This last week I was at a customer site in South Carolina. The customer is a brand name manufacturer (M1) that produces a well known product (could be the market leader, but I am not sure) that is used across the world, and this plant produces said product for the US. While its not a physically imposing plant it does produce on order of 1.8 _Billion_ units a year - which is not too shabby for a product that is about the size of a little finger.

Recently M1 was sold by its multi-national parent company (P1) to another multi-national company (P2). Now P1 and P2 were of course on different IP networks and company policy for P2 was that the manufacturing plant bring the IP addresses of all equipment (computers, PLCs, etc) into alignment with P2's network. So late last year a project was set up to change over all of the IP addresses during December when they were having one of their maintenance shutdowns. Money was approved, two contractors I know were booked to do the work and things were looking good.

In order to save money, P2 has contracted out the maintenance of its entire global network system (voice and data) to British Telecom (BT). That means that while P2 and M1 own and control all of the equipment that hooks into the network, BT controls the actual network infrastructure such as switches. And of course there a number of switches scattered through M1's plant that would need to be reconfigured as part of the IP re-addressing. And this is where the WTF starts - The Costa Rican division of BT is in control of M1's network infrastructure, and no one could either speak to the right people or get BT organized to do the work in time. As a result the project had to be scrubbed at the last moment and the contractors were left hanging with no work for that period of December.

Now the project has to be re-worked so that they can figure how to re-assign parts of the network at a time while keeping the remaining part up and alive while the plant continues to pump out parts. Outsourcing can be cost effective but when it impacts on your core business there should be some big questions raised as to how applicable it is.

OzPeter:

While its not a physically imposing plant it does produce on order of 1.8 _Billion_ units a year - which is not too shabby for a product that is about the size of a little finger.

Sigarettes?

 

OzPeter:

Outsourcing can be cost effective but when it impacts on your core business there should be some big questions raised as to how applicable it is.

I'd think TRWTF would be if those machines couldn't work without a network, or can they?

OzPeter:

Outsourcing can be cost effective but when it impacts on your core business there should be some big questions raised as to how applicable it is.

I've personally only seen two situations in which outsourcing can really be cost effective:

• If the outsourcing company has no clue how to perform the work which they are outsourcing.  (In this situation, the effects will be random, as we're talking about people who also have no ability to determine who does know how to perform the work - if they could determine that, they would've hired competent people already.)
• If the company to which the work is outsourced hires its workers at substantially lower wages.  For example, call centers in Elbownia.  Note that Elbownians are frequently not known for their tech savvy.  Also note that those Elbownians who *are* tech savvy tend to cost a lot to hire, and so are usually avoided by companies trying to make the biggest buck in their outsourcing arrangements.

This is, of course, not a statement that outsourcing can't be cost effective in other situations; I just haven't seen an example of it.

dtech:

Sigarettes?

Actually not cigarettes .. you have to think rabbit with a drum .. but not that brand name. One of the sad things about seeing this product made is when you realize that pretty well every unit is eventually destined for landfill.

dtech:

I'd think TRWTF would be if those machines couldn't work without a network, or can they?

As a self contained machine they can work within the confines of their local network. But all information about production bubbles up through the network to the management systems and beyond. So no network means no production logging which means no reports saying that Fault X occurred Y times and caused Z downtime over the last (shift, week, month, year etc). And thats of the main reasons to have an integrated system is to report on things like that in order to spot trends that affect the bottom line.

As a system the plant is running fine at the moment. Its just that the IP addresses aren't in line with what company policy says they should be.

tgape:

I've personally only seen two situations in which outsourcing can really be cost effective:

I'm sure you haven't, but there are others. One big one is where the outsourcee has better economies of scale. If you want to send letters to 1000 of your customers, you're probably going to outsource a greater or lesser part of this unless you have a lot of bored sales staff and a relaxed time scale.

Similarly, a non-IT company which needs a piece of custom software shouldn't assemble the equipment and staff needed to create it when someone else already has staff, equipment and the source for a program which could be turned into something which is close enough to the software in the specification to be usable fairly quickly.

 two RWTF here:

1. Not changing the policy to keep the existing site in line.

2. Outsourcing anything to BT

_moz:

Similarly, a non-IT company which needs a piece of custom software shouldn't assemble the equipment and staff needed to create it when someone else already has staff, equipment and the source for a program which could be turned into something which is close enough to the software in the specification to be usable fairly quickly.

In theory you're right but in the company I work for they even screwed that one up. The IT department bought a license for some "close enough" software from a another company, got some of their consultants on a 6 month project to customize the application, only to discover half-way through the project that the software wasn't as customizable as promised and couldn't deliver the required performance.

And now they have blown the remainder of the budget on work-arounds and remedies for the failing application, they are stuck with an application that no one wants: nor the IT staff ("not customized enough for compliance with internal policies") or the consultants ("too customized well-beyond the original scope of the application") want to support it. The business people are not happy with the functionality and performance. And none of the clients who tested the beta decided to go with it so they are not making any money of it.

Sadly, I had worked with this IT project lead before on assessing software and he wasn't afraid to reject bad applications. I guess he was pressured by management to go with the initially cheaper outsourcing.

 

bjolling:

...The IT department bought a license for some "close enough" software from a another company, got some of their consultants on a 6 month project to customize the application, only to discover half-way through the project that the software wasn't as customizable as promised and couldn't deliver the required performance.

 

 How does it take 3 months to figure out that the software sucks ? No wait... you mean it took the _management_ 3 months to figure out it didn't work, then they go along for the rest of the time anyway.

Ren:

bjolling:

...The IT department bought a license for some "close enough" software from a another company, got some of their consultants on a 6 month project to customize the application, only to discover half-way through the project that the software wasn't as customizable as promised and couldn't deliver the required performance.

 

 How does it take 3 months to figure out that the software sucks ? No wait... you mean it took the _management_ 3 months to figure out it didn't work, then they go along for the rest of the time anyway.

I was the technical test lead at the time and management didn't care too much about the application until we started reporting on the performance issues. It took the system 8 hours to process 250k trades for one B2B relation, but management was hoping to process about 10 times these volumes. So they took these numbers back to the consultants who were reluctant to do anything about it.

We, the testers, actually created indexes and optimized the SQL statements improving the performance by a factor 4. All testers in the team including myself actually are consultants from a company that is specialized in .NET development, so we wanted to showcase the abilities of our people.

We created a migration plan to move this application to something we would custom build. Everything went fine until my supervisor wanted to prove to the client that we had an extremely good vision for the future and gave away our recommendations for free. The client of course took these recommendations and presented them to the other company who promptly promised that could do the same. Of course they never did and in the end the budget dried up and the project stopped.

There were some many WTF's going on at so many levels I think this story would deserve its own thread.

_moz:

tgape:

I've personally only seen two situations in which outsourcing can really be cost effective:

Similarly, a non-IT company which needs a piece of custom software shouldn't assemble the equipment and staff needed to create it when someone else already has staff, equipment and the source for a program which could be turned into something which is close enough to the software in the specification to be usable fairly quickly.

 

This sounds like a blanket recommendation.  It helps not to ignore context:

• Is the function of the software fundamental to the business itself?
• Are there strict requirements on response/turnaround times?
• What is the difference in magnitude between the outsourced price and a reasonable estimate for in-house development?
• Are your outsourcing options geographically proximate?  Can you make in-person visits if necessary?
  
• Are there any other aspects (i.e. language) which could impede communication?
• Do these shops have experience developing the specific kind of software required, or even experience with your overall industry?
• Do you plan to outsource just the coding?  The design?  The specifications?  The initial requirements gathering?
• Will you need extensive documentation, and if so, is that part of the contract?
• Will you need interoperability with other systems, and if so, is that part of the contract?
• Are you confident in the operational and financial security of the potential partners?
• Are you confident in your ability to recover if the project does fail and you have to start over from scratch?

 

Simply being a non-IT company does not eliminate the need for a competent IT and development staff, especially when custom software is involved.  Outsourcing may be a good option, depending on the nature and position of your company, the nature and importance of the IT work, the local conditions, and various other factors; however, automatically assuming that it IS a good option for no other reason than "it's not what our business is about" is folly.  Do your research.