Our company is considering a new ERP system. What are the benefits in choosing pre-written software over developing our own system in-house?

Simply the COST. That's cost in dollars, cost in time, effort, and opportunity cost.

Purchasing a pre-written product saves much of that labor by not having to re-invent the wheel.

Look at closed (and it sounds like you should look carefully at) open-sourced ERP ssytems.

Cost is certain a very important consideration. I agree that it would be impossible to invest the kind of money that would be necessary to program something that could compete with any of the commercially available ERP systems.

Outside of the initial development costs, the cost of maintenance and support is also a significant consideration. I don't believe you can match this from a cost basis either.

There are some other considerations as well, such as having a system that is SOX compliant if you happen to be a public company.

I will agree with the cost and support comments but I will add risk as a major concern. You can right a great ERP system that perfectly fits your business for a price. Now let us assume there is a major shift in technology, your business expands dramatically, your business model changes or any number of things. That ERP system will almost certainly not be flexible enough to handle it. You might have to start from scatch, but first you will hobble your business for a few years with bad patches and indecision before biting the bullet. You do not mention the size of your business or IT staff but another seemingly silly issue appears repeatedly with custom software. The person who wrote it leaves, used some great tricks to get you what you want and know one else can figure them out. This last one doesn't really seem to matter if its a SMB with the local "guy" or big-IT that used a big-C for help. The only difference is the price.

Big ERP systems are big, expensive and may not give you exactly what you want. Counter to that they are designed to do more things, grow and change. Many years ago I was working with a F100 company putting in an SAP system. The CIO told me the big consulting bills hit when they tried to customize the processes. Hindsight told him that that the canned processes they didn't like were better than many paid extra to changed to. Putting a new system in is a good opportnity to take a new look at your existing processes. Are they really the best approach?

I would concur that cost is a key driver but also flexibility with processes. If you are going with off-the-shelf products, you have to be willing to align your process to the technology. If you don't have the luxury of being flexible, you will end up customizing the technology and may cost you just as much as building one from scratch. The downside to in-house development is the investment of resources needed to maintain and evolve the technology to continuously meet the business needs.

Cost certainty is certainly a factor. You also know it works, but "configuration" will be required and is powerfull in most products. You may have to change some of your processes or methods to accomodate the way the system does things. Watch for trade-offs between configuration choices and operating complexiity. Configuration choices will affect implementation cost.

when I was working in the IT department for Dell, we traded off the costs and benefits of building our own systems versus buying. In the end, we took a hybrid of taking code we could modify and making it work for the Dell supply chain. It really paid off in terms of making the supply chain nimble and a compteitive advantage.

That said, it wasn't a cheap endeavor (the team for shipping was over 40 people). The key question I would recommend you need to answer is how many years you are committed to keep programmers and business analysts and project managers on your payroll. If it is less than 5 to 7 years, I would recommend you choose a very good software and implementation partner.

If you know of an ERP system that has functionality that matches your needs in a certain industry, I would most definitely go with a purchased ERP system. This assumes that you will not customize it. The last thing you need is to do is buy an ERP system and then customize it to death so everyone in your company feels it does what your old system did. I have seen a lot of money wasted in this process. When I had the chance to control it, we did no customiziation, which meant we kep it updated and lost no functionality over time.

I agree with Patrick that you need to research ERP systems being used in your industry to find the closest 'fit' for your business, but I believe you still need to make sure there is an option to customize the software, if needed. However, make sure you seriously evaluate the basic package and attempt to align your processes to match before making any changes. I worked as a business analyst for an IT firm for several years and had numerous clients who wasted so much money making the system match their processes. Funny thing was, most of those processes existed to fit their old software (which they hated), but it was comfortable and familiar so they wanted to duplicate it. I'd frequently end up asking why they bought new software if they wanted it to work exactly like the old one. That said, if you do have needs which are truly not addressed in the base system, it is good to have the option to customize.