I am a internal web developer for a somewhat medium-sized media company. I built our enterprise intranet and am constantly building applications…products…that employees use and rely on daily. These tools are internal only and require such authorization and so on. Most of them were built from the ground up. Sure, I took various libraries and scripts from open source sources, but the application as a whole was built, not purchased. In a recent meeting with other internal “developers”, I found out that internal development is possibly rare…?
I am a one-man department when it comes to internal web applications. I’ve built many. I work for one IT department…more like a faction…within our enterprise, and the other faction has several developers. For the first time (in 5 years) I was in a meeting with most of these counterparts. The topic was version control, but that aside, I learned that they don’t build applications.
While discussing organizing code repositories, I asked why we don’t label folders “by product”. Someone quickly asked, “what do you mean, by product?”
That response intriqued me. Even though I don’t sell the programs I make to the public, I still consider them products because they were built for a specific purpose, customized. Regardless of the semantics of the conversation, I found out that most of them don’t build full-fledged applications, but build various middleware and parts-and-pieces for other, large, purchased systems.
Where as people come to me looking for a system and I build them one, these guys (and girls) build database interfaces and little add-on tools for many large systems that our company purcahsed (for a ton of money).
This context explained to me why he questioned the “product” wording. To them, they’re not building products. They purchased the product, they’re just making the product do more specific tasks by way of little add-on tools.
I’m not taking away each our skillsets to program and develop, but until that meeting, I was under the assumption that there were developers all over our enterprise making in-house software. Apparently that isn’t true.
This then got me thinking how rare is internal software development? I would like to think that most companies have internal software people making products for internal use, as it would save a ton of money.
We tried to develop IT products in house a while ago, but it’s not worth it.
It took to long “we need it yesterday”, was to expensive and wasn’t strong enough.
We will purchase a major piece of software (be it $50,000 or $5,000,000) and work with the vendors to custimize it to our spcific needs. It cuts cost, but most importantly time. We have money and need time and since TIME=MONEY it works out.
In addition to what Thee said… Liability is also to be considered (Why I never sold web hosting).
If everything is built in house, there is nowhere to point the finger when something breaks or self destructs or gets hacked.
When I worked in the IT department at the library, I was initially suprised that when it came time for a new server (for the catalog system database) we didn’t just build one ourselves. Instead, we got quotes from a number of different vendors, and didn’t even consider that we would build it ourselves.
The same thing went for simple file servers. My boss loved “network appliances”.
In both cases, the reason was the same… Sure, we can build it ourselves. While it might seem to cost less at first, there is a value to the support contract. We needed the ability to simply make a phone call and let somebody else be responsible for fixing the problem.
There’s frequently too much on the line to not be able to point the finger at a vendor and say “fix it”, and then tell the boss “they’re fixing it now”.
G+
Good points.
In my case, there isn’t money. Sure, we’re not clawing at the walls for money, but budgets are always getting cut every other month, department by department. The last thing each of them want to hear is a $100,000 price tag on software - they are more than willing to have it made inhouse. That’s not the real issue.
Time IS the key. However, I’ve watched our entire IT department work on a software launch that ends up taking several months just to install and then many more months of follow-up and fixing and customizing.
Sure, we can buy it and have someone else install it, but it doesn’t get installed overnight no matter who makes it - so I only half believe the time excuse. I can get a (big-ish) product done in 2 months, testing, training and all.
Inhouse software is also customized out of the gate. There isn’t any need to spend months after the install to have people make middleware and interfaces, it’s already been done - it’s in the design! Much like a tailored suit, it fits better, looks better, and works better than something you buy off the rack.
And, if you ask me, if a company gets in a “need it yesterday” situation, then they didn’t think far enough ahead to begin with. People are only in those situations if existing things fail and fail hard. Physical computers are one thing - they can put together in one day - but software systems are a different story because you can litteraly build them from scratch out of raw resources.
I will also challenge almost any support staff to a “fix off”. People mail or call me with an issue and I fix it while I’m on the phone. They hit ‘refresh’ and the problem is solved. Done deal. Last time I had to contact a vendor support person it was first a operator, then some dude who passed me off to another guy. That guy had to start a ticket that would get assigned to a tech and be done within the next 48-hours. Far from efficient.
Building products inhouse means the developers are right there all the time, hence the “in” part. If I had the knowledge to build a car, I would because then it would be free to fix, shy of parts. With code, there are no parts, just text, so the overhead is a lot lower. It’s like having a butler…why pay for a support contract to someone far away when the butler, whom you are already paying for, can do it and do it fast.
There is also the fact that for $5 million we can purchase software that cost $100 million to develop.
Right now we are getting ready to install a new base operations software that easily cost more than $100M and years and years to develop. There is no way we could duplicate this.
Well of course not. I’m not really thinking in terms of global fluxcapacitor transport systems. I’m talking department-level or even building-level products.
> “medium-sized media company”
I’m certainly not including gazillion dollar international companies in my thoughts. With that kind of stash they can do what they want.
The large company I worked for built everything in house. I think there was some paranoia on their part though. If something broke and someone from some other company had to fix it, that meant that they might see really sensitive information in the database.