Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
From lunch to acquisition: How Atlassian bought Trello (techcrunch.com)
134 points by andytolt on Feb 21, 2017 | hide | past | favorite | 74 comments


"When we sat down, it was all about the vision and what we wanted to build,"

Blah blah blah. It's all about them benjamins. Who likes atlassian software? Shittiest software I've ever used. They have a shitty version of everything you could want.

Bitbucket, JIRA, HipChat. All suck. RIP Trello.


This is an incredibly short-sighted comment.

Two CEOs sit down, culture had better be the #1 thing they discuss before one writes a check to the other for the better part of half a billion dollars.

Good M&A adds tons of value by combining two companies who are better together than separate. Bad M&A (Google's acquisition of Nest comes to mind) means fired CEO, delayed/ruined product launches, and incineration of nine-figure amounts of shareholder value. It all comes down to whether the two companies want the same thing.

I won't argue that Atlassian software has no rough edges UX-wise, but that isn't the point. The point is that, when one company buys another, they'd better have roughly the same idea of where they want to go, otherwise you may as well fire up an industrial blast furnace fed by shareholder capital.


This is all 100% true yet still quite naive.

Bad M&As happen all the time. Good ones are the exceptions not the norms.

"Companies" are groups of people, and their leadership.

What the leadership wants very often has nothing to do with the company's long-term future.

This is all observable behavior happening all around us.


>Good ones are the exceptions not the norms.

Is that really true? Here are lists of acquisitions by Amazon and Google:

+ https://en.wikipedia.org/wiki/List_of_mergers_and_acquisitio...

+ https://en.wikipedia.org/wiki/List_of_mergers_and_acquisitio...

The vast majority of acquisitions are uneventful and added positive value to the acquirer. Maybe you're thinking of high-profile multi-billion dollar failures like Daimler+Chrysler, HP+Compaq, Microsoft+Nokia. It's possible that the amount of negative press those failures receive distorts the perception of the (quiet) successful acquisitions.


The vast majority of acquisitions are uneventful and added positive value to the acquirer.

How many add positive value to the acquired?


Most if not all of them because they were all friendly deals and the acquired wanted the deal to happen. (In other words, the acquired's owners wanted BigCo's money more than they wanted to remain independent.)

Most of the targeted businesses were privately owned which means that the big acquirer cannot execute a hostile takeover and buy the company against the Board of Director's wishes. Amazon's acquisitions of Zappos shoes, dpreview camera website, etc were all friendly deals. Same situation with Facebook acquisitions of Whatsapp and Instagram.


Jira doesn't suck. It is extremely complex, especially if there is a (misguided) desire for agile plugins and such, by Jira on its own is quite usable. You must have not used their competitors to consider it crap.

Jira's default state in many companies is more likely proof of broken processes. Something which no tool can solve.


The corollary of Jira being "extremely complex" turns out to be that most Jira instances are not set up optimally - meaning that the experience for a lot of people is "this is crap".

I'm not sure the blame for that can be placed on Atlassian thouh - the problem space they're trying to solve is large, varied, and complex (perhaps _that's_ Atlassian's fault, trying to do too much?) I've yet to come across any alternative to Jira that solves a large subset (or superset) of the problems Jira solves, that isn't equally difficult and resource intensive to set up and customise for a specific company/team. Trello's a great example of a _much_ simpler (and on the surface much less "crap") piece of software, but it falls _way_ short of doing everything a typical Jira shop needs Jira to do... A 5-10 person bike shop building custom bikes? Yeah, Trello will probably crush it. A 10-20 (or50-100) strong dev team working on multiple complex projects? Nah - Trello won't solve all your problems.

Jira can _easily_ suck (and does, way too often). But it's almost certainly a config/training problem rather than a software problem.


Many companies have JIRA set up like a clock which is too tightly wound - not enough slack in their processes and permissions and too wedded to a development or organizational ideology.

JIRA is amazing if the overriding philosophy is openness and slack in workflows.


But it's almost certainly a config/training problem rather than a software problem.

We can probably boil this analysis down a little bit, right? What's Jira's default UX for new installations/projects, does it _easily suck_ by default?


Change a few words and that becomes a defense of SAP.

The fact is, no software can make everyone happy all the time and the software will suck if it tries to do it. Yet we keep falling into the trap of thinking crap like JIRA and SAP can be customized just right.


Sure, and I'd say the same (with a little less personal experience) for SAP, Salesforce, Microsoft Dynamics and a whole bunch of other "enterprise platform" scale tools. There'll be heaps of poorly thought out and configured installations, and a few much less talked about companies where they got it set up properly and trained the users - who're busy raking in money executing their company business instead of killing productivity and morale fighting with their software...


> and a few much less talked about companies where they got it set up properly and trained the users - who're busy raking in money executing their company business instead of killing productivity and morale fighting with their software

I've never seen evidence that these companies exist apart from a few CTO's blabbing on about how great their latest implementation is, even though they never have to deal with the problems.

The only ones it ever works for are those with very generic problems.


I _suspect_ this is because you don't move in the same circles as the people who benefit most here.

Spend some time talking to the sales people who's haircuts look like they cost more than you make in a week, or who's suits probably cost more than your car. They're _all_ got finely tuned Salesforce (or equivalent) setups and can probably at any time tell you off the top of their head the number of leads and prospects in their pipeline, the size of their expected deals, their recent conversion rate, the estimated total value of their prospects/leads (and the precise dollar value of their commission and bonus...)

I've seen the same with large logistics operations - the ops/management people know exactly where all their assets are and the schedule of movements compared to the upcoming work requests.

You won't hear these stories at a Ruby On Rails meetup or in a weekend hackathon. It's not part of most tech/IT stories. But it's 100% happening out there.


If you like lightweight, fast apps then you hate Jira. If you like being able to tweak absolutely every single feature, then you love Jira.

It's nice to imagine an app with all the flexibility of Jira and all the speed of Trello, but it's probably a trade-off.


> tweak absolutely every single feature

Unfortunately, you can't. For example, let's say I want to have a different set of priorities or resolutions for a project. I have to run a whole new JIRA server.

Twelve year-old issue with 2,319 votes https://jira.atlassian.com/browse/JRA-3821


The problem with software X is that it's far too complex and it doesn't have my favourite feature Y!


Or thousands of other people's favorite feature.


Favourite?


Ask your parents for a trip to Europe as a graduation present.


Resolutions by issue type? Talk about bikeshedding to the extreme

Atlassian is right in not addressing this kind of feature request


Complex != Good

Have to admit, I haven't used any of their competitor's software, but just because you can change every single setting, doesn't mean people will want to. I'm pretty sure no one in my company fully understands it. Settings are all over the place, and even small changes require many clicks to get to where you need.

I know it's a different product, but if you look at Zendesk, it's so simple to get setup, and there are still many integrations to achieve what you need but all at the click of a button, in a friendly interface. It just works™...


Sure. vi, grep, and bash are all great tools, and you can do amazing software dev work using them - but that doesn't mean things like IntelliJ aren't also great. My Groovy/Grails devs think I'm some sort of Wizard when I pull out grep and pipe things to awk then pipe that to a Perl one liner, but I watch them using their IDE and think "Wow, that's _so_ cool!" as well...

I'm old-school, deeply entrenched into the Unix "do one small thing well" philosophy. But there's no string of terminal command line tools that's going to do all the things Jira does for the entire group of people in a typical business that use Jira.

(I find it quite amusing when devs here discover/get-shown for the first time that SourceTree is - underneath - pretty just typing git commands into a terminal for them!)


Occams razor is slightly more sophisticated than 'complex != Good'.

Distinguish between inherent and accidental complexity. You'll see that every human endeavor has many known and unknown inherent complexities. Just ignoring those will hamper your ability to abstract and ascend.


Complex Jira setups and bad management seem to go hand in hand. A bad manager will try and fix problems by introducing more process and avoiding the real issues. It lets them feel like they are doing something useful. And Jira is more than happy to let them do that. That's not a fault of the software, though.


Well, Jira has an 80/20 problem: it tries to do the 20% that satisfies 80% of the users (cases.) so far so good, but when you trip over one of the details that make up the remaining 80% you're SOL. JIRA-1330 is my favorite :)


Atlassian products are about one step improved from corporate financial management software. Think about the last craptastic time keeping system you've used. They all seem to be terrible in some weird way.

For example, the system at my company has unexplained checkboxes everywhere, and for some reason disables copying of charge codes once their entered onto a sheet? Why? Nobody knows, it's probably not a feature, just some nonsense produced by some compatibility requirement for an old browser or something. Who knows. Other fun things: - search doesn't work right to find codes - it won't verify if you can charge to a code until you've tried it - it won't tell you codes have expired when contracts are over - it's a huge process to review old time entries, especially old sheets filed by employees - etc.

compared to this, the Atlassian products are at least usable when they're configured correctly. The problem is that they're almost never entirely configured correctly, or they'll somehow mysteriously fall out of correct configuration after months of working just fine. Oh, and this time keeping system is actually a step up from the old one we used to use.

I guess what I'm saying is that they aren't state of the art w/r to usability, but they're also not complete bottom of the pile garbage.

Trello is also a step or two above the normal Atlassian stuff, but the inability to self host and configure in certain ways also restricted the size of projects you could tackle effectively with Trello. Atlassian's normal project management stuff seems to work best at a scale just beyond Trello. And in that respect is superior.


Definitely see this with Vanguard - great index fund, terrible software as a user.


> Who likes atlassian software?

I do, and have been using it since Atlassian was a tiny startup.

Don't know any project management software as flexible and powerful, with integration with enterprise source control and bug tracking systems as Jira.

I really don't have any reason to suggest anything else.


"Shitty" is too far. I'd call it "adequate".

Atlassian's genius is having lots of adequate enterprise collaboration software under the same brand.


What makes you think Jira is worse than say BMC Remedy or IBM Rational ? - JIRA is certainly better than most of their direct competitors.


Oh the nightmares of using ClearQuest back in early 2000's.


Remedy is one of the worst software packages I've had the misfortune to work with, so saying something is better than Remedy is damning it with faint praise.....


Ha ha... but under the hood there's some utopia lingering.

I've seen its exceptions and I figured it's made of some sort of auto-generated code translation of an Eclipse based think client.

Reminds me of the bad old days of round-trip engineering, UML and Corba...


Atlassian found their niche: enterprise sales. Because the buyer is not the user usually so it's way easier to sell shitty products.


From my experience Hacker News community dislike post like this but I have to agree, RIP Trello. Atlassian bought 17 startups and for better or worse they're not thriving.


So BitBucket, HipChat, and so on, you say they're not thriving? Bb didn't even have git repos when they sold!


I had to manage our corporate JIRA for a few years after the previous admins managed to get it stuck in an impossible configuration. My experience was it worked good enough if you stuck to the defaults. Spent a few weeks ripping out all kinds of weird workflows and started trusting the users instead. The thing was not perfect, but compared to the other stuff we had i'd give it 8/10 (or a B- for you Americans out there)

Then a bunch of contractors came in with a paper full of boxes and arrows, titled Enterprise Workflow, and demanded admin access to do everything themselves. One month later stability is down the drain for everyone, and a few of their issues are jailed in a state, unable to go back and forth in the workflow, never to be closed or disappear from their status report. They declared Jira a non-agile, buggy product and bought something different. Our users now hate both products and track issues in excel.


Crappiest software I used: IBM

Atlassian might be not great but it's not sold on golf courses and actually works


Personally, I think Trello is pretty crappy. Unlike you, I'll give some examples:

- Not usable without browser plugins

- Referencing cards from other systems impossible (no stable urls)

- Referencing other systems from cards difficult. For example, try referencing a commit in a repo

- Not all actions on cards are logged

- Referencing cards from within cards in a stable way is impossible (hello web!)

- Limited keyboard control

- Changing certain things concurrently does not warn user

- Only two part hierarchy

- The complexity of over engineering processes shows up in a year in both JIRA and Trello. But in Trello you can't blame the program.

- Drag and drop frustrations in Firefox

- Side scrolling nightmare when pasting long logs because of narrow design of cards

- Endless gimmicks

- Advertisements in cards for plugins (at least I think it was)

I can go on, but I think I've made my point. Could you give specific reasons why you don't like JIRA?


- What are you missing that browser plugins add?

- Cards do have stable URLs. Not sure why you think they don't.

- A Github url to a commit will link to that URL. Powerups can be used if you want tighter integration.

- Very few people actually need this. If you do, then you need a fully auditable system.

- Copy paste the card URL. Again, card URLs are stable and Trello will detect that it's a card URL.

- There are fairly extensive keyboard shortcuts: https://trello.com/shortcuts

- Such as?

- Boards -> Lists -> Cards -> Description/Comments/Check Lists/Polls/etc. At least 4 levels.

- Is this a prediction??

- Sounds like a Firefox problem.

- Are you putting logs in the titles of the cards?

- Such as?

- Haven't ever seen an ad in a card.

It sounds like you set out with the intention of not liking it and found (or made up?) reasons.


- names of tags on cards, card id's for referencing from git commits, smarter use of screen estate. Everyone I work with uses at least a couple of plugins.

- Not when you move them from board to board

- But that's not linking a commit, that's linking to github. In JIRA I had full integration and even auto linking of commits to issues. You give a repo url and the rest is easy.

- I recall (can't test this) that moving a card is not logged. I think that's pretty basic functionality and it helps in asynchronous communication.

- card urls are not stable over different boards.

- moving cards over columns.

- Yes, but I cannot nest cards. Simple divide and conquer strategy will scatter takes over many cards. To-do lists are not sufficient.

- seems like a nasty ui decision, dragging with scrolling.

- No, in the description. But the pre has no word wrap and the pop-up only takes a quarter of the width of my screen (without browser plugin)

- large smilies on cards for example, or whatever these things are.

- I'm absolutely sure I saw advertisement in a card. Maybe it was caused by a browser plugin?

I've used many task management systems. I understand the appeal of Trello. For me, it is just not sophisticated enough and misses essentials.


> Not when you move them from board to board

That's not true. Did you try it?

> I recall (can't test this) that moving a card is not logged.

Moves between boards and lists are logged. Moves up and down a list are logged, but not shown in the UI (only visible via the api). All moves were shown at one point, but it was just way too noisy.

> card urls are not stable over different boards.

That's still not true.

> moving cards over columns.

The card moves physically each time and each move appears in the log.

> I'm absolutely sure I saw advertisement in a card.

Trello suggests turning on a power-up if it looks like it would help (like github power-up if you attach a github link). "Advertisement" seems like a strong word here?


Well, things change, also at Trello. This does not seem to be a thing anymore, but was when we were setting up our workflow (https://trello.com/c/sSldoiVf/46-deprecated-mentioning-cards...).

Actually, it seems changes in descriptions are not logged. Even though I normally would not want to see this, it sometimes comes in handy. Just a simple line: "Edited: John Doe, 2017-02-22".

Well, suggesting to turn on power-ups on each and every card, styled in a way as if it is part of the card.

W.r.t. moving cards over column: if I move a card from A to B and the move is not synchronised (because of lag or a Trello Server outage), a co-worker might move the card as well (say, from to-do to in-progress). Once the situation stabilises, there is (or was, the last time I noticed this issue) no visual feedback of this situation. It's a minor annoyance, but I've noticed that the 'eventually consistent' UI makes it possible that inconsistent states are reached.


The complexity of over engineering processes shows up in a year in both JIRA and Trello. But in Trello you can't blame the program.

Not a Trello user, but sounds like a huge positive to me. If you can't blame the software, there might be some chance of pushing back against the process-advocates.


I have quite the opposite opinion. I really like working with Jira (as opposed to e.g. bugzilla, clearquest).

And I really prefer Jira agile plugin to Trello, some PO tried to force Trello on us (developers) but we had to resist for some time, until they stopped pushing it. Trello looks like a sticky notes app and nothing more to me. Not to mention it is not "on-premises" so if a given company cares about security it won't use it.

Bitbucket Server (formerly Stash) is also quite nice, but I don't have comparison to GitLab or Github Enterprise, so I don't know what I'm missing.


It's an Australian company publicly traded, so they must be doing something right.

The only thing they suffer from is NIH syndrome.

I am a happy user of Bitbucket and a sad user of GitHub, so there is that.


Can you elaborate on why you are a sad user of GitHub?


Because I have to pay for private repos and teams however small they are (bitbucket private repos are free).

Also because I generally tend to prefer the silver (underdog) candidates like Pepsi, the Falcons or Hillary.


Github is a pretty decent product. It seems churlish to criticise it on the basis that it comes at a price.


Same applies for Bitbucket.


Didn't GitHub update their pricing a couple months ago to make private repos free too?


From a 10-second look at their website, the answer is no.


How is Bitbucket bad?


Bitbucket, unlike github, clung to a design of many servers connected to a big disk. GitHub instead went for many servers with their own disk and a distributed storage system. This means that bitbucket ends up being a lot slower in many operations than github.

Aside from that, it's fine.


What are those operations that are slow in bitbucket?


I do. I love it actually. Separately they make good software but using multiple atlassian products together makes it amazing software.

Jira, Bitbucket, Confluence, Bamboo all amazing tools


My 7th year using it and I have nothing but good things to say about bitbucket (switched after using github for two years)


> Who likes atlassian software?

I do, just because you dont like it does not mean others will have the same opinion.

If you think you can do a better job you should start your own company.


Did I miss something or is this a content free article?

Mike fly's from Australia to America. Has chat over meal saying they want to work together. Fly's back to Australia. Agree on a price. Deal done.

What am I supposed to learn from this?


You are supposed to learn that content free articles can be good PR. They can make it to hn front page even.


Most corporate puff pieces are content free.


> “When we sat down, it was all about the vision and what we wanted to build,” Pryor said. “The kind of people we hired, the kind of company we wanted to create.”

Soon as I read that line I knew this article wasn't going to be of any worth. Just a PR puff piece.


I've always wondered what the Fog Creek people think about this. After all, their little baby got scooped up by their biggest competitor.

Is that good or bad?


I just hope they keep the free Trello up and running forever, and it stays simple on the surface at least.

I have this nightmare where one day I open Trello and it looks like Jira.


You are not alone in that thought.


I guess "it's good". With the traction they were having, I think they could have raised more money and stay independent. So I don't think it's "it got scooped up", but more Trello decided to sell to someone they liked.


I suspect the owners of Fog Creek are happy with the financial windfall.


I'm proud of MHP and Trellists.

Seeing how successful Trello has become makes me sad for the lost potential of our former promising stars, like Copilot and Kiln.


Of course it's good or they wouldn't have sold.


Remember Trello is separate from Fog Creek. Until the acquisition, they were complementary products. Now they are still complementary but being offered by a direct competitor.

So the question is how do the people who were left behind at Fog Creek feel about something that was spun out of them now going to make their biggest competitor stronger.


If anyone else thinks Atlassian are going to run Trello into the ground, there is an open source alternative that's sprung up called Wekan. https://wekan.io/

I haven't used it personally, but it appears to look and work just like Trello!


A much better account of an attempted Atlassian acquisition is told by Peldi from Balsamiq. Fascinating story with lots of juicy details about why it didn't happen. AFAIK Peldi tells the story at conferences but asks for it not to be recorded, to respect Atlassian's wishes.


Please tell me the restaurant was mad dog.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: