Atlassian is 20 years old and unprofitable • Hacker News Board

Re: Atlassian is 20 years old and unprofitable

#341

Post by stu2b50 » Thu, Jun 30, 2022, 2:27 PM UTC

zeroonetwothree wrote: post #314

carlsborg wrote: post #303

I suspect the following is built into valuations of many software companies with subscription revenue models: they can reduce headcount dramatically and keep a relatively tiny headcount of staff around to sustain the existing recurring revenue and be highly profitable even in a slowing economy. For example how many people would MongoDB, with close to a billion dollars of annual revenue, need to keep the lights on, as…

I think this is not as straightforward as you make it sound. 1. Often the way companies are structured requires specific teams and processes to exist and unwinding that is not easy. 2. It can be very hard to identify which people are actually critical. If you get it wrong it’s a big problem. 3. Additionally, firing a large percentage of employees will create a very negative feeling in the remainder so they might leav…

AirBNB? They were flailing at the beginning of covid, taking a massive 2 billion loan at 10% APR to maintain cash flow. They laid off a substantial portion of their workforce, hunkered down, and it paid off - they went on to successfully IPO a year later, still in the pandemic.

Re: Atlassian is 20 years old and unprofitable

#342

Post by ratww » Thu, Jun 30, 2022, 2:36 PM UTC

zeroonetwothree wrote: post #314

carlsborg wrote: post #303

I suspect the following is built into valuations of many software companies with subscription revenue models: they can reduce headcount dramatically and keep a relatively tiny headcount of staff around to sustain the existing recurring revenue and be highly profitable even in a slowing economy. For example how many people would MongoDB, with close to a billion dollars of annual revenue, need to keep the lights on, as…

I think this is not as straightforward as you make it sound. 1. Often the way companies are structured requires specific teams and processes to exist and unwinding that is not easy. 2. It can be very hard to identify which people are actually critical. If you get it wrong it’s a big problem. 3. Additionally, firing a large percentage of employees will create a very negative feeling in the remainder so they might leav…

I've worked (as in actively worked) in a couple restructuring efforts, and it went fine, but it wasn't on tech, it was on education. It went massively well. The universities were financially healthier afterwards, and all the remaining employees had safer (union!) jobs.

The trick for this being successful is most of the time by challenging the processes themselves, rather than the employees. Employees aren't useless, but process are. The reason you need 500 people rather than 250 or 100 or 50 to do something might be simply because you said so somewhere or sometime.

So, companies should review and change processes BEFORE reducing headcount. But alas, companies don't ever do it, because it requires introspection, egolessness and actual work, the sorta stuff that is rare in upper management.

Re: Atlassian is 20 years old and unprofitable

#343

Post by fencepost » Thu, Jun 30, 2022, 2:37 PM UTC

papito wrote: post #311

ryanackley wrote: post #107

Worked there from 2008-2013. So early-ish employee here. I still have a chunk of shares from the original ESOP program. Saying that for full disclosure because I'm biased and this feels like a hit piece to me. When I worked there, they disclosed all financials in our internal Confluence site. There was a policy of full transparency with the employees. The owners (Mike and Scott) had impressive business discipline. Th…

It's like that scene from Silicon Valley (which is almost always accurate). Seriously, when did it become normal in capitalism that you should not prove that your business can make money? That there a customers willing to pay for your product? That the promise of future maybe profit sells better than actual profit now ? This is insane.

If I have a company with good cash flow, all employees (including me) are compensated well, and I'm spending a chunk of that cash flow on R&D and growth but have only marginal profits if any am I doing something wrong? Should I cut R&D to pay dividends? How about if the company is privately held?

There's a difference between 'not profitable' and 'losing money.'

Re: Atlassian is 20 years old and unprofitable

#344

Post by acdha » Thu, Jun 30, 2022, 2:38 PM UTC

Aeolun wrote: post #312

acdha wrote: post #278

Earlier quoted context omitted.

GitLab issues are better for all but very large projects. Jira will win on features, of course, but unless you have 50+ people working on the same thing, multiple dedicated PMs, etc. you’re almost certainly not seeing enough value to be worth the frictional cost and temptation to divert time away from the product to build Potemkin project plans in Jira. (This is the same failure mode as many Agile-as-practiced projec…

Oddly enough, if your team is small enough that it doesn’t need a dedicated PM, you generally have no issues with Jira either. It’s almost as if the PM is the problem instead of the solution.

You have fewer problems with Jira. You’re still paying more for a slower application with a clumsy UI.

I also wouldn’t be so quick to condemn the PM. I’ve seen plenty of developers insist on complex workflows and custom fields because they think they can automate everything, only later to realize they’ve created a second job. The problem is basically the lack of a functioning feedback mechanism to ask “do we need this enough to pay for its care and feeding?”, and I’ve seen people from every role on either side of that, although it’s certainly more common to see business people underestimate the complexity of what they’re asking for.

Re: Atlassian is 20 years old and unprofitable

#345

Post by upupandup » Thu, Jun 30, 2022, 2:39 PM UTC

ryanackley wrote: post #107

Worked there from 2008-2013. So early-ish employee here. I still have a chunk of shares from the original ESOP program. Saying that for full disclosure because I'm biased and this feels like a hit piece to me. When I worked there, they disclosed all financials in our internal Confluence site. There was a policy of full transparency with the employees. The owners (Mike and Scott) had impressive business discipline. Th…

> it's actually bad to be profitable when you're growing.

I've seen this happen so many times:

- profitable, down to earth founders generating net profits

- VC approaches said outfit and tries to gather as much info

- VC realizes said outfit is difficult to emulate and break into market

- VC invests and starts demanding they run at a loss to grow quickly

This works well when interest rates are low, and VC is not under pressure to deliver returns. However, when capital suddenly becomes slightly expensive, the whole house of cards start to crumble taking down the good business with it.

Neither strong prudence or optimism helps here. Once you start running at a loss post-VC money, you no longer control the destiny of your own company you started.

Let this cyclical downturn (likely to last for 5 or more years) be a lesson that the previous generation learned. If you are not making growing net profit (Revenue - COGS), you are no making period. Not so while ago people were arguing that debt is an asset and cash is a liability. It's always amusing to me how quickly people denounce gravity are impacted by it.

The good times are over for VC backed SaaS

Re: Atlassian is 20 years old and unprofitable

#346

Post by papito » Thu, Jun 30, 2022, 2:41 PM UTC

fencepost wrote: post #343

papito wrote: post #311

Earlier quoted context omitted.

It's like that scene from Silicon Valley (which is almost always accurate). Seriously, when did it become normal in capitalism that you should not prove that your business can make money? That there a customers willing to pay for your product? That the promise of future maybe profit sells better than actual profit now ? This is insane.

If I have a company with good cash flow, all employees (including me) are compensated well, and I'm spending a chunk of that cash flow on R&D and growth but have only marginal profits if any am I doing something wrong? Should I cut R&D to pay dividends? How about if the company is privately held? There's a difference between 'not profitable' and 'losing money.'

I worked for a company that rented a new expensive floor in the heart of NYC to show "growth" to other investors. We were supposed to spread out for the optics, when potential investors did a walk-through.

Re: Atlassian is 20 years old and unprofitable

#347

Post by tablespoon » Thu, Jun 30, 2022, 2:42 PM UTC

ryanackley wrote: post #107

Worked there from 2008-2013. So early-ish employee here. I still have a chunk of shares from the original ESOP program. Saying that for full disclosure because I'm biased and this feels like a hit piece to me. When I worked there, they disclosed all financials in our internal Confluence site. There was a policy of full transparency with the employees. The owners (Mike and Scott) had impressive business discipline. Th…

> Doug Berman, the founder of Great Plains software

I believe that's Doug Burgum (now governor of North Dakota).

Re: Atlassian is 20 years old and unprofitable

#348

Post by ProAm » Thu, Jun 30, 2022, 2:47 PM UTC

zeroonetwothree wrote: post #323

sien wrote: post #28

So with Slack there is Mattermost and Zulip. What are the equivalents for Jira ? Is there an openish source piece of similar software ? Jira cops a lot of hate on HN, but really to paraphrase Churchill on Democracy - "Jira is the worst issue tracking software, except for all the others". The comparison on Wikipedia has so many of them : https://en.wikipedia.org/wiki/Comparison_of_issue-tracking_s... Are any of them a…

Linear and Asana are far better. I refuse to work on any project using Jira. It’s the first question I ask in interviews.

Asana is hot garbage. As much as I dislike Jira, Asana could be argued as worse.

Re: Atlassian is 20 years old and unprofitable

#349

Post by jokethrowaway » Thu, Jun 30, 2022, 2:53 PM UTC

yobbo wrote: post #117

ramraj07 wrote: post #72

Earlier quoted context omitted.

The actual lines of code in an OS has balooned exponentially. Of course we can’t discount the abilities of the average software engineer in the 90s vs now. I’d guess back then you had to be fairly driven and smart to be employed there, which is not necessarily true anymore. If you have a pulse and can barely pass fizz buzz you’re employed with 200k salary, so yeah.

> If you have a pulse and can barely pass fizz buzz you’re employed with 200k salary If this was true; it is not an "efficient market" we are observing, neither in SaaS or developer employment. These companies would be outcompeted. The underlying mechanism that enables this is probably SaaS/etc lock-in. Will be interesting to see what/how overturns whatever paradigm this is.

Every once and then they buy a startup to inject themselves with new blood and maintain their dominance. They also need to have decent leadershipt to make the right decisions (eg. I'm sure Ballmer would have killed M$).

They can afford all those useless employees because software allow you to get lots of profit with little work.

Re: Atlassian is 20 years old and unprofitable

#350

Post by upupandup » Thu, Jun 30, 2022, 2:55 PM UTC

carlsborg wrote: post #303

I suspect the following is built into valuations of many software companies with subscription revenue models: they can reduce headcount dramatically and keep a relatively tiny headcount of staff around to sustain the existing recurring revenue and be highly profitable even in a slowing economy. For example how many people would MongoDB, with close to a billion dollars of annual revenue, need to keep the lights on, as…

I just don't see companies being able to justify subscription revenue models in the coming downturn. Many opt to acquire software for a fixed price if they are financially sound, those who cannot, aren't going to be able to keep paying subscription fees methinks, i could be wrong.