in January 2017, the first Email in our inbox said that Atlassian had acquired Trello for $425m.

The lightweight project management app is a good addition to Atlassian’s product portfolio because of its consumer-friendliness and simplicity. Whereas Jira is a full-blown project management tool that can be used by huge organizations with >10,000 employees, it can also be used by 10. That versatility comes with complexion, which Trello doesn’t have. In fact, it’s so easy to use that it’s more of a “Prosumer” app.

After digesting the news, we started thinking about how we could best integrate Trello into the Atlassian ecosystem. We knew that it would be stupid to block the Trello brand in any way, whether from a domain-strategic nor from product point of view.

Acquired companies are just as tough to integrate as they spur Growth. There’s more to it than just adding a logo to your site. You have to think about the best domain strategy, responsibilities, budgets, and cannibalization. Atlassian has a very passive approach to company integration. It often leaves brands relatively autonomous, integrates analytics, and budget planning, and leaves the team intact.

Trello was exciting because it works so different from a user-acquisition perspective. It has public instances that are indexed by Google and serve as sign-up vehicles, just like Pinterest. Some companies host really interesting stuff on Trello boards, such as public roadmaps, recipes, or templates. The Trello team has a long-standing history of producing fantastic content and thought-leadership that paid off in traffic and links.

What we learned from the acquisition was this:

  • There’s a lot to learn from competing brands
  • Acquisitions do pay off but are high-effort
  • You can benefit a lot from buying complementary brand

There are different degrees of integrations. A 100% integration is not always favorable.


Lesson 8: If you can imagine spam you will find some

Whenever can spam be imagined, it will be found. That’s especially the case when a site has some form of public instance, which can be leveraged for spam. That was the case at Dailymotion, where I worked before, and it’s the case at Atlassian. Dailymotion had lots of illegally uploaded movies and songs. Atlassian has products with public instances like Trello or Bitbucket.

Luckily, I had a great team of developers and product marketers. Together, we decreased spam significantly, though not completely. It’s near to impossible to get fully rid of spam unless you’re willing to make public instances unavailable to Google. But, as you see in the graph below, there’s a lot you can do.

We started with light actions, such as making all outgoing links nofollow, kick/banning IPs of clearly identifiable spam accounts, and much more. An open-source project like Github can leverage its users to identify and report spam, but when you have a business product you need to find scalable anti-spam methods.

Fighting spam always means fighting with spammers and they realize when they’re being fought. Thus, you need to be aware that it’s not a one-off thing, but a continuous fight.


Lesson 9: From startup to scaleup

When I joined Atlassian, many processes were undefined and lots of work was “just done”. It was a startupy feeling with a healthy degree of chaos but high-speed. When I left, we had fully-fledged systems in place that allow the company to scale processes end to end. It felt a bit more corporate but definitely more structured. The challenge in scaling a company is often to keeping work efficient but structured. Failing to minimize process overhead and hire the right people means building a “lame giant”.

When scaling, we added quarterly capacity planning, standardized processes for procurement/integrated marketing/hiring and lots of managers to take care of all of this. An Agile culture helps to minimize the downsides but they’re not completely preventable. It’s important to not simply implant a whole organization from another company, which happens fast when key hires bring their former friends and colleagues. Scaling could be as difficult a growing after Product/Market-Fit.


Lesson 10: SEO and paid performance #aligned

One major project I worked on and that I think way more companies need to tackle is to draw synergies between organic and paid search. What sounds like buzzword bingo is actually an effective way to optimize ROAS (return on ad spend) and compensate for underperforming SEO at the same time.

The thought process is as follows:

  • Queries that are important for the business and that are underperforming in terms of organic rankings are compensated for by bidding more on ranking in paid search.
  • Queries that are well occupied in organic search are less or not at all bid on in paid search.

In reality, that is a very simplified view of the matter. You need to adjust your attribution model accordingly because you’re missing the query referrer, for example. You need to adjust bid and spend at least on a weekly basis and create separate ad groups for that tactic. I’ll certainly write a full blog-article on how to do it, or better said, how we did it.


What a good company can do for your career

My time at Atlassian was incredibly valuable from a skill and self-development standpoint of view. I’m very proud of our accomplishments and the projects I was able to work on.

From a career perspective, Atlassian couldn’t have been better for me. The reason isn’t just that the brand is something like a myth – which I honestly didn’t even know before I joined – but because I was able to do so many career-defining projects. Pulling together cross-functional teams, defining much of the content strategy, and building a team are projects that are so valuable for career development!

There’s only so much I can say about Atlassian but if you have more questions feel free to reach out on Twitter or shoot me an Email.

%d bloggers like this: