Hacker News new | past | comments | ask | show | jobs | submit login
Business Operations – Tech Stack (about.gitlab.com)
223 points by tosh on Oct 23, 2020 | hide | past | favorite | 84 comments



Seems like a fairly concise list of Saas tools for a company of around 1,000 employees.

If you've ever worked for a company 10-100X that size, you'll find this list quite small.

To be honest, I clicked on the comments of this post to check my thesis that HN users would be astonished and was immediately validated.

It's always amusing to me how people who only work in one area of the company (ie. engineers, designers, sales, etc.) think they are the only ones doing "real" work. Everybody else is doing bullshit and wasting money. Without fail, every single time, if you give rank-and-file the opportunity to express this, they will.

Very few people in larger companies understand how the sausage is actually made, how leads are generated for the sausages, how sausage deals are closed, and how the sausage customers are nurtured for continued sausage business.


> It's always amusing to me how people who only work in one area of the company (ie. engineers, designers, sales, etc.) think they are the only ones doing "real" work. Everybody else is doing bullshit and wasting money. Without fail, every single time, if you give rank-and-file the opportunity to express this, they will.

Same here. I've spent my career hopping between functionally different areas - supply chain, IT, business operations, marketing, BI/Analytics, and a few hybrid/in-between roles.

As you said, every single time, each group tends to trivialize the contributions and efforts of the others.

They also tend to have very little awareness of their own unique privileges. When I'm on a business team, even if I'm doing technical work it takes an absurd amount of effort and approvals to even get local admin access to my computer, whereas when I'm on a technical team it gets rubber stamped and is usually part of the onboarding process. Same with infrastructure - if I need to set up a recurring process/script, it's pretty trivial finding/accessing infrastructure somewhere that I can plop that onto when I'm on a technical team. But god forbid I have the same need while working on a business team - you have to move mountains and requisitions to be able to accomplish the same thing.

There are privileges in the other direction, too. In particular, sales/marketing/account people expense an absurd amount of stuff in the daily happenings of their job, and can choose to do so with little fanfare. Whereas on the IT side, it can be near impossible to spend $10/month on something if there isn't a clear business unit to charge it back to and wasn't in the annual IT budget.

At the end of the day, every group tends to have different internal skillsets and organizational capabilities and restrictions. And all are working just as hard as the other to keep the sausage flowing, within the capacity of their abilities and role. The number of posts and comments on HN about devs struggling with marketing/selling their single person SaaS should be an indicator that you shouldn't trivialize any group's contributions before taking the time to fully understand them.


> you have to move mountains and requisitions to be able to accomplish the same thing

This is one of the (many) reasons it's good for techies to get out of their bubble. In many cases, knowing someone in another department makes it as easy as "hey, Susan, when you have time could you please..." and it's done tomorrow with zero paperwork.

I worked for a large corp for 17 years and I saw this a lot. After being there for a while I knew a lot of people. People would struggle to figure out how to get something done and my answer was usually something like, "come with me. I'll introduce you to Charlie in Technical Publications and he'll get it done for you."


Just sounds like people have't been able to walk a mile in anyone's shoes


Pretty much. What's curious is just how universally true it tends to be, though.

A personal hunch of mine is that it's also one of the implicit strengths of technical people starting companies. Software engineering teaches you to develop mental frameworks around systems and integrations and abstractions. And once someone with that experience walks a mile on the business side of the house, they're able to develop organizational structures that maximizes the organizational effectiveness as a whole rather than the local maximums within each group that are traditionally found.

Although that's not always what happens, it's an outcome that's unique to such circumstances.


Working at a fortune 500 company it feels like the tools that we have for planning alone is almost as long as that list. I'm exaggerating some but this is totally a small list compared to what it could be. The amount of SAAS tools that you want seems to scale pretty linearly with organizational complexity.


I'm fairly astonished how few duplicate applications are here.

I work for a company the same size, and we have many more applications than this, with lots of overlapping and competing products.

Kudos to management at Github.


GitLab?


I just moved closer to the business operations side of a 2000 person business that's owned by a 10x the size business.

Your comment has been my learning in the last six months. Holy mother of god there's so many things that keep the company running.


What about those of us that think our own work is mostly bullshit



Very, very familiar to the Tech Stack at our SaaS company with a similar amount of employees. They are also using our products too. :)


For anyone who is staggered by the amount of stuff here, it's worth noting that they apparently have 1,276 employees (which is a lot more than I thought they had). I'm not sure that really excuses having "an email signature marketing platform used to standardize brand consistency", but it does at least explain ADP, Salesforce, the pro LinkedIn stuff, and all the other things a little startup would never dream of spending money on.


Most people are only aware of the SaaS products used by their team or department. If you were to list out every single tool or service used across many 100-1000 person companies, it wouldn’t be much shorter than this list.

If you’ve never worked in sales, you’d be surprised at how many different SaaS products and services can go into running a successful sales department.

Sales people tend to be more sticky to their preferred SaaS products than engineers. You may end up with a lot of duplicated functionality across sales tools just because the VP of sales has always used certain tools for certain operations and has no desire to change that.


> Sales people tend to be more sticky to their preferred SaaS products than engineers.

Maybe sales get to choose their (SaaS) tools while engineers are forced into it by the sales department to a greater extent?

At my last gig Perforce was changed to Gitlab to "save money". Gitlab didn't work too well the legacy mess of 100s of scripts and 1000s of release branches so we got stuck with both. A really industry specific tool was deprecated by IT to save money with a cheaper replacement that did something else, that ofcourse didn't work out so we got stuck with both.

Etc.


Wait, how was it that the sales department made that happen?


As the broker of SaaS products for all departments in an all-remote SaaS startup, I agree with you.

Even with only a few dozen employees the inventory is confusingly-large.


I got a wake up call this week in just how much my company spends on “email signature marketing platforms” and it nearly made my hair go white, along with causing a strange series of emotions, none of them very positive-given some of the battles I’m fighting as a manager strapped for funding, and watching my team and other teams go through “covid related pay cuts”.


It's because all of the budgeting has been shifted to Marketing and Sales departments.

Someone complained about a bill from IT and we told the vendor to suspend the account for 1 month... it took three days before that same person had to come to us and ask why it wasn't working.


It's because all of the budgeting has been shifted to Marketing and Sales departments.

I'm sure. It's made for some painful discussions that I've just given up trying to have, and now I just point back to all of the saved email chains when interrogated for answers by superiors.

We're talking in the tens of thousands of dollars for email signatures.


That's per year? Sounds like a lot.

The execs think email signatures are obviously important because that's in front of their nose more often? Or why are they ok with paying a lot for that


I honestly don’t know. I asked my executive about it, and he seemed equally alarmed, but resigned to it? If that makes sense?


As a marketer, why are you so fixated on the cost instead of the comparative value it may add? Do you know what incremental value it is adding to the mix? Has anyone tracked the time saved from sales people and the impact of the touch point from an attribution perspective (presumably with Bizible, GA, or any other tools and data)?

It could very well be negligible lift, but if it stops a large sales team from having inconsistent, inaccurate, and off-brand signatures, automating and centralizing that could be worth it on its own merits.

Be cautious about being so quick to dismiss a tool just because it's expensive.


Be cautious about being so quick to dismiss a tool just because it's expensive.

How much caution, expressed in dollar amounts is warranted? If you had the visibility into our accounting that I do, would it change your mind?


Well that depends, right? If you’ve analyzed it thoroughly and come to the understanding that the ROI is poor and the budget would have been better spent elsewhere, then obviously you’re probably correct and someone else is not doing their job right.

But if it’s more of a shallow, knee-jerk reaction like ”this much spent on EMAIL SIGNATURES!? This has to be a waste of money!”, then perhaps you’re wrong to jump to conclusions quite so fast. You haven’t really elaborated, so we don’t know which one it is.


I didn't think it was wise, nor really my place to elaborate on the specifics of my company's finances, beyond to express-as a ranking manager-my surprise at the costs incurred weighed against common indicators a bit more relevant to a wide audience such as HN.


So, I made a point about ROI, which could very well justify it in the broader marketing and business context. You haven't addressed that point. Instead you've just made some vague claim about something you have visibility into about cost incurred.

Please elaborate as to why considering the return of the tool in question instead of just its cost is not relevant.


DVTRN is right, even if with attribution every marketing initiative was worth undertaking cash is still constrained from a fixed standpoint so you have to force rank prioritize what to purchase.


> (presumably with Bizible, GA, or any other tools and data)

You mean Bizible, GA, AND every other SAAS marketing attribution tool under the sun


I work for a huge multinational corporation and I'm going to count myself lucky that I have absolutely no idea what "an email signature marketing platform used to standardize brand consistency" is. Whatever it is I can't imagine spending money on it.


It's useful for customer-facing roles. From a sales context, it means that you can have a dynamic signature that includes a) up to date logos and branding and b) a dynamic list of upcoming marketing events like webinars, trade shows, recent blog articles, etc. put automatically in email signatures.

We don't use one (just have folks copy over the new signature to Outlook every couple weeks), but it means we have a highly manual process that is prone to error for non-technical users. It's definitely worth some money to get right.


My definition of a "useful" email signature varies highly from anyone who thinks that a list of upcoming events belongs in them. What a waste of bytes.


Look, people spend money on it because people DO look at these things. I don't think a big splashy banner in my signature looks good, but if I get even a handful of extra people attending my webinar then it DOES pay for itself.

This isn't some sort of idealistic view of the "way the world should be", it's just people doing stuff to optimize for their marketing & sales funnel, whether or not it annoys you or is a "waste of bytes"


Ah, so that’s where those unsolicited spam snippets at the end of each email come from!


You realize that Gmail blocks most of that by default, right?


Look, a lot of people insist advertising "doesn't work on them", but the whole hundreds of billions of $ industry works because it DOES have a reliable ROI.

I'm just explaining the world how it is, not necessarily how it should be.


Of those 1,276 employees, over 500 are sales or marketing people [0]. That's why you see emphasis on stuff like brand consistency and marketing. It's a surprisingly large portion of their business.

0: https://about.gitlab.com/company/team/org-chart/


Not that surprising. If you look at a bunch of SaaS S-1 documents at the time of IPO, marketing and sales are a material portion of the P&L and (presumably) an even bigger portion of the total cash outlay. Just seems to be what it takes to grow a public tech company.


Some exec sees an email go out with an old signature, or a signature they don't like, or who knows...

"I want everyone's signature to look like mine"

It then costs $x/month so it gets approved. As a bonus they get link tracking etc.


Typically spend that is tied to an anticipated incremental revenue is easier to approve ("this $20k investment will be worth it if we only get one $50k booking out of it"). That's also why salesperson variable comp tied to sales is easier to approve at higher figures- if a salesperson is driving multiple $M worth of sales, it's easy to justify a salary in the high 6 figures.


Title would be more accurate if it said something like "Gitlab: Business Operations - Tech Stack". I was confused for a moment why stuff like Rails and Vue.js weren't mentioned until I realized it's strictly about Business Operations.


Yep, maybe the mods can change the title to reflect that this is the software GitLab the company uses.

This was probably submitted to Hacker News after someone saw my tweet https://mobile.twitter.com/sytses/status/1319325053217992704 which has some questions and answers in the replies.

John already linked to the architecture of GitLab the application in https://news.ycombinator.com/item?id=24869339 in case you are interested in that.

Happy to answer any questions about either.


This might be of interest to you - https://docs.gitlab.com/ee/development/architecture


In the IT/Tech DD world, we call this "Back Office" vs "Customer Facing".

Some of these applications appear to span both.


> "Back Office"

Interesting, I thought "back office" was a super-set including both those (ie. CRM-y) tools and coded backend logic when a user makes a requests.


Most customers don't actually interact with CRM-y tools (usually). There might be integrations to (revenue cycle management for example) from customer facing to back-office, but that's usually the delineation.

But to your point, there's some grey area here.


I think you're confusing it with backend. Generally in traditional businesses you have reception/front desk where the customer interacts with the business and the back office where all the employees are doing various work.


Do you let the customers enter the back office in real life :-)


Ok, changed. Thanks!


Thank you!


They mention AWS, Azure, Cloudflare and DigitalOcean so I don't think that's strictly true.


They're paying money to (or have a formal business relationship with) those cloud providers, so it makes more sense.


Very interesting. The best part is I didn't know about the swag shop. Heading over there now...

I'm a huge fan of GitLab. Their managed service in particular is worth every penny. The sales people I've interacted with were mediocre and left a bad taste in my mouth though. Sadly typical for enterprise sales. Luckily the product sells itself.


It’s interesting to see how they use so many duplicative tools.

From CDN, transactional mail, eSignature and more.

I wonder why they don’t standardize in tools more.


For eSignature I speculated on Twitter https://twitter.com/sytses/status/1319360157113970688?s=21

For CDNs we switched from Fastly to Cloudflare because we needed support for SSH (port 22)

And for email sending I think you want multiple providers so you have a hot one (already know on the internet to send a large volume of mail in your name) in cause a provider has problems with delivery.


Lot of consolidation opportunity in the marketing/email/CRM space.

Could fairly easily collapse all of the following into your SFDC stack if you add their Marketing Cloud (fka ExactTarget) module:

MailChimp

MailGun

Mandrill

Marketo - basically a CRM like SFDC

Outreach.io

Sigstr - believe SFDC is building (or already has) similar functionality in their CRM

SurveyMonkey & Qualtrics - both survey tools that you could consolidate or potentially migrate to built in SFDC/MC survey tools


Because it's hard, and there are more valuable things to do. Everyone above a certain size duplicates lots of stuff.

There are entire consulting practices - big ones! - built just around helping companies standardize applications.


I am sorta writing a book "The Devmanual" - describing what I think is the basic set of "how we do things around here" - the idea is to procrastinate on writing the internal tools of my business empire before i actually hold my business empire !

The point I am trying to make is that all the tools for a business should interoperate and work with some degree of cross-programmability.

A list this long feels like it has broken past this concept.

It's not that gitlab does it get value from these but that there is more value in a programmable API level.

I think I mean that there should be a way to :

   x = get_hr_people_from_hrsaas()
   for p in x:
       use_sales_saas.calculate_commission(p)
It's still a stretch but i feel there is something there beyond IFTT


What do they use as file storage solution for their customer repositories on gitlab.com?



Correct, GitLab running on GCP file servers for git repositories and GCP object store for most of the rest (container registry, artifacts, etc.)



Thanks for catching that. I created a merge request to remove the .html extension from the URL. https://gitlab.com/gitlab-com/www-gitlab-com/-/merge_request...

You can see the BCP here: https://about.gitlab.com/handbook/business-ops/gitlab-busine...


That's a staggering amount of SaaS usage. Would be interested to see what the spend in total.


Our spend with vendors isn’t public because that vendors don’t want discounts published.


Sid, I'm amazed about how much detail do you know about your business.I've watched a lof of your unfiltered videos. Answers like this, sometimes, are not easy for CEO of +1000 employees. How do you do it?


Thanks! I think that this is possible because of the transparent culture we created. Other CEOs have the same level of knowledge. There is just no appetite in the company to risk saying things in public.

Yesterday someone showed us the new tech stack page in a meeting. We're all proud because it replaces a bunch of spreadsheets where each department (finance, security, people) tried to maintain the same list of apps with a single yml file. I ask if anyone has a concern if I tweet it, everyone is OK with it since it is already public anyway.

That takes a value of transparency https://about.gitlab.com/handbook/values/#transparency which has us be public by default https://about.gitlab.com/handbook/values/#public-by-default even when it is hard https://about.gitlab.com/handbook/values/#transparency-is-on...

And it takes us reinforcing these values now in 17 different ways https://about.gitlab.com/handbook/values/#how-do-we-reinforc...


you are amazing how consistently you live your values. thank you for being an inspiration to the rest of us.


Watching you on two or three unfiltered videos in different parts of the org (sales, product, etc) has sort of left me in awe of what a (as you self describe)”product ceo” is. Not sure how to say thank you, but thank you.


Thank you


So I see 1Password ordered first in the list and I wonder how many of the other services actually are having their credentials managed through it by a whole team, and not simply managed by individuals holding superpowers outside a 1Password vault.


Okta is also on the list and many of these SaaS support SSO.


Surprising how few of the items in the stack are open source.


Just curious do GitLab actively recruit new grads compared to other companies or do they only hire experienced devs



Cool. But now I want to know the tech stacks of other big players. Is there anywhere a list?


I talked to a developer from Mux a few months ago on my podcast that focuses on talking about tech stacks used by both small and large apps.

Mux serves billions of requests a month and supply one of the largest video APIs in the world.

That episode is at: https://runninginproduction.com/podcast/31-mux-is-an-api-bas...


Wow, I'm surprised at the amount of sales-related software that is used.


Why surprised? Sales is what's driving those big $ enterprise contracts


Nearly half the company (500+ of 1,276) is sales and marketing people, so it kinda does make sense.


I'm surprised there isn't more marketing and sales tech tbh.


Interesting that no anti-virus type software is mentioned.


I also don't see stuff like MDM software or Active Directory type licenses etc. Are they counting IT management as an outside service?


We use Okta (which is in the list) for AD purposes and JAMF (which I'll ask to add to the list but is on https://about.gitlab.com/handbook/business-ops/team-member-e... ) for Mobile Device Management (MDM).




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

Search: