> My guess is that the “back-end” supply chain / software / inventory management problems are just not as sexy as the “front-end” consumer-facing problems, or maybe the consumer-facing problems are just more intuitive.
Absolutely! I wrote MRP systems for a pharma manufacturing in Florida and the kind of stuff I had to do on a daily basis far surpassed the complexity of your breathtakingly-beautiful but typical project-management webapp or customized T-shirt webstore. Let me be clear, complexity has absolutely nothing to do with the merit of one product vs. another. KhanAcademy code could be simple as 2nd grade math but nothing I ever write will ever be as beneficial to the world. However, complexity is expensive, takes time and dedication, and rarely pays off in the short-term.
There is a reason most supply chain software installations run in the millions to tens of millions. Here's an example of something I wrote while ago: A drag & drop scheduler in JS, kinda like Google Calendar that lets you schedule production jobs on different equipment, across different labor teams. When you change a single job on the schedule, it auto-calculates the entire requirement for the entire company. Moving one production job up (say shampoo for customer A) could end up in the company losing $500k because one of the ingredients that went into shampoo for A also goes into conditioner for customer B. This particular item has a lead time of 3m from China. And since order for B is significantly larger in amount, every single day of delay is money actually lost because you bought all the other raw materials for B on credit from the bank and now have to pay interest on it, while it just sits in the warehouse waiting for the raw material to be flown in from China.
Of course, this is something you want to avoid in the planning stage itself. And that's what the software does and warns the user within seconds of making any changes. One tiny bug in the code, say it doesn't correctly factor in the internal lead time from QC (this particular chemical needs to be sampled for microbiological contamination) and you just delay the project by a week. Putting up a pretty website is hard work but it is nothing compared to hiring 20 people to spend 3 months mapping out the multi-stage routings for 500 different SKUs. No tech-VC wants to invest in businesses that require a tremendous amount of operational labor. That's what banks are for.
Every time you see someone announcing a CRUD app written over the weekend you pretty much know that there is hardly any business logic involved.
The example you give is a great one, it only gets more interesting when it is not just money on the line but lives. Think avionics software and the like.
That stuff ranks amongst the most expensive software produced per line of code because of all the audits and certifications that try to ensure that there won't be hidden bugs that somehow make it into production.
Working on a project like that sounds like a lot of fun. I guess there must not be very much money in it though because every company requires custom software, eh?
*If you're good at direct enterprise sales... Most startup guys I know prefer PR, metrics, A/B testing, etc. over this type of meat and potatoes type of work.
Having done a LOT of work in that space (historically), I can assure you it's all SAP, Oracle, PeopleSoft, AS400 and epic quantities of turds. The largest turd being EDI and all sorts of nastyness.
It's not fun. It's actually painful. It takes a special kind of person to do it: Someone who doesn't care about what they do, just the money.
> It's not fun. It's actually painful. It takes a special kind of person to do it: Someone who doesn't care about what they do, just the money.
In large companies, absolutely. But for rapidly growing companies, this is one of the most challenging jobs out there and you really start to care about it on a personal level. I can make one change on the barcode scanners that saves the six warehouse guys over two hours a day. It actually improves their quality of work life by reducing potential for conflicts and blame. You won't get mentions on TechCrunch but you will help out individuals on a personal basis and the company on an efficiency basis.
Unfortunately the money is way better everywhere else for someone with skills to do this right.
Fair point - I'll give you that. Ironically I did a spot with Intermec barcode scanners and printers for asset management on top of Oracle so I understand where you are coming with.
I actually quit because I was the only one who did care. Everyone else was just after vendor backhanders.
Absolutely! I wrote MRP systems for a pharma manufacturing in Florida and the kind of stuff I had to do on a daily basis far surpassed the complexity of your breathtakingly-beautiful but typical project-management webapp or customized T-shirt webstore. Let me be clear, complexity has absolutely nothing to do with the merit of one product vs. another. KhanAcademy code could be simple as 2nd grade math but nothing I ever write will ever be as beneficial to the world. However, complexity is expensive, takes time and dedication, and rarely pays off in the short-term.
There is a reason most supply chain software installations run in the millions to tens of millions. Here's an example of something I wrote while ago: A drag & drop scheduler in JS, kinda like Google Calendar that lets you schedule production jobs on different equipment, across different labor teams. When you change a single job on the schedule, it auto-calculates the entire requirement for the entire company. Moving one production job up (say shampoo for customer A) could end up in the company losing $500k because one of the ingredients that went into shampoo for A also goes into conditioner for customer B. This particular item has a lead time of 3m from China. And since order for B is significantly larger in amount, every single day of delay is money actually lost because you bought all the other raw materials for B on credit from the bank and now have to pay interest on it, while it just sits in the warehouse waiting for the raw material to be flown in from China.
Of course, this is something you want to avoid in the planning stage itself. And that's what the software does and warns the user within seconds of making any changes. One tiny bug in the code, say it doesn't correctly factor in the internal lead time from QC (this particular chemical needs to be sampled for microbiological contamination) and you just delay the project by a week. Putting up a pretty website is hard work but it is nothing compared to hiring 20 people to spend 3 months mapping out the multi-stage routings for 500 different SKUs. No tech-VC wants to invest in businesses that require a tremendous amount of operational labor. That's what banks are for.