SAP BW|BOBJ|Project Management Blog

COTS vs Custom Software Development - Decision Tree Analysis

Written by Lonnie D. Ayers, PMP | Fri, Aug, 06, 2021 @ 12:49 PM

An Advanced Decision Tree Analysis Calculator

Have you ever needed a way to make a decision as to whether you should buy Commercial Off-the-Shelf (known as COTS) software or develop your own Custom Software Solutions using the custom development approach?  As it happens, we are often faced with customers who are trying to make this COTS software vs custom developed software decision.  They often engage us to perform decision tree consulting, though they may not always call it that.  

 

Decision Tree Analysis - Not Always Easy to Use

 

In my role as an SAP Project Manager, I am often faced with this decision, as just about all COTS projects, SAP (COTS is known as Commercial, Off-the-Shelf software, of which SAP is one) or otherwise, have a procurement aspect to them that must be managed by the project manager

 

Decision Tree Analysis

 

One of the tools in the Project Managers toolkit is known as Decision Tree Analysis normally using some form of Decision Tree Analysis Software.   These are deceptively simple tools that have been around forever.  What they boil down to, at least in this context, which in my context is usually around SAP custom development or Hubspot Custom Development, for specific, unique business processes which requires bespoke software (meaning a custom solution), is providing a way to evaluate the 'value at risk' and providing a dollar based view of the risk adjusted value at risk.  

 

 

 

 

There are two major challenges to using this approach.  The first is getting anybody to agree on the risk figures you want to use.  The second is getting anybody to agree on the expected payoff from the project.  Which is a problem, because technically, when you start your SAP project, one of the very first deliverables is the business case, which is expected to contain the expected project payoff.  Of course, the exercise to get to that number can be very challenging to pull off, in both time and money.  

 

In my experience, most customers don't actually 'accept' the numbers developed by the consultancy, but if there is a number available, you're ahead of a lot of projects.  However, I was looking for a 'shortcut' I could use to overcome this challenge, a sort of 'rules of thumb' approach I could use to come up with some sort of sanity check.

 

Ask Dr. Google

I was looking for a way to come up with a quick estimate of what any given project's payoff might be.  What I found was that a shortcut is to use X dollars of ROI for every dollar of project cost.  Thus, for an ERP project, though there are literally hundreds of different numbers one can quickly find out there, a 7 to 1 project ROI popped up frequently.  However, I was looking for a more generalized approach to use 'in the heat of the battle'.  Looking at the 7 to 1 number, it roughly translates into Cost * ROI.

 

Using such a formulaic approach means you can explore different assumptions.  Now the math part is simple, and it comes down to getting everybody to agree on the assumption.

 

COTS vs Custom Development

Taking a wild leap of faith (but based on valid experience), if you make the assumption the transformation project under evaluation, whether a COTS or Custom Development approach, results in the same ROI, you can then use the same basic formula to arrive at a top level value at risk number.  

 

Risk Adjusted Value at Risk

 

In our example Decision Tree, we have assigned a probability of failure to both the COTS approach and the Custom Development approach.  The working assumption is that the COTS approach has a lower risk of failure than the custom development approach.  Of course, this assumption could, has and will again, kick off all kinds of 'discussions' among customers, vendors, and developers, all of whom have strongly held 'points of view'.

 

Making Rapid Make vs Buy Decisions

In order to help my customers with this very difficult process, I decided to make an online calculator that allows you to play with your assumptions and have immediate feedback.

 

The breakthrough with this advanced Decision Tree calculator is that it:

  • Allows you to explore various scenarios
    • Explore whatever assumption you want to make about project ROI
    • Typical issues to consider but often overlooked:
      • Risk of vendor lock in.
      • On-going software and systems maintenance cost.
      • On-going required Software Updates for both the COTS and Custom Solution.
      • Knock-on effects and cost on core systems.  This can be especially problematic.
      • Integration cost and effort with your existing systems.
      • Which approach provides the best competitive advantage.
  • Allows you to set the cost equal to either decision - Make vs Buy - and see what the Probability Weighted Loss Function is.  

Would be glad to walk you through how I came up with this.

 

 

 

 

 

 

 

What Sort of Questions Can You Answer With This?

Of course, no evaluation is ever a simple matter of asking a single question and that's it.  With this calculator you can ask and answer the following questions:

 

  • What is the minimum ROI you must make from the COTS approach for it to make more financial sense than developing your solution in house.
  • You can play around with your loss probabilities to find out at what level of risk changes your outcome.  Perhaps you feel your inhouse development capabilities are good enough to overcome the inherent advantages of the COTS solution.
  • You can combine this analysis with our business case builder tool.  This tool not only provides you a quick way to develop a business case based on equivalent sales, it provides you with an extremely advanced Profit Proposal calculator so you can take this analysis much deeper.