Every software system used at Gitlab | Hacker News Reader

Every software system used at Gitlab

about.gitlab.com

283 pointsksj2114118 comments

I'm so sick of Gitlab support spam in every thread that mentions any issue, making it so that critical conversations are interrupted with self serving pitches!

>I once saw Gitlab do this thing.

>Hey, Bob from Gitlab thing here. We are really trying to make thing great for our customers . Our next release of thing will do foo which we hope will fix this. Here is a link to a marketing post comparing us to GitHub. Hope that helps. Please don't mention us in a comment, or I have to do this. It hurts to live.

Hey, Jimmy here, vending machine manager at GitLab. I'm sorry that our posts here are concerning. I'm going to run this one right up the ladder for you. Thank you so much for your concern and sorry about your experience with GitLab!

Their CEO is the worst offender. Shows up on every gitlab thread with these generic/uninteresting comments.

The thing is - I get it. I thi k it's cool when people show up to talk about their product or code that is mentioned on a forum. It's endearing. But when it's systematized by a large corporation, it's no longer so.

It is hilarious to me that this comment is at the top, the rest of the thread is rife with comments from GitLab staff.

I've noticed a similar thing with Stripe

Would you rather have say, Google’s style of replying on issues with deafening silence? I see your point though. I think it’s better this way. A company can be helpful or shoot themselves in the foot while they figure out to communicate. Too much salesy self-serving talk means they don’t take you seriously so why should you return the favor? Now you can find out easily.

I feel the same way when grumpy customers spam posts about interesting topics with their complaints.

Can we all agree to just talk about the interesting thing that was posted, and skip the one-off customer stuff? (I predict: no.)

We've just had a bad experience with GitLab annual renewal, to the point where we didn't know if they'd cancelled our account or not.

The billing accounts system seems to be completely disconnected from user accounts. A notification banner was spammed to users in the GitLab.com user interface, saying that our account was going to be cancelled becuase we didn't have auto-renew. The users who got the banner didn't have permission to act on it in the billing system. The billing system said that we _were_ on track to renew, which disagreed with the banner. Eventually it transpired that the end-of-year "truing up" meant that our account was on hold, but we weren't notified of this.

And when we _did_ pay the bill we get a banner saying "you've been downgraded to the free plan" when the gitlab.com interface says we hadn't.

It really undermined confidence in how well these systems are connected.

On the flip side, the fact that everything is open meant that you can see customer support reps raising internal bugs about this stuff. The principle of open-everything makes it more pallatable, but it doesn't make up for lost confidence.

fault confessed is half redressed

Sorry that this happened. We still have a long way to go to improve our billing system. We have allocated more people to it since a few months and are making much faster progress. But it isn’t we’re it should be at the moment.

Thanks. No worries, these things happen. It's just a little alarming to see messages like "we've downgraded you to the free plan". Even a bit of help text in the banner, linking to a page on situations under which the message might be raised / not be 100% accurate, might ease the experience.

For sure. Please also see the response of our product manager in https://news.ycombinator.com/item?id=23167232

I am a product manager at GitLab and my focus is on making renewals as easy and pain-free as possible for customers. I am very sorry that you had such a confusing and frustrating experience here. We are working on a number of issues that will improve this process, including:

* Making Gitlab.com aware of who the payment owner is so we can more accurately target the renewal banner - https://gitlab.com/gitlab-org/growth/product/-/issues/1551

* Updating the renewal banner copy for group owners that aren't payment owners so they don't attempt to renew if they do not have the right permissions - https://gitlab.com/gitlab-org/growth/product/-/issues/1553

* Moving the renewal behavior from the customers portal into Gitlab.com so payment owners can more easily manage their subscription and renewal - https://gitlab.com/gitlab-org/growth/product/-/issues/1528

* Adjusting how we do "true-ups" so they are done quarterly and pro-rated to the end of the subscription instead of in arrears - https://gitlab.com/gitlab-org/customers-gitlab-com/-/issues/...

I would love to chat with you to learn more about the challenges you experienced and to review some of our plans to make things better if you're up for it. You can reach me directly at mkarampalas[at]gitlab.com

Thanks, it's great to see that. Incremental true-ups would be very useful. My colleague is talking to an account manager directly, but I'll pass your message on.

I'm not surprised. They currently refuse to have a sane policy around licensing people who just want to put in issues and view repos vs actual developers. They make all users count as a full developer, which makes no sense. Their reasoning is basically "its just easier for us" which is not the point. [0][1]

0. https://gitlab.com/gitlab-org/gitlab/-/issues/3320#note_3351... 1. https://about.gitlab.com/handbook/ceo/pricing/#reporter-user...

That's an entirely separate issue, so this seems like an ad-hominem. And possibly a bit off-topic for a thread about their system integration.

But that said, I do wholeheartedly agree! The top-tier account type allows free reporter-level users. But for every other tier (e.g. silver), you have to pay for them. I think that's big mistake, as the cost-benefit of paying 200 USD per year for each member leads to work-arounds, which compromises the user experience.

Gitlab has been pretty great so far, but the code review tooling is so poor I think I need to look elsewhere. The pull request pages take forever to load, they can't handle pull requests with over 800 lines well at all. It will collapse files with over 100-200 lines of changes for seemingly no reason (hiding the comments within them). You can't ignore whitespace in diffs, and the comments disappear from the diff if they push new commits, you need to look at them in the overview again.

I can relate. We use GitLab at our company.

I still think the merge request UI is great. Its main problems are its slowness and the fact that files with a long diff are collapsed. As the full merge request is quite slow, I will usually open the diff of each commit one by one, then put review comments as a "per commit" basis.

Thanks for the comment.

I've left a few replies in this thread already with more detail of work we have done, is currently in flight and work that is planned to keep improving the overall situation.

As to the "per commit" review, we're working on something to make this navigation easier and easier and shipping it very soon. https://gitlab.com/gitlab-org/gitlab/-/merge_requests/28596

As to the collapsed, unfortunately if we loaded all diff lines expanded, the large changesets would make the browser unusable. We're working to lower the memory footprint and have better ways to deal with these large MRs.

One example I can leave you with that we're trying to figure out a way to make this useful as an additional option, is this: * File by file mode of reviewing MRs, including the concept of "unread diffs": https://gitlab.com/groups/gitlab-org/-/epics/516

We're aware of several of the pain points (we use GitLab to build GitLab ;) ), but if you have further feedback, please leave a comment in the issues and epics I've left here (or create new ones) and feel free to tag me @andr3.

Thanks!

Yeah, this is definitely one of GitLab's weakest points at the moment and seriously impacts day-to-day work here as well. Simple features like ticking off files you've already reviewed are missing and even mid-sized MRs are a pain to review even on a beefy machine.

This feature already exists, I use it all the time.

In GitLab? Maybe we're using an outdated GitLab version here, but I can't find it. And it doesn't seem to be mentioned in the offical Reviewing documentation: https://docs.gitlab.com/ce/user/project/merge_requests/revie...

If you mean the "collapsing" feature, this is really not equivalent, since the state does not persist between multiple review sessions.

I don't have it either. I wish we could "check" things off. I also wish I had more ability to do settings "by group", rather than by project.

Community Advocate at GitLab here, it'd be great if you can add this feedback to this issue: https://gitlab.com/gitlab-org/gitlab/-/issues/24629 Thanks!

Hi! I'm an Engineering Manager at GitLab in Source Code team.

This is on our horizon. We have an Epic to track an overall revamp and adding an option to review file-by-file which will include developing something like what you just described.

Epic: https://gitlab.com/groups/gitlab-org/-/epics/516

Issue for the checking of file seen (including the concept of "unread diffs"): https://gitlab.com/gitlab-org/gitlab/-/issues/24629

If you could please add your thoughts in that issue, we could use that feedback while we develop the feature.

Thanks a bunch for caring and voicing your concerns. It helps us get better.

Thanks a lot for taking the time to answer! I hope my criticism didn't come off too strong, overall I love working with GitLab, the parent just struck a chord with one of my pain points. Glad to see you're working on it!

This sounds as it it would be a great improvement over Bitbucket. :-)

In my experience, if you are reviewing 800+ line PRs you should be in an IDE. I believe the average human being can not review an 800 line diff reliably.

Such large PRs would also probably benefit from live discussions over video (or in the good old days -- in person).

Interesting, I’ve only went through the effort of checking out a pull request for review a couple times. What’s the workflow like? Any tips for reducing friction? I avoid checking out PRs because it’s much higher friction and because I trust CI to catch most bugs so the crappy code forge interfaces are good enough.

Personally, I would kill for a one click to step review interface in IntelliJ that synced comments back to the review.

I use gitlens extension on Visual studio.Pretty great so far.

I've reviewed 4000 line MRs without problems using other tools. It is possible.

Here's a little script I use occasionally, it forces the browser to download all hidden files, and to display them all:

  document.querySelectorAll('.js-file-title').forEach(t => t.click());
  document.querySelectorAll('.diff-content').forEach(d => d.style.display="block");
  document.querySelectorAll('.nothing-here-block').forEach(d => d.style.display="none");

I don’t understand. This describes basically every code review tool I’ve ever used. If there’s something aside from Gitlab that I should try that still integrates, I’d love to know.

For huge diffs, there’s not much that beats `git checkout <branch> && git diff HEAD^`.

For leaving comments, pretty much every system I’ve used has issues. At work we use phabricator; the default settings are atrocious but they can be tuned to something workable.

I sometimes wonder how hard it would be to get diff reviews to be a first (or even second) class citizen of the git ecosystem. That way you can finally get the bliss of leaving review comments in the IDE. They already have “notes” which are vaguely similar, maybe it’s something they’re considering?

> [git ecosystem] have “notes” which are vaguely similar, maybe it’s something they’re considering?

If you're referring to git-notes, they are ... kinda special. They are objects in their own namespace that reference the repository objects the notes themselves are attached to.

Sure enough, you can attach a note to any git-object. And you can view them either directly or from the git-log view. But the discoverability is lacking, and the UI of doing pretty much anything with git-notes is absolutely atrocious. Because they are in their own namespace, you don't automatically get them when you do a pull or fetch, which can get pretty confusing.

For the curious ones, the magic invocation needed to pull in updates to notes from the remote repo is:

    git fetch origin +refs/notes/*:refs/notes/*

Notes are not versioned, and you certainly don't get nice N-way merges between local and remote content. Semantics are different from regular git, and stuff can get lost.

So from where I'm looking at things, git-notes are not suitable for plumbing in a code review flow.

You're right, I was referring to git-notes and they aren't suitable as plumbing for code review :)

I was meaning to point out that we've developed plumbing for objects with "weaker" guarantees than source code, which can still be attached to refs. To me, this is indicative that the maintainers are willing to accept patches on things that are outside the "core mission" of versioning source code, and might mean that they're open to the idea of accepting a well-architected code review feature.

Although, you're also right, in that there's still a lot of plumbing to make sure that authors can only edit their own comments, and that sort of thing. The more I think about it, there's a lot of work to be done and it isn't easy.

But, if you made a proposal, I think git-notes is precedent for them accepting your contributions. :)

I always use IntelliJ for code reviews, works great.

I mean load a pull request from Github (takes <1s) Load one from Gitlab and it takes 10+ sec. Maybe I'm spoiled from using Google's tools previously.

Hi, Engineering Manager at GitLab here, Source Code team.

Thanks for the comment.

Indeed that is something we've had in our minds for a long time. Over the past months/years, we have been shipping out incremental improvements in the way we handle and render large MR diffs.

Recently we made the loading happen in batches which resulted in a significant improvement in how fast we start showing the diffs. https://gitlab.com/groups/gitlab-org/-/epics/1816

We plan to continue improving the performance of the MR page, lowering memory footprint this Q2 as well as providing other ways to enhance the experience.

Some examples:

* Commit navigation, to help commit by commit review: https://gitlab.com/gitlab-org/gitlab/-/merge_requests/28596

* File by file mode of reviewing MRs, including the concept of "unread diffs": https://gitlab.com/groups/gitlab-org/-/epics/516

We're constantly looking to improve and learn more about the struggles of our users. If you have further thoughts, please drop a comment in one of the issues and epics above and tag me (@andr3).

If some of your problems are not covered by these, feel free to open an issue (https://gitlab.com/gitlab-org/gitlab) and tag me as well.

Again, thanks for the feedback.

> It will collapse files with over 100-200 lines of changes for seemingly no reason (hiding the comments within them)

OH MY GOD this so much. This one specific anti-feature has no doubt caused countless bugs to be missed during code review. I can quote the issue thread by memory at this point i've revisited it so many times the last year(s?)

For goodness sake, if the trade off is between 'seeing all of the changes during the review' vs 'having the page load quickly', the only people who choose 'quickly' are those who haven't been burned by human error missing a collapsed diff. Gitlab made the wrong trade off here in good/fast/cheap triangle.

I don't know if any Gitlabbers are in the comments but I'd be interested to know: Is this basically your internal asset register open sourced?

I suppose they're not particularly top secret but there could be a version that has some more sensitive data and this is a safe derivative for example. Similarly, such an asset register would likely have non-technical assets as well while this just seems to be tangible, technical things?

I figure any large company with a risk management function will already have/need such a spreadsheet (or document), complete with some sort of data classifications to exist for auditing reasons so the interesting step is maintaining it out in the public.

Kudos to Gitlab for having the culture to allow this sort of stuff :)

Err, what is an internal asset register?

An asset register is how companies keep track of their assets. Any software license, computer, monitor, keyboard, copy machine, etc. the company claims as an asset is tracked in the register.

Basically an inventory system.

You'd be surprised the scale of companies that do not have this data. Or how many companies think they know where their data is but really don't. Especially with the proliferation of the SaaS services like you see in Gitlab's diagram. It's very easy for someone in the accounting or marketing department to sign up for something on their department's PCard and the IT / security / etc departments have no idea that company data is being kept there.

On the go to market side the registry is a Google Sheet. On the GitLab.com infrastructure side it is an app we made that isn’t integrated into GitLab.

I’m seeing if we can move the sheet to the handbook and integrate the app into GitLab itself.

And HN users only worry about “cloud lock-in”. I’ve said dozens of times that the typical largish company is locked in to dozens of SAAS offerings that would be a major pain and expense to migrate away from.

Once you choose your infrastructure, you might as well go all in. Just look at their dependence on SalesForce.

I used to love GitLab, and I still do in a lot of ways but I will no longer recommend them.

If you're looking for a shiny pretty interface for all your users, then Gitea can serve the same purpose. Personally, I find cgit to look fine too. I think GitLab simply tries to do too much, and they've built their entire app on a poorly scalable and bloated tech stack.

I think SourceHut probably provides the best platform for teaching developers a more native way to use Git that provides most of the same features as GitLab. For CI/CD there are so many better options outside of GitLab that integrate into Kubernetes and you won't regret it.

Why is salesforce the center of attention?

Because this is the Business Operations tech stack.

Correct! If someone is more interested in what GitLab itself consists of please see https://docs.gitlab.com/ee/development/architecture.html#com...

The diagrams are nice, but it would be so much better if the nodes were linked to documentation.

I agree. BTW I’m not sure that mermaid allows you to link items in a diagram to the docs.

Because it is the source of truth for all of your customers both current and future.

And surely customers should be the centre of everyone's business.

Like OP, I'd love to understand better what Salesforce does (so well, it seems).

Its a bit hard to imagine, just like SAP software (for me), since all your experience is on industries far from that world.

In their documentation Gitlab explains the usage they have of Salesforce, you may get a non-Gitlab specific view of it, for example: https://about.gitlab.com/handbook/customer-success/using-sal...

How much money do you wager that Salesforce license costs?

I know there's the whole build-vs-buy debate, but with the cost of Salesforce, might it not be better to just stick everything in a database and throw a MySQL WYSIWYG tool like phpmyadmin in front in order to let non-engineers see the schema and build forms?

I suppose if you're earning money hand over fist (or keep getting VC funds) then the cost of your Salesforce license doesn't matter. But I bet that hurts. And unlike a lot of the other software in use, Salesforce comes with steep lock-in.

You severely, severely, underestimate the complexity and business value of a CRM

Been there, done that. Agree with dtech.

You also get to a point in hiring where the field of sales/ops people judge you & your maturity based on whether you use a commercial tool like Salesforce/HubSpot vs build your own.

Can you imagine joining a tech org and they have a half baked phpadmin wysiwyg as a standin for github + code reviews?

While yes, you could calculate out how much the subscription is vs development time to build a simple CRM, the maintenance & mental overload is righteously terrible. Better to spend that effort integrating your foos to your bars than building a sales crm.

I was thinking the same. Nice vendor lock-in... I would love to know how much that costs yearly.

Marketo eh? My wife mentioned wanting that in her small business after seeing a shiny video but I said the price is probably not going to be appropriate (some research seemed to support this). Does anyone have experience using it? I have mixed feelings about Adobe but don't know how much influence they wield.

If she wants to sell to enterprise Marketo is a good option because of the deal size / volume. Google uses it for selling Google Cloud Platform. Otherwise Hubspot is probably a better idea.

Enterprise is definitely not on the menu, and Hubspot or one of the similar alternatives does sound more around the level she'd need. I had some fun messing around with Mautic and it could be fun to learn, but I'm not sure how its future looks.

I’m not a Mautic user, but its future looks brighter now that they were acquired by the company behind Drupal.

https://www.acquia.com/about-us/newsroom/press-releases/acqu...

Acquisitions can go both ways and apparently the guy who invented and put a lot of sweat into it (dbhurley) has parted ways with the project. Also they used to produce a lot of media on using it, but it's all a couple of years old now. Maybe that means they got the fundamentals right back then but it does send a mixed message. I would love to try it, I've set it up and it seems to be working, but I'm just not sure what'll happen if something goes wrong (their forum moves a little slow too). I have no special knowledge, just a gut feel and outsider impressions.

I’m wondering where the GL is, for when we outgrow Xero (which was about six months ago).

Looks like they use NetSuite.

This isn't a recommendation or advice but if you're after a ledger that's larger than what Xero can offer to a big org, Netsuite seems to be what Gitlab uses and is ironically enough also what Xero uses. Unfortunately (and somewhat ironically?) we're too big to actually have Xero run on Xero

If you’d like some feedback, for us it’s less a volume issue and more a case of design limitations. We run subledgers per customer in our own code for clearing and settlement (we’re a two-sided network), and posting the rollups back to Xero sometimes runs into edge cases and limitations. Recognition timing is one pain point, another is API-driven reconciliation.

What I’d struggle to give up is the ease of reporting. Xero’s report designer is right in my sweet spot for usability, flexibility, and domain-specific correctness, probably my favourite reporting tool in three decades of shaking down business software.

Thanks for your feedback!

While I don't work on the product (directly) or speak in any official capacity, I can definitely route your pain points and compliments to the respective teams internally :)

Gracious. That's quite a lot to manage.

Doesn't seem that much for Gitlab's size. And they are doing an effort to document it all and write it down instead of having it lie distributed in the minds of employees managing those services.

I do find it a bit weird you released internal handles and employee names.

I was amazed to see they doxxed basically every admin they employ... Why bother securing any data when you can just publish it?

Surprisingly, does not look too messy

Curious about whether Salesforce represents a single point of failure. If it goes down or is unavailable, what impact does that have on the rest of the organization?

That's why they bought OwnBackup

In some architectures it can be, which is a risk. A while ago I was in a day long kick off for a project to replatform a division of a large company, with Salesforce at the core. As it happens, one of the other division's Saleforce instance was down the whole day, bringing down everything. Internal systems, web, sales... everything broken.

We listened to the lesson and the architecture we designed ended up not having Salesforce as a dependency for any critical or money producing customer facing functionality.

We used Alteryx as a ETL tool to keep Salesforce in sync in almost real time. The customer facing sites and services ran off their independent databases. This also meant that the developers working on that didn't need to know the details about how Salesforce worked, we just mapped Salesforce data to their database.

Surprised to see Shopify in the list (used for marketing?).

Yes, our web store for swag.

This is a great stack to learn from.

Does anybody know some other companies that have their business stack available?

How did GitLab track and create this diagram? Is there a service that does it automatically?

It looks very much like OmniGraffle.

And adding a new system is a pretty big deal at a large company so keeping it up to date isn't that difficult.

I meant how do they track the data? If they're just doing it manually that makes sense though it seems inefficient...

See https://app.diagrams.net/ -- you'll recognize the style.

I use the same service, but I always knew it as draw.io. Never noticed when they started redirecting to this domain.

This is very valuable to me as a Linux user in a Microsoft Platinum partner org because I can't run Visio natively.

so that seems to be: https://github.com/jgraph/drawio

can such a diagram be generated using text markup ? i want to have this in revision control and then generate this in a static page (Nextjs or Gatsbyjs)

It’s a bit clunky, but Graphviz and its DOT language[1] might do the trick.

[1] https://en.m.wikipedia.org/wiki/Graphviz

mermaidjs [0] can do class diagrams that look kind of like this. It integrates into markdown via Jekyll easily, so I assume it will work well with gatsby.

It’s a basic text only structure so it actually diffs pretty well so I can see a merge request with a new node or new relationship.

[0] https://mermaid-js.github.io/mermaid/#/

I notice that Nanoc, the static site generator they use for their docs pages, isn't mentioned there. Maybe it's just too small to matter?

Very nice! Thank you for sharing, I wonder though, no fancy big-screen dashboard service in there? Do people not value that kind of thing?

They are a remote distributed team. I don't think their employees would value having to arrange their workspaces to accommodate a big screen set up high and at a distance from their workstation.

Dashboards are nice to have even on remote teams I think. Anyone can put up the dashboard or make a dashboard for their own screens if they wanted to.

No, it’s just a BO stack

Nice, just curious counting the filled and dotted lines - how this software is integrated together?

As a complete remote company, I wonder if GitLab's business affected much by COVID-19?

People still have kids who’d normally go to school, family to take care of, etc.

And there's the mental health and anxiety issues a pandemic and quarantine bring

Thanks for asking! I work at GitLab as Senior Manager, People Operations

Our e-team recently sent out a company wide reminder that we are family and friends first and work second. As people start settling into “new normals”, they also encourage team members to take the time they need. My direct manager is reminding me to take the necessary time of to help school my kids during our 1:1's and continuously check if I need any further support.

I work at GitLab and don't have children. My routine is pretty much... unchanged since COVID-19. That said, my stress is through the roof, as are other members of the team, not because of GitLab or deadlines or work, but life.

Those with kids, I know that it can be even more challenging or exhausting. so in some ways "unaffected" in others "feeling it". Pretty much like everyone else.

I can't comment specifics on finances, but I do count my lucky stars that GitLab raised when it did as it personally gives me a little more reassurance that we may be less affected. That said, obvs anything can happen to me or GitLab.

hmm.. something's missing. does the dev team itself not use any systems?? hard to believe

Is Salesforce still a big things?

Gigantic and for dev jobs it pays extremely well for easy work.

For a company that makes and promotes FOSS, their tech stack sure is dominated by proprietary SASS vendor lock in solutions.

As someone who has a B2B startup there is almost nothing in the open source space for Marketing/Sales Ops. And anything that does exist is really low quality.

And it's because (a) most developers aren't exposed to non-Engineering parts of the company so they have no understanding of what problem to solve and (b) it is incredibly lucrative to build a proprietary SaaS solution.

> there is almost nothing in the open source space for Marketing/Sales Ops

There is https://erpnext.com/ which is completely FOSS built on Frappe framework. If anything, it is rather higher quality than most SASS solutions out there.

It's not entirely clear on their website, but I assume the full software is at https://github.com/frappe/erpnext, and they have a model similar to Sentry?

Word of advice: stay away from Erpnext, don't get fooled by the apparent "clean" interface.

They might have some good conceptual ideas, but the implementation is horrific, their ops/delivery infrastructure a shitshow if even there's one, bugs all over the place and a product management that basically says yes to everything and finishes nothing.

They are in part victims of their own success, but just look at their Github issues page.

That kind of software is not fun or sexy, so people have to be paid to write it. Same goes for accounting where there are few open source solutions. Boring as hell, and profitable.

There is Odoo which can be self-hosted too and has a fairly generous open source edition.

s/This website uses cookies/This website got mugu'd by European policymakers/

Funny how they don't seem to really have any non-specialised tools to coordinate, manage and track work in a structured way across the company (like smaller team tools like Trello, or larger ones like Asana). The only cross company collaboration tool I saw on that diagram was Slack!

Makes sense considering Gitlab _is_a cross-company collaboration tool; they eat their own dogfood. They have a whole set of features that compete directly with Trello/Asana

They use Gitlab for that e.g. https://gitlab.com/gitlab-org/gitlab/-/issues/16653

It has kanban, issues, epics, portfolios etc and they use comments for capturing knowledge e.g. similar to Confluence.

You can watch how they use it here: https://www.youtube.com/watch?v=wpvaqtJjXd0

Wow, I didn't expect so much!

GitLab, here is my proposal: remove 80~90% of this. Keep the essential. Prove that you are a really innovative company. Stop following analytics. Engage with your users.

Prove that you can control entropy and recover your agility rather than going the bureaucratic enterprise software way, just waiting for a more hungry and foolish player to cut the grass under your feet.

Reply on news.ycombinator.com