Part 1 of this blog series highlighted the comparison of best-of-breed (BoB) and enterprise resource planning (ERP) solutions, and when BoB is the best choice for supply chains. Part 2 of the series dove deeper into the specific needs of supply chain planning and execution and the tightly vs. loosely coupled interoperability debate. In this part of the series you’ll learn more about where to run your BoB software solution.

You’ve Finally Selected Your BoB Software Solution. Now What?

Whew, you’ve evaluated a dozen options, you’ve finally selected your BoB software solution and the implementation partner to get it setup.  You’re done, right?  Well, not quite, you still need to figure out how to run it!

In days gone by, the next step would be turn the project over to your IT department to host and manage your software and wait for your turn in the queue both for capital appropriations, purchasing and then resources. Those days are past, and there are now multiple ways to get the final step done and they deserve some careful consideration.

You’ve picked BoB software, so I recommend you look closely at choosing a BoB managed cloud offering.

Deploying and Managing Internally

It is still an option at most companies to follow the traditional IT path. There are certainly many IT departments well equipped to purchase and setup the software.  They will participate in the project and accept the ongoing maintenance requirements. This can still be an effective and successful process for go live. Day one your system is providing ROI.

The long term is where it gets tricky.  Modern IT departments are facing year over year budget cuts or at best they remain flat – which is a cut with real inflation. It is not uncommon to find an internal IT group spending 80 percent or more of their budget simply to maintain what they have. Since they support many strategic and critical systems, it is rare to be able to assign full time resources to any one product and almost as impossible to expect those resources assigned to be able to keep up with the changes in the software. Requests that can bring real results get into the IT queue, where it is not uncommon to wait months, or even years. Upgrades almost require an act of congress to happen. And as people shift jobs and even leave the company, critical information leaves with them.

If you face this challenge, the result is often a system that peaked in value the day it went live.  Changes in the business, as well as changes in the software, just don’t keep up.  Processes get set in stone, even processes that no longer make sense in the current business model. It is not unusual for me to talk with a company who has 5+ year installs and is too afraid to change anything because they don’t remember the original reason it was requested.

The other tricky part is it is very difficult to calculate the actual cost of running the software.  Running internally is generally going to be the least expensive option on paper and in reality, but you really need to dig in to discover this. Since so many items are blended together with other areas, you get an allocation based on something, such as hardware invested, at best. Given the complexity differences between products, this tends to move everyone to the mean, so it really is difficult to compare. When looking at costs, you really need to look at skills that need to be maintained, like database administrators (DBAs) and application experts, how you can keep those resources content and how their knowledge can be maintained when they leave. And while it is difficult to judge some opportunity costs, they are very real. Not changing and upgrading in a reasonable time period is leaving real money on the table. Staff attrition means an investment to get skills back to acceptable levels. Opportunity costs are hard to quantify, but they should absolutely be part of the equation.

Managing Internally: Deploying on a Private/Public Cloud

There are many cloud options available today, such as public, private and hybrid. When deciding whether to host it internally or externally, but manage it internally, you will need to look at these options and determine what will work best.

There are positives to these options, especially if your software supports automatic elasticity.  You can add capacity for peak periods with a simple change. Your options for compute, input/output (I/O), backup, disaster recovery and capacity are very broad. Availability from the largest vendors is generally very good, if their definition of availability fits your definition (these are worth examining closely).  Price per unit of capacity is generally very competitive and continuing to drop as competition grows and long term storage of large amounts of historical data is especially cost competitive.

However, there are some “gotcha’s” to watch out for.  If you need always on capacity, it is not always clearly less expensive than other options.  Calculating your actual fees is difficult, as you are charged on multiple dimensions, such as network, storage, compute, I/O patterns, etc.  It is generally not a predictable bill and unpleasant surprises where utilization and costs shoot up are typically found out after the fact, when you already owe the vendor.  You have a lot of choices on how to set things up – meaning you need to choose.  This is an area we here at JDA have seen consistently under estimated.  Working with a cloud vendor ends up teaching you a lot about how to really get what you want done, whether it is implementing a data retention policy or correctly choosing the right configuration.  You want to be careful that the lesson is not a painful and expensive one.  Lastly, make sure that the vendors resolution policy for situations other than a complete down are sufficient for your business.  Most cloud vendors do not have any commitment time to fix something like a system being unusably slow or intermittent network drops.  And your ability to look into the layers under the surface is very limited – so you need their help.

It goes without saying that all the other questions around skillsets regarding the software, database, etc. still apply.  The cloud option still requires your IT department to be involved, though less resources are needed on a day to day basis.  Your bill may fluctuate from month to month which can lead to some unpleasant surprises.  And, while cost for hardware is lower, I strongly recommend you look at how much manpower it will take to get those cost savings.  I try to never trade hours of labor for cheaper hardware.  Labor is always going up in cost and a one-time purchase of infrastructure amortized over 5 years is often a lot less than non-IT folks expect.

Part 4 of this blog series will dive deeper into detail around deploying on a BoB software vendor’s managed cloud and making the choice for a BoB managed cloud.

How JDA Can Help with Managed Cloud Offerings

JDA understands the importance of focusing on business results, meeting project deadlines and a managed cloud offering to support them. That’s why JDA has invested in developing a robust, versatile and high-performance integration capability that not only provides an accelerated time-to-value through pre-built interfaces and easy-to-use configuration tools, but also establishes a high degree of flexibility and adaptability, permitting support and evolution of this integration over the long term. To learn more about how JDA can help, contact us today.