Hacker News Reader

Comment by _bxg1 | Hacker News Reader

I remember reading this article about SAP a few months ago, and it just blew my mind: https://retool.com/blog/erp-for-engineers/

Here was an entire market segment I'd never heard of, and just the market leader is worth $163B. And on top of that, the technology is so totally alien to anything I'd ever heard of before. It's like an alternate-history kind of thing. Like if we discovered a continent somewhere in the Pacific that had never been in contact with our civilization and had steam-powered airships or some crap.

And I'm sure this is just one of many such industries, even just within the umbrella of software.

View on HN

That is fascinating to me, because virtually all my friends have heard of SAP outside of IT - as users or recipients. My engineer friend uses it for supply chain management, a beer store hq clerk helped implement it, construction accountant has used it, etc.

At the same time, I guess you have to be at a company of a certain size for it to be relevant to your life... But that perhaps proves the point - HN is very starup focused, and occasionally in a bit of a bubble (which is fine - I come here to get away from my daily mainstream IT grind and see excited people talk about cool things:), yet a hundred reasonably successful startups won't employ or touch near as many people as a single "quietly" successful traditional behemoth.

And if it helps understand the ecosystem and its seeking oddness, this is the crucial part of article:

"Implementing ERP isn’t just a purchasing decision: it’s committing to redoing how you handle operations. Installing the software is the easy part; adjusting the whole company’s workflow is the most of the work."

Don't think of erp and implementing it as a software development challenge. Sure, Developers can fscrew up*, an erp implementation but they cannot make it succeed - it is fundamentally a business transformation endeavour and must be undertaken as such to have a chance.

A good way for your organization to completely botch an ERP implementation is to fight the system every step of the way because your business is 'special'. Many people end up abusing customizations to the point of trying to emulate their old system in the new one, and nobody wins.

You just brought back some painful memories. At an old company of mine, I helped migrate a legacy system to SAP (of all things!), and they fought it every step of the way.

Well, management did. Every other phrase they uttered was "The old system did it this way" or "I used to just press F2". The pushback to learning something new was unreal.

Nevermind the fact that multi-step processes were reduced to a simple button click. Or that inventory and customer management were orders of magnitude better. The old system!! The old system did it another way!

I feel like you could replace "ERP implementation" with "business transformation project involving vendor software" and have a new Law.

The same with the parent commenter's thoughts, "Sure, developers can screw up [business transformation projects], but they cannot make them succeed."

> A good way for your organization to completely botch an ERP implementation is to fight the system every step of the way because your business is 'special'.

More generally, a good way to botch a business transformation project is to select technical solutions without a keen grasp of how they fit your business need, and to bullheadedly stick with those solutions as friction is revealed with the business needs.

Both distorting the business around vendor software and ending up continuously fighting with the vendor software throughout th effort (both of which are commonly cited “how to fail” routes) are clear (though superficially opposite) symptoms of that same failure occurring upstream in the process.

Written like a successful salesperson for competing ERP solution X, well done!

This is why it's important to tackle this from both ends:

- From the ERP end: you want it to be as flexible as possible so that it can reflect how the business actually operates and should operate in the future

- From the organization's end: you want to identify not only how the business currently operates, but how you would want it to operate in an ideal world

This reminds me of something I read while helping a business major friend with a case study a while back. One of the readings she had as background mentioned that companies like SAP that sell ERP are almost selling the process/workflow more than the actual software.

Part of me wonders if these things are more of a "no one ever got fired for buying IBM" thing than anything else. Are people buying the workflow because it's what they really need, or are they buying it because everyone else uses it?

1. You're absolutely right - you're buying software, a set of best practices, workflows, and a proven way to implement and use them.

2. HR, Finance, all the backoffice stuff - that's not your competitive advantage. It's not your core business. You just want to do it as efficiently and quietly as possible.

3. Therefore, in ideal case, a company realizes it is not a special snowflake, buys an ERP, implements it as vanilla as possible, and quietly goes on focusing on what IS their competitive advantage / core business... rather than spending years, millions of dollars, and tremendous risk, by trying to customize ERP to some obscure different process they just "must" have :-/

that happens a lot with enterprise software, too much customization and you end up painting yourself in a corner or run into roadblock after roadblock of platform gotchas. I do SF implementations these days and have seen it go great and have seen it turn into a total nightmare.

however, not once have I ever heard of an SAP implementation going well. It's a massive undertaking and takes total commitment from so many different people and roles in an enterprise. You need genius PMs and BAs along with a genius client team before you even talk about the actual technical people.

>>not once have I ever heard of an SAP implementation going well.

Sometimes it depends who you ask; I found that hands-on people on the ground and SMEs tend to be very negative about it, for personally valid reasons - it's a lot of work, it doesn't go smoothly, and it disrupts and changes what and how they do it.

However, there are quite a few ERP implementations where the top-level stakeholders feel the results are worthwhile the disruption and pain, and from that perspective it "went well" - the business objectives were overall met, whether it be capabilities or speed or reporting or integration/centralization or headcount or complexity reduction, all of it ultimately boiling to money at some level...

Now you're describing an implementation that is disruptive to the majority of employees "on the ground" and only worthwhile to the top-level stakeholders.

Only the top-level stakeholders felt the results were worthwhile the disruption and pain -- that they did not suffer.

But have you ever heard of one going well ?

As in, not just worthwhile to a tiny minority of the people on the top?

Yeah I bet it makes stuff more efficient, but clearly there's a disparity between the people suffering "disruptions and pain" and the ones who ultimately feel it has been worthwhile.

Another way to look at it is, the top-level stakeholders are responsible for the business and not just themselves or the lives of their employees. An effective leader is thinking about what is best for the business. Not just themselves or any one part of the business.

And maybe, just maybe, what is right for the business is bad for the majority of the staff. At least in the short term. I know that is a bitter pill to swallow, but it is reality.

> And maybe, just maybe, what is right for the business is bad for the majority of the staff. At least in the short term. I know that is a bitter pill to swallow, but it is reality.

Perhaps an example more palatable on HN would be Uber replacing all its drivers with software. Clearly better for the business.

The process often derails at exactly the “implement it as vanilla as possible” stage. Usually because someone doesn’t want to change what they do and someone else doesn’t care enough to make them change.

Another view of this - if the true purchase is of a group of workflows, are they purchasing for the things they know they need or the assurance that behemoth IBM/SAP/whatever has a process for whatever they didn't think about? Purchasing to minimise some future operational risk.

This is probably an understated benefit a lot of us in tech don't think about.

If you're selling widgets, and selling 100 units a month, there are growing pains on the path to 100 units a day and 100 units a minute that you can't anticipate, and that billion-dollar retailers have already figured out. So I think to the extent that physical B2C sales is pretty constant across industries, you know that if you're big enough to justify the investment in SAP or a similar ERP platform, there exists a process that you'll need 0.5, 1, or 5 years from now that they have and you can't really articulate today.

Sounds like Agile to me.

Some industries need your system to be certified for a software system to consume their data/services. You don't have much choice.

A lot of HN topics are about tooling for startups. Those tools are sometimes startups themselves, which can can make HN feel like a bit of a bubble I think.

Software for software developers can be a huge focus here, but it's just scratching the surface of software for the rest of the world. I was startled when I applied for a job at a company in my city, and they told me their product was entirely API connectors for the local banks. In fact all of the companies I've worked for here have been making software for local industries! From solar installers, energy efficient lighting manufacturers, local commerce and POS systems, mechanics software, creative market places, local lead generation.

I think if you want to start a business from scratch and you're not in SV, it's worth targeting those problems instead of problems that SV/startups have. You might not get sexy VC capital, but you might be able to jumpstart with some clients up front.

You won't need sexy VC capital if you can build a business that runs a good profit. Honestly I'm getting more and more cynical of the SF bubble that doesn't seem to be about making great software, but about getting a lot of investor money and going public - prioritizing short-term profits at the expense of the consumer and employees.

SAP was particularly successful because not only did they make money, but so did the Systems Integration partners.

At Accenture, for example, there were projects where 30-40 full time consultants were flown in from all over the country to do big SAP implementations.

...so you could see how Systems Integration partners heavily sold SAP products.

But my Blockchain Deep Learning Dog Walking app is going to change the world!! I went to Stanford so I must be right! /s

Dog walking is not a viable matchmaking platform business model because successful matches cut out the matchmaker, you didn't get the memo?

Humor fella.. he is kidding!

Have you seen your local software? It is rarely about making great software.. good enough.

Open source tries to make great software. Does the best job.

When good enough software solves the right business problem, it frees up resources to solve the next most important business problem. Good enough software runs many great businesses.

The one unicorn I worked for had so much cash to spend, there were at least 3 teams of 8-10 engineers on the same floor trying to solve similar business and infrastructure problems with different combinations of open source software.

Agreed. Running a company with a sustainable business model and healthy profits, not being propped up by VC money would probably do the local economy of one such a city perhaps more good.

"My engineer friend uses it for supply chain management, a beer store hq clerk helped implement it, construction accountant has used it, etc."

Is SAP something that can be somehow scaled down? if this is used for a beer store, it has to be really large store, if that's what is written in the cited article is true:

"A basic installation of SAP has 20,000 database tables, 3,000 of which are configuration tables. In those tables, there are ~8,000 configuration decisions you need before even getting started. And that’s why SAP Configuration Specialist is an actual job title!"

Ah,I guess context would help - "The Beer Store" is Ontario's [Corrected From "Canada"] monopoly province-wide beer retail chain.

You are absolutely right, a single retail store would not use such an ERP :)

Er, an Ontario-wide beer retail chain, owned by the province-wide beer wholesale monopoly, in turn owned by a cartel of foreign breweries.

Let me guess, you live in Toronto?

> an Ontario-wide beer retail chain, owned by the province-wide beer wholesale monopoly, in turn owned by a cartel of foreign breweries.

good thing we have SAP to facilitate the development of large scale business, such as monopolies and cartels.

I don't think the existence of SAP is particularly relevant to a monopoly that exists because of statute. Although I don't live in such a state, I know there are places in the US where all the liquor stores are government run.

Corrected! Thx :)

SAP has two scaled-down ERP products:

https://www.acclimation.com/au/solutions/sap-bydesign-vs-bus...

To add some context (for your curiosity), I'm not in The Valley and I'd never worked at a true "startup" until very recently, but I have only ever worked at small (<100 people) companies. Except, that is, for the summer I interned at Hewlett-Packard. Not sure how it never came up there; maybe they just have their own equivalent system.

I had actually heard the name SAP come up in passing at my last company, who had lots of ginormous clients. But all I'd really gathered before reading the article was "it's some kind of database, or something".

> Except, that is, for the summer I interned at Hewlett-Packard. Not sure how it never came up there; maybe they just have their own equivalent system.

When was this? Leo Apotheker became CEO of HP in 2010, after 22 years at SAP (two as CEO). He then presided over a sequence of colossal fuck-ups, lost 40% market cap, and was fired less than a year later.

HP definitely knows about SAP, but I wouldn't be surprised if they're not big fans.

Correction: 22 years at SAP, not as CEO.

He ran sales arms of SAP for 20 years, not SAP itself. After he became CEO of SAP itself he was fired after 2 years, then he joined HP amd he was fired from that after one year. Pretty rough end to his career there..

... or maybe it's the perfect ending?

I'm not sure about his SAP earnings, but he sold HP an 11 month CEO for upwards of 27.5M[0]. That sounds like the perfect way to end your sales career ;-)

[0]https://www.nytimes.com/2011/10/01/business/lets-stop-reward...

Promoted to the level of his incompetence perhaps.

Both hp and SAP are absolutely gargantuan companies (try guessing how many employees each had by the end of his tenure- unless you actually know I'll bet you'll underestimate it). They are also old (for computer related companies at least) and have lots of products for lots of clients.

Not to say he wasn't bad at it (I honestly don't know) but it's a hard job to get right.

SAP goes deep. It goes as far as to the factory floor when the technician is measuring if the piece is within tolerances and from Mitutoyo caliper sends the data into some kind of database, or something ;)And there it is, on a screen from 1994, SAP proudly on the screen. It's everywhere.

I worked for EDS from late 90s to early 00s and SAP was _everywhere_. Projects were scoped and framed as "how can we fit this into SAP?" - and not "how can SAP help this project?"

SAP is/was virulent in nature. If you used it for one thing it damn near forced you to use it for everything. A significant chunk of the work EDS were undertaking (as IT outsource) was migrating to SAP.

Our timesheets, project plans, data, hell even a huge chunk of our app development was done in SAP. We were paying through the nose for SAP contractors too.

That... explains a bit of the herding cats commercial.

> At the same time, I guess you have to be at a company of a certain size for it to be relevant to your life... But that perhaps proves the point - HN is very starup focused

That’s pretty dismissive. I work at a company with ~80B in revenue, I never interacted or heard of SAP in work context. In fact it is only incidentally that I’ve heard of them previously. There are hundred of thousands of engineers working in companies with 10-11 digit capitalizations that will never interact or hear about SAP.

Edit: to clarify because I am getting a bunch of downvotes. My point is not about SAP’s penetration, it is that lots of engineers (even in companies that use SAP) might not encounter SAP products. The perspective “of you haven’t worked with SAP you are in a small org” is wrong.

-“ The perspective “of you haven’t worked with SAP you are in a small org” is wrong.”

I would fully agree with that, my generalization was kind of the opposite - if you work at a small company you are unlikely to use or work with SAP.

My other generalization was that HN was focused on startups, their technical culture, the cool and interesting, and almost by the definition not on boring bread and butter aspects. Does not mean that people/users themselves do not work there - I have spent twenty years on PeopleSoft/Oracle myself, but that is not why i come here and yes, the overall zeitgeist on HN toward e.g. Oracle is pretty consistent :->

FWIW, comment was not meant to be dismissive. I was genuinely fascinated, late at the night, at the notion that IT industry professionals might not be exposed to a behemoth such as SAP or even ERPs at large, and was pondering as to how that might come to be; there's absolutely personal bias & hubris in it, which is why I tried to understand where it came from and what's the diff between background of my friends and colleagues, vs perhaps some other IT professionals :)

> The perspective “of you haven’t worked with SAP you are in a small org” is wrong.

The fact that you are arguing against this might be why you are getting downvoted. The original comment was saying “if you are in a small org you probably haven’t worked with SAP”, which is not the same thing.

Edit for further clarification: I can see how it might look like they were arguing what you stated, but it was more of a Bayesian inference. Many HN readers are at a small company, therefore a plausible reason that an arbitrary HN reader might not have heard of SAP would be that they are in the large pool of HN readers that are {in small companies and, therefore, do not work with SAP}

You know, I actually like your Bayesian interpretation. My main dislike was the dismissive tone in OPs phrasing, but I think his proposition might be OK.

There is nothing remotely dismissive in the (100% accurate) sentiment that small organizations have no need or desire to use large Enterprise-among-Enterprises applications such as SAP.

Your firm (KP) does use SAP's BusinessObjects business intelligence software along with SAS analytics tools and to "support its data analysis activities against it's Oracle EHR system".

Since you are in the machine learning silo, your perspective may be a little skewed.

To be fair BusinessObjects is not SAP. It's just one of its acquisitions.

Well, if you look at it that way then Powerpoint is just one of microsoft’s acquisitions. After a certain period an acquisition becomes fully native to that business. In my view BusinessObjects passed that point with SAP a while ago.

Sorry, you are right, I should have been more specific. What I meant was BusinessObjects is only the reporting component they acquired to complement their product suite. It's not "the product" SAP sells (i.e. ERP).

"SAP" isn't a product either, it's a company name. When certain people talk about SAP they actually mean R3 or ECC. R3 is dead and the sun is setting on ECC which is being replaced by S4/HANA (which some people will still call SAP). SAP has a lot of products and it's main products have a lot of modules so two people using "SAP" may not even recognise each others versions.

It would not surprise me if KP were using ECC for their finances or some other area the parent comment is not familiar with.

Sure, but lots of people don’t interact with it and would not know about it. The fact a company uses it doesn’t mean most engineers would know about it.

Companies that make things often have components managed by the ERP system. This flows down to the engineers. A software shop never deals with such things.

There are also alternate vendors of ERP. Microsoft Dynamics NAV for example. At my current shop I manage INFOR VISUAL ERP.

...Microsoft Dynamics GP, Microsoft Dynamics AX, all entirely different products for different sectors. Then there is Netsuite, Sage (lots of different products), SAP Business One (small SAP), Oracle ERP, JD Edwards...plus 1000 niche providers who are often migrating to NAV rather than their own product. It may be a shock to a FAANG, but the ethos is very much, 'move slow and don't break anything'. 10 year old versions still supported, that sort of thing. Another reason Windows is popular in enterprises is that ERP clients written in VB6 (or whatever) still run on Windows 10!

Windows is popular in enterprises for like, a million reasons. As a Windows Systems Administrator, The entire Windows ecosystem is built around business and business processes and management.

And was that a FAAMNG? Presumably the larger, "new-wave" tech companies roll their own solutions of some kind

While Microsoft does sell their own SAP competitor they actually use SAP for their main accounting.

https://www.erpsoftwareblog.com/2012/11/why-does-microsoft-h...

I'm pretty sure I've seen job adverts from Google for people with experience of Oracle Hyperion.

What would be the point of Google writing their own financial consolidation software?

You'd think so... I thought so once... but again, perhaps not necessarily - ERP is NOT their core competency.

I.e. HR & Payroll legislative rules and best practices, business & organizational process transformation etc, are not what Google Engineers and IT necessarily have experience in, and if they allow hubris to override that fact, it might not go great.

Internets indicate that Google uses Oracle, Apple uses SAP.

This info may be outdated - for example, Facebook claims they built their own ERP just recently; I do wonder what that "backend database" part is though, or how much HR/Pay/Finance, as opposed to inventory/datacentre management, is custom: https://techcrunch.com/2019/08/15/how-facebook-does-it/

I work for a large healthcare company. My point is not whether we use it (as another poster pointed out we apparently use a product that was acquired by SAP), it’s that lots of people might not know all of the business software that is used for various reasons.

How many more letters are we going to add to this acronym?

Do you fill in a timesheet using a computer ? If so, there will be SAP or another ERP system behind it.

SO what ERP system are you using? What do you sell? How do you manufacture? How do you handle inventory and finances?

You're either using SAP and you don't know it or you're using Oracle.

EDIT: just looked at your linkedin profile. I can see that your company literally has SAP consultants working for them full time. I don't know which SAP products you are using but you definitely are an SAP customer.

> You're either using SAP and you don't know it or you're using Oracle.

Or you're using NetSuite (which nowadays is Oracle, granted, but a rather different product from what most people think of when they talk about Oracle or SAP in an ERP context), or you're using Microsoft's ERP (Dynamics something or other? Never used it), or you're using something developed in-house (very likely running on an AS/400). Or hell, you might be running multiple ERPs in a weird limbo state mid-transition (e.g. Thermo Fisher, which was using AS/400 and SAP (or maybe it was Manhattan? Or both? My memory's fuzzy...) - the latter of which might very well have been partially implemented on AS/400 in order to integrate with the existing ERP - as of a couple years ago when I was working for them).

It's worth noting, too, that there are plenty of large and arguably-dysfunctional companies not using an ERP at all, or (like with my current employer) it's relegated to very specific uses because other functionality got offloaded to other systems (for example, it's relatively common to run a WMS separate from the ERP and integrate between the two, as it is with an eCommerce system; in my employer's case, all three are separate).

I have heard of sap in tons of contexts in undergrad business classes, news articles, and things like that. Though I have never come across an implementation in the wild. Though my experience in large enterprise is limited.

I'm in Germany and of course everyone here has at least heard of SAP and knows it's huge, but I guess not many people know just how huge. SAP consultancy is a big market here with lots and lots of job openings. I knew a few people who did that, it's usually a very well paid job, but it's well paid because otherwise no one in his right mind would do it. From what I heard, it's incredibly boring while at the same time incredibly stressful when you're on-site. Also, I've never met anyone who actually liked SAP.

I've heard very similar things about being a SAP consultant here in Europe - basically, it sounds like a job for someone who:

- understands technology, but is not a big fan of it

- can communicate with clients

- can handle stress well

- is into salary maximization

Doesn't mind living out of a suitcase.

A friend was telling me how they make over $200 an hour. For a job that doesn't require a technical chops of a fullstack dev, that sounds crazy to me but doable.

SAP consulting is like some sort of mythical club,where only the chosen ones can join. SAP isn't something one can learn from his bedroom and then go around advising multinationals.You have to somehow get trained on a job where it is used. So the vast majority of SAP consultants pretty much owe to the fact that they lucked into it.It is however, pretty lucrative though.

I liked SAP, especially compared to what I have used since. We have never met, so your point is still true! It might be some time before Germany lets us English types in again!

Want your mind to be blown even more? I worked for a multinational company, I would not be surpised that this company is among the top 10 SAP customers worldwide. Well, everybody hated hated hated SAP, it was a huge hindrance instead of any help.Any use beyond of a "dumb" database was a huge PITA, and I am not fully convinced that all in all it helped more than damaged the business. It was clunky, unintutive, slow, ugly, archaic and a long etc. There were these guys, that were trained as "experts" but they frankly behave more like witch doctors. SAP is coasting just by its name and legacy but from my POV is ripe for the taking. It is not easy though, ERP systems are hard, and changing them ever harder, but in my fantasy world I would like to develop some program and would be very happy to skim 0.00001% of their income.

SAP's ugly green, old style-UI screens are wonderfully optimised for data-entry. You can be ultra-fast. It's like VIM. It's like driving at 300 mph.

Its all the new fangled dumbed-down, web-ui's that is sheer crap and leaves your head scratching.

Well..the old UI's were coded by old-school, C programmers who started their career in the 70s and 80s. They even wrote a multi-process Erlang style kernel to back the application programming. Performance came naturally from their gut.

Nowadays, we get Electron and congratulate ourselves.

Saw a big write-up somewhere about this. It's easy to learn the muscle memories for various patterns on a text only UI, and you can become extremely fast.

But almost no one is fast with a mouse based GUI

Excel.

Indeed.

I did once get an explanation of how SAP screens work when a former employer was considering buying SAP ERP - it reminded me at the time of how IBM mainframe terminals work.

I worked at an $8bn multinational and we ran the whole operation through SAP from quote to cash... and we really liked it. It was consistent, and the 'ugly' green-screen style data entry is wonderfully efficient. I hear this all the time though, about all ERPs. It is all about how you use them.

Same experience here I work as a SAP Developer at a €14bn multinational.

I had to work in the store for a month and got good feed back from the workers there.

We try the smooth the user experience with web apps for tasks that are done often so it's not to confusing.

But the impotent data entry task like order entry and so on are done the old interface because it's more efficient.

Same with accounts generation.

Almost certainly most operations accounts and tax filings could be fully auto generated from their bookkeeping - if the bookkeeping was good enough.

You could do that by constraining the bookkeeping and adding some intelligence to the reporting.

So why do bookkeeping firms like Xero stop at basic bookkeeping? Because they sell via accountants - who make a lot of money piping data from one computer system into another to change the format a bit.

Enabling certain people to maintain their mystique is lower hanging fruit than pushing for full automation.

Plus who wants to program accounts systems...

'Almost certainly most operations accounts and tax filings could be fully auto generated from their bookkeeping - if the bookkeeping was good enough.'

I firmly disagree. I am supposed to be reviewing accounts right now....accruals, prepayments, deferred income, the difference between capital and expense, capital allowances, R&D allowances, dividends vs salaries, consolidation of groups. The average bookkeeper is not expected to work at this level. That all requires judgement and experience and expertise. That is why Xero stop at the easy bit

That sounds all pretty trivial.

I wonder if you have worked as an accountant? Or at least took some intro courses in the area? Or at a bare minimum took a look at common corporate financial statements?

This is such a typical dismissive reaction by a person who has ZERO insights into the domain.

If it is so trivial, why don't you do it and disrupt the market? Some very trivial millions are waiting for you to just come and get them, what's holding you back?

I don't do it because I care more about my time these days than about making millions :-) And why would I want to do something that is so obviously boring?

Trivial stuff takes what, an hour, a week, a month at maximum to do? So you are trading the opportunity to be financially independent for the rest of your life for what, your free time on a span of one month? Yeah, that does sound reasonable.

Good, then we agree :-)

Ostensibly you’d rather spend your time on HN than making millions.

Definitely! There is no doubt about that :-)

Trivial in the sense of integer multiplication being easy; non-trivial in that it's full of weird & wonderful business rules, many of which are barely written, and strange interactions between different markets and contracts etc.

Half of it is just the business & people skills to extract the necessary information. Harder than you'd think, and that's just asking people for stuff!

For basic bookkeeping, yes, you can generate all the needed reports and tax filings from the books. I'm in Europe and I've had a couple of companies where I had no accountant - I just entered all the transactions in the software and it generated all the necessary filings for me. Payroll is also auto-generated, taxes handled etc.

For complex accounting and analysis you still need accountants and business analysts. It's a different type of job in that case though: instead of handling the common cases, you handle the edge cases that the software is unable to deal with.

Harder than it sounds. A complete drop-in replacement requires an almost unbelievable amount of internationalisation and research. And even if you can get all of that lined up, B2B sales and marketing - and inertia - are a huge challenge.

It's not that it couldn't be done, but it's very much a moon shot and it's not going to happen without someone charismatic and experienced driving it and selling it.

Specific niche feature SAP-a-likes are more plausible on a smaller scale, but sales and marketing will always be the real challenge.

Good luck! You're time to market and the sunk cost fallacy are overwhelmingly against you.

There are smaller ERPs too, but I think often the real alternative is building something in-house tailored to your specific business.

That's basically what Salesforce did, but they ended up skimming a lot higher % than that

I've never seen an et cetera qualified :D

> but in my fantasy world I would like to develop some program and would be very happy to skim 0.00001% of their income.

Which is exactly what SFDC, NetSuite, etc have done. There's definitely room here.

Workday is in the process of doing this. I wish them all the best, but I can't help thinking that we're just replacing the old goat with a new goat.

I did that and was successful to some degree, but it's not easy. I would've done something else if I knew that back then.

You skimmed a small part of SAP’s market? Like, a specific niche that you built a better tool for?

I can only imagine that if you create any tool worth competing with SAP, your customers also have to be huge companies.

Yes, we competed with SAP in a specific industry and honestly our tools were better and more user friendly, and no, they were not all huge companies, which is one of the reasons we were able to eat SAP's cake because we were much cheaper. I believe we got the job 90% of the time when we were shortlisted with SAP.

We lost some customers to SAP later on because they become bigger and "consultants" told them that they should switch to SAP because, you know, all big companies use SAP. I haven't heard of any success stories from those customers yet (this is not entirely SAP's fault though).

Which SAP product did you compete with?

Coming from the opposite direction (mgmt consulting for big corps -> startup), I was very familiar with the ERP concept, but am still super confused by the ERP market. Every so often I look for an ERP for our company, but am always disappointed by the options. It seems there's an uncanny valley between Quickbooks and SAP, but maybe the issue is just one of discoverability.

We've slowly stitched together an assortment of SaaS tools, custom built modules, and Excel files in the exact way I swore we would never do. If anyone's been through this hell before, I would appreciate some help.

I have been doing ERP consulting for about a decade now and I tend to disagree. I think we have a lot ph ERP options now in the paid space apart from just Quickbooks and SAP. You need to pick and choose based upon what is the most complex operation for a company- be it multi currency financials, complex supply chain or some other parameter. Then you can pick from lowest complexity to highest in the following manner

* Quick books

* Sage

* Dynamics / NetSuite

* JD Edwards / Infor / Manhattan

* Workday

* Oracle EBS / SAP

Each of those on the same line are similar in features and complexity.

This is obviously not a complete list - just the ones I have dealt with.

Quickbooks will do you fine up to $100M in revenue or 8-10+ years in operation. When you cross either of those bridges, you should think about implementing a "real" ERP.

But this doesn't happen until your team (accounting/finance/CEO/CFO) start running into roadblocks in either piecing together financial statements, or the month close process starts taking weeks of manual labor, or there's a HUGE accounting issue and everyone discovers the forecast is way off and the company has been burning cash faster than expected.

At this point you have the organizational momentum to do the crazy amount of work required in getting an ERP up and running, which requires getting most of your business flowing into it so that it'll even be remotely useful (otherwise everyone goes back to Excel).

What needs to flow into it? Everything you're doing by hand to update your three sheets. You need to get revenue/income in there (sales orders, purchase orders etc.), product costs if you have them, payroll, bills (internet, office supplies). Hopefully you use something like Bill.com, and this becomes a (slightly) simpler process.

Maybe you have vendors in Europe where you pay in Euro, so you now also have to figure out how you're going to convert currencies. Maybe you have warranty on your products, so you have to balance accrual accounts. This list goes on, and on, and on.

Every company is different, and you might not even recognize the overlap between two ERP launches besides the fact that they're using SAP.

>Quickbooks will do you fine up to $100M in revenue or 8-10+ years in operation. When you cross either of those bridges, you should think about implementing a "real" ERP.

I think that very much depends on what market you are in. Sure if you are a Saas, but not if you are a European distributor with stock in two countries with different currencies. Stock adds massive complexity. You might be reaching for a lower end ERP at $5m of turnover.

Stock in two countries is nothing that can't be managed by hand without an ERP. I know, because I've done it before.

You usually reach in inflection point where your manual process sans-ERP is costing the company real money, and the ability to make quick(er) decisions is hampered by your current process.

It also depends to an extent on your legal accounting / reporting requirements.

Friends who had businesses in both Poland and Singapore tell me that the difference was like night and day.

Wait until you start accounting for stock compensation.

Yeah I thought the same, I've worked with, against and for SAP since I turned 21. My whole career has spent working on ERP on various platforms.

SAP is a German company so here everybody knows about it. There is even a startup strategy called "unbundling SAP" where you pick one feature of SAP, do it 10x better and target a smaller but still huge market. Examples for such companies are: Personio (500M valuation), Cellonis (2.5B valuation)

Sounds similar to the "unbundling Zapier" article that made its rounds here on HN a couple days ago. Definitely seems viable.

mind sharing a link of this article, can't seem to find it

https://kamerontanseli.ghost.io/first-it-was-craiglist-next-... http://news.ycombinator.com/item?id=23126455

A friend of mine worked for Sun in the late 90's and mentioned SAP to me.

He told me companies would buy top-of-the-line (overpriced) hardware and pay a premium to get SAP up and running, because they would plug their company into it and recoup their costs almost overnight. It would just do say currency conversions or other super-boring non-core-business things better and save a pile of cash.

You can't save yourself rich, but sometimes the difference between a profitable quarter and a horrendous loss might be your accounting software or your tax software...

Oh yes, there are tons of giants creating software for health applications you never heard of, defense purposes you cannot hear about, logistic you would never consider (managing competitions, port services, private jet traffic, truck fleets...) and so on.

And those don't use $LATEST_TECH, $FANCY_FRAMEWORK or $EXOTIC_LANGUAGE.

They use $BORING_TECH, $OLD_LANGUAGE, $DEPRECATED_FRAMEWORK.

They are not in the cloud, don't update often, have to deal with legacy stacks, stacked on a stack of legacy stacks, and administrative burden from left to right.

Those are the ones that pay for an Oracle licence, and sync it to dozens of excel sheets from an FTP and a lone Access DB, where all auth is manage by a Tomcat service actually proxying to Active Directory and their custom CRM.

And they certainly don't have a team of 10x programmers. In fact, most are windows dev that never used the terminals and don't have admin rights on their machine.

I should know, I make half of my income going from company to company to train their team.

They are everywhere.

They make billions.

SAP is amongst the least fun software category which is ERP systems.

Highly customised, to the point that upgrades are impossible.

Extremely complex systems. So complex that vendor lock in is assured, and the lock in means big prices.

And yet a good business because ERP systems are at the core of any medium to big business.

It's grindy, boring, challenging without being fun work, often using non standard technologies that have been developed by the ERP systems vendor, which help ensure lock in.

ERP companies like SAP and Oracle are the absolute best at making money on services - they know how to charge huge huge amounts and get away with it. There's TONS of money there.

I wouldn't want to work in that area though.

My sister-in-law graduated with a degree in Industrial Engineering, and got exposed-to/proficient-in SAP at one of her first jobs. Now, she does SAP consulting work, much of it from home (even before COVID-19), at a significant pay bump and highly flexible hours.

The crazy thing is for most of the world people use Windows and SAP and firewalled intranets and no external email. Apple laptops and free access to the internet and SV culture is a tiny minority of what the world's computing looks like.

I’m nominally a DevOps engineer at a really big Germany-based auto parts (among other things) manufacturer, specializing in Azure and Kubernetes.

My real job many days is helping devs make peace with our various web proxies - even the special, relatively open, non-SSL-intercepting “dev proxy” requires CNTLM, which just celebrated its eighth anniversary on its current build. And of course, my little darlings take dependencies on it for things for which the masses absolutely not have access to it.

as a designer working in enterprise side application development - so much this. So many design portfolios are geared towards consumer side apps but the world is powered by the most dry boring stuff.

[I finally got around to making an account after lurking for so long. Your post was the catalyst ]

I'd imagine the smart startup engineer folks will never know o this market because they're either working at some place like Google or doing their own thing, and one would need to have worked at a less sophisticated blue chip company to be exposed to these flawed systems.

Some people also worked in finance. That can pay well, but has plenty of SAP or similar crap to deal with.

When I was starting school, my dad had informed me that SAP was HUGE, and consultants with SAP skills were pulling in 200k, back when even google employees were still waiting to IPO.

I've got a - maybe silly - question: Do you think that software engineers read about the business of technology?

I mean, some of these behemoths (or used-to-be behemoths) of technology, like Oracle and Intel and Microsoft (Ok - everyone knows them) and Dell and Apple (mostly the 2nd one) and places like SAP (which is arguably the biggest database and logistics and supply chain (sans the EDI middle-ware vendors on the planet) and Synnix (which is arguably the largest ODM supplier of white label servers to the largest data centers in the world), and, well, Wistron? BBG? (Submarine - https://www.visualcapitalist.com/submarine-cables/) Satellites? Aerospace? What it takes to wire our world together?

This is not a criticism - it is simply curiousity - as a tech person - you don't read up on (or do you think people do or do not read up on?) the way the world works technologically?

My son graduated with an engineering degree recently and his friends always want me to take them to lunch (free food!), but also so I can tell them how the world is connected (power, telecom, SDN, Satellites (launch - cradle-to-grave), and I'm always a little amazed because I grew up libraries of books and they've got google, but want to be lectured?

I guess I don't understand. If you have a moment, could you edumacate me?

>I grew up libraries of books and they've got google, but want to be lectured?

Your son and his friends know you. Your experience is lived, and real, as real as you sitting next to them. Reading books is abstract, the language can be obscure, archaic, verbose, flowery, dense, etc. Your son and friends know how to read your emotions and communicate effectively with you. An hour of your stories could likely replace dozens to hundreds of hours of reading because you can tailor your stories on fly the in a way no author ever can.

Thank you for that. (I'm not being glib.)

EDIT: I think your learning style and mine are much different. I remember when I had to do a biological proposal to DARPA (I guess I didn't have to), I went to the University of Texas bookstore and got every book they had on molecular biology - and read them over two days at a Starbucks on the deck (beautiful day, but hot), and then wrote the proposal and had the interview. (I'm an EE,CS and Physics guy who knows jack-shit about molecular biology). I never figured I ever had time to talk to anybody as long as I had 30 vertical centimeters of books on the floor that had to be processed by dawn. Maybe it's a generational thing. Maybe I'm just broken.

Whatever you are, you're not broken if you can actually process 30cm of books in a few hours.

I do read about these things, but I can also spend my time with lots of other stimulating/important stuff. Priorities, maybe?

Fair enough - I was just curious if the average modern software developer read books by Andy Grove ("Paranoid"), or Crossing the Chasm, or the biography of Steve Jobs, or kept track of what's up with the top 250 global tech companies on planet Earth every week for 20 years.

Now, that's a little unfair because I read really fast.

But - I guess I was just curious about how a tech person couldn't know about SAP.

NOT as a criticism, btw -- just I don't know how this could be? They are not a secret company. I just (apparently) don't know how young software engineers are wired? (Yeah yeah - Ok boomer, I get it)

But, because people actually do matter to me - I would like to understand better how they are wired...

EDIT: I am also stimulated by my healthcare projects and my energy projects and my stupid stupid project to build a carbon-nanotube elevator to the sky (which may have more funding than practical deliverables!), but I am completely stimulated. I assumed you were too, so it wasn't intended to be an either/or.

EDIT2: If you don't know the answer to the question I was asking, don't worry about it. Nobody else appears to know either. Just because I have a question doesn't mean that there is an answer. Thank you for engaging and have a good night.

I think it’s a bit of a personality type thing. I’m 30, know a fair bit about SAP and ERP systems in general despite never having really used any (small electronics company I work for uses a really crap MRP from a tiny company in New Zealand, which I suspect was probably selected because the operations manager was mates with somebody who implemented it it and/or because it was cheap and our accountant is a penny-pincher).

But for me, I just find my self researching all sorts of things that aren’t even related to my job. From aviation to renewable energy to database systems to carrier network hardware to finance and economics to chemical process engineering, etc. I just find myself almost infinitely curious, and meet others like that occasionally but it doesn’t seem to be how most people are wired. I don’t think it needs to be the norm to be for most jobs, but I find it’s a huge advantage for the kind of higher level product architect/systems engineering roles I’m doing more of.

In some ways it’s detrimental - I find it really difficult to keep motivated to do the deeper-level nitty-gritty engineering tasks because I’m always more interested in everything going on at a higher level in our products, so when involved in hiring I don’t try and select for people like me. I think tech companies need some but also people who are happy to focus on deeper-level work.

Thank you for that.

I would guess - at least for a lot of developers in my cohort - we came into tech at a time when every few years there was some big improvement to software productivity. Staying current with language features, new languages, and new paradigms could be the difference between getting that next job or not. Especially since people younger in their career are often expected (yes, even in tech) to be current with all the latest and greatest. A new graduate with only C experience is going to have a much much smaller pool of potential jobs than one with React and NodeJS. We've gotten trained (and paid) to assume the old isn't effective, and the new must be better. Obviously this isn't always the case, but I think it leads to a culture of obsessing about the new hotness and ignoring everything else.

That makes perfect sense. They are looking for jobs. (Which is not bad!)

"We" were going to build a new world. So I guess we were some kind of idiots. After I sold my first "game" on an Apple 2 in 1984 (they paid me a lot), I went to work a semiconductor manufacture, and AT&T and Apple and General Magic and Motorola were going to re-invent the world. Good engineering work - bad vision. Didn't work.

Meanwhile in Japan, where I worked, I translated Nintendo Games (In addition to the adobe and lucas arts stuff, which was more for Sega) to the 6502 (hey, I'm old), and made money trying to make the meet the new.

I think it was easier back then. I think there was less intergenerational hostility.

Why do we have inter-generational hostility? I know how to bring the old arcades alive! :-) (Actually I know how to activate satellites - if we want low-cost back-haul - packets here - packets there - There are 0.1% utilized GPRS networks out there.

AND Now we are getting into topics that you don't want to hear about.

I do really appreciate your willingness to share tonight. You broke open a couple of pre-conceptions. And I guess that's the reason you talk to people.

Thank you.

This is something a lot of us lack. Do you know of any resources which can help in learning about the business side of the technology?

Start here: https://www.andykessler.com/andy_kessler/excerpts/How_We_Got...

Message me after. I am at wmarchibald@gmail.com

EDIT: Oh! Also "Eat People" by same author - Kessler : https://www.amazon.com/dp/B0049U4KC6/

Semi-edited: Everybody else will say Zero to whatever, and the Hard things about Hard Things, and Jimmie's book Good to Great and Crossing the Chasm, and you might call me mad (and I may be!), I am the only person who will recommend a book called "Eat People". (It is a good book, though)

EDIT2: U - akjha - Please let me know what you thought. You have my email address.

Please let me know how those two resources treated you and we can go from there. (Part of it will require that I know where you want you career to go)

Hey! I'm David, the editor of the post you linked to. If there are any other topics you're interested in learning about, please let me know! We wrote another one on Salesforce (https://retool.com/blog/salesforce-for-engineers/) that was well received on HN too (https://news.ycombinator.com/item?id=20277115). We were thinking about writing about Marketo / marketing automation solutions next. (I'm david AT retool.com)

(As an aside — when we started Retool, we ourselves didn't expect the "custom internal tools" to be such a big market. As it turns out, we now have customers that spend $400M+ / year on building custom, homegrown, internal tools! While we [or any single vendor] certainly can't capture all of it, our goal is to capture some percent of it.)

I worked at NetSuite for a little over 4.5 years (2 years before the Oracle acquisition), and it also blew my mind how much companies were spending on NetSuite customizations. It's a huge world out there for sure.

Great article David! Thanks for taking the time to include some screenshots and videos. This is the kind of company-funded-blog-content that we all love to see.

Where is most of your adoption coming from? Which verticals? Which horizontal groups?

Did you ever look at pega, nintex and Appian?

You really have to be living in an alternate universe if you haven't heard about SAP. To me it feels like every other job involves some work with it.

SAP was just nominated (together with Deutsche Telekom) the contractor for the official German Vivid tracking app (Covid-warn-App), just saying

I really don't get why they did not award an agency with a proven track record of delivering apps. If you choose one of the big ones, you still have high rates, but at least there is a chance of actually getting it done.

SAP is known for incredibly long ongoing projects and Deutsche Telekom has a track record of destroying acquired online businesses. I would bet 10:1, that this project will fail.

Probably because SAP is a huge political force in Germany more than anything.

I work in B2B software and almost all of our clients use SAP. It's absolutely massive.

My company is using SAP so it is very familiar to me. But before I joined my company, I have never heard of SAP. There are some good industry process knowledge that gives them an advantage and the software is good once you get used to it. But it can definitely be 10X better (there is a big opportunity to make it easier to customize or integrate with other software) . However, once you become their customer, the switching cost is very high even though there is better software in the market. That makes it difficult for startup to break into this market.

most people would be similarly surprised that $1T-market-cap microsoft is a business & enterprise software company, rather than a consumer one. it's ~80-90% of their $130B in yearly revenue (hard to get more precise because of the way they report their numbers).

That's not surprising to me personally, but I can see how it might be to the average person

yes, but a decent chunk of it competes directly against SAP.

> ..78% of the world’s food all flow through SAP?

Having worked in food industry as IT about 20y ago, we used to have fragmented system portfolio, and were struggling to pull data from one system to another. The around 2002 management decided to switch to SAP. I remember working on the storage/capacity/CPU aspect of it (and mid-migration I left the company and completely missed on the SAP train in my career). That food company was providing milk, yogurt, cheese, juice, ice-cream to about 25mil people (customer base). That's SAP for you right there. So that 78% doesn't surprise me at all. I believe the remaining 23% is too small to invest that amount of $$$ nad just hasn't done it yet.

When you get the ability to "track every cucumber" from pick to sell, cut down on fraud, errors, be able to calculate every bit of profit, loss, why would you go anywhere else? I understand the initial cost may be hefty, but once you set it up (good and well), you can stop focusing on that, and see how to grow the business parts of your business, and not the admin.

See this sounds good, and is why many execs are sold on it.

But because it is so massive, it's impossible to find talent that can actually implement it properly. That's IF your company doesn't require a ridiculous amount of customization.

Projects always run over-budget and over-time. Vendor lock-in becomes a huge problem. These companies solely exist to keep customers using their software and paying their service fees. They then provide just enough marketing buzzwords to justify it to execs who justify it to other execs.

If SAP is done right, it is very powerful, but it is seldom done right.

> Oracle and SAP are so dominant that even Microsoft uses SAP instead of their own ERP offering, Microsoft Dynamics.

This is insane. If it's hard for a company like Microsoft to move away from SAP and shift to their own solution (hence saving face in addition to the $$$), I can't imagine anyone switches software in the ERP world

Not so mad, MS Dynamics GP and Nav don't play in the same league as SAP/Oracle. AX is closer, but not there.

I work for SAP and it's surprising to me that not even other engineers in the Bay Area have heard of the company despite its campus in Palo Alto and their sponsorship of the "SAP Center" in San Jose. Do you file expenses? You probably use SAP Concur. Do you order office materials or equipment? You may use SAP Ariba. Have you taken part in interviewing a candidate coworker, or have you done yearly career goal planning? You may have used SuccessFactors for that. And this doesn't even count SAP proper's offerings (HANA/S4, etc).

Concur, Ariba and SuccessFactors are acquisitions, and only SuccessFactors is fully integrated under the SAP Umbrella. Here in Prague Concur and Ariba have their standalone subsidiary companies, with Separate HQs in different parts of town.

In my opinion the civilization in question does not have steam-powered airships unfortunately, but it does have some really overpriced crap.

Believe it or not, there was business software in the 90s. They were called *RP. ERP = Enterprise Resource Planning, MRP = Material Resource Planning, CRM = Customer Relationship Management. Home-grown code hacks to run a business cannot do what these companies can do, that’s why they’re expensive.

It seems that the market is big enough to allow unicorns making "plugins" like https://en.m.wikipedia.org/wiki/Robotic_process_automation to emerge.

Thanks for sharing, very interesting to read.

> basic installation of SAP has 20,000 database tables, 3,000 of which are configuration tables. In those tables, there are ~8,000 configuration decisions you need before even getting started

Wow, this is nuts. 20,000 database tables?

They've been adding tables since 1980...

I did an internship at the SAP HQ in Germany more than a decade back.

The work was exceedingly boring. But they had excellent lunch. Better than most Google offices.

You work in IT and you've never heard of SAP?

That's pretty wild, like not knowing who Microsoft or IBM are.

I haven’t read of such things in my country but am hopeful that this will happen soon here in India.

TfL, BP, Shell etc run on SAP. Had good fun messing around with SAP over there

ExxonMobil runs on it as well.

I clicked on the link not having heard of ERP or SAP and expecting to have my mind blown, too.

But, isn't that software almost the reason why computers and databases and programming happened? To help businesses with there boring business stuff, like inventory and orders and accounting?

This is kind of how I felt when I started to learn Haskell

It's like Salesforce, that crap is ubiquitous in large companies. Have you ever used it? It's a nightmarishly clunky piece of trash and yet Salesforce is a 150B behemoth.

Why would you use ERP instead of a regular SQL database?

edit Got a lot of vote downs. This was a genuine question from someone completely ignorant on what ERP is. Not trying to say that SQL is better, but I'm asking this because I'm ignorant of what ERP is and the description makes it appear as if it's a database.

What's the difference? would be better wording I guess.

Try getting your average office worker to deal with an SQL database directly. No chance unless it only has a single table. A database without an interface tailored for a specific purpose is useless for office staff. I know this from experience having once tried to get office staff using pgAdmin to update the shiny new database I had designed for them.

What's funny is that this is what SQL was originally designed for. That's why it reads so much like English. Of course, CLIs were too.

Because it's your business logic. And your GUI builder. And your tax system. And your legal report generator. Etc.

Yes I've heard that often businesses mould themselves to the software, rather than the other way around. The argument is that decades of exposure means that, if not entirely optimal, it at least represents competent processes for running a very large and complicated business. Rolling your own multinational tax compliance in SQL would be madness.

Why would you use an ecommerce platform instead of a regular SQL database?

When you're doing eCommerce.

My point exactly.

I don't know about SAP in particular, but ERP solutions frequently use SQL on the backend. I work with Microsoft Dynamics and it's just a SQL database with a special front-end for constructing workflows. I suspect Oracle ERP would be similar and have Oracle on the backend as well.

Because the ERP solution comes with a gaggle of sales people, sales eng, and bizdev folks to convince your co's decision makers that it is best thing since butter.

SQL doesn't.

"SQL doesn't"... you are correct, SQL can't figure out how much tax you need to pay across 9 tax jurisdictions. SQL can't tell me the best way for my warehouse picking robots to move across the floor based on the layout of the warehouse and what others products are being picked right this second plus over the next few minutes. The list is really endless of what SAP will do and SQL won't. It would take decades to replicate what SAP does.

I mean I guess you can throw all that functionality in a bunch of triggers but that would slow down your database to a crawl every time one record updated.

Have the triggers publish to service broker!

Sounds like SAP does custom App building. Do they use a custom underlying database technology as well?

Would it be equivalent to contracting some software system to a a company that would build the system using SQL?

Reply on news.ycombinator.com