»
S
I
D
E
B
A
R
«
To Integrate Or Not To Integrate?
March 13th, 2008 by Anand Sethupathy

Almost every time we encounter an organization with multiple applications running their operations, we always hear the same request for a fully integrated real time data solution with a CRM solution acting as the glue holding everything together. While this is technically possible, we rarely end up delivering this type of solution after going through a cost-benefit analysis with the organization. Below is some of what we consider and cover in that decision making process:

Step 1: Determine the Value Proposition of Integration

The first and most important consideration is to determine the real value that this integration will bring to your organization. In most cases we find that organizations are already generating integrated views of information even if they come from separate applications. This is often done with some manual analysis using Excel and other tools. The real value of a systems integration project can be measured in the following ways:

  • Costs Savings – How much are you saving in labor costs by not making your employees do the extensive manual analysis to reconcile data from multiple applications?
  • Time Savings – What benefit does having this data in real-time or near real-time provide to your organization?
  • Accuracy – Does systems integration result in more accurate reporting? If so, how much in savings are you achieving through better information.

Step 2: Evaluate the Integration Approach Alternatives

Generally speaking there are three integration approaches we often discuss with organizations.

  • Manual Integration – Allows you to synchronize data back and forth between systems, however, it is driven by a manual process. This is typically done on a weekly, bi-weekly, or monthly basis often with several tedious steps.
  • Uni-Directional Integration – In this situation, we make a determination that one Application will be the primary and another one will be the secondary. The Primary application always sends data to the secondary application(s), however, it never receives any data back. Uni-directional integration can either be real-time or conducted in batches (such as nightly updates).
  • Bi-Directional Integration – This is the most complex type of integration and also the most often requested. In this situation applications transact data (often in real-time) in both directions. For this to work effectively rules must be in place to resolve conflicts. A conflict is a situation in which both applications attempt to update the same information at roughly the same time. In this situation, we need to be able to establish an order of precedence as well as a transaction log so we can correct any mistakes.

Step 3:  Perform an ROI Analysis of the Alternatives

For each of the integration alternatives consider the Return on Investment (ROI). While you can consider an investment over a longer time horizon most IT investments are often evaluated on a 3 to 5 year time horizon.

 

 

 

The above chart is meant to be a hypothetical analysis of cost/benefit and ROI for an organization. Actual costs will depend on the complexity of your existing applications and the integration solutions you decide to go with. The above chart is also not meant to be an endorsement of Manual Integration. For many organizations this might be the right approach, however, that determination can only be made after a careful analysis of your integration options and your internal cost structures.  


2 Responses  
Evan Callahan writes:
March 13th, 2008 at 11:40 pm

Thanks, Anand! I have always simply asked clients to take me at my word: “Trust me,” I say. “You don’t want to integrate that.”

Even though I doubt you can ever accurately predict and quantify the efforts involved, this kind of analysis will be helpful as a way to give some relative weight to the choices.

Norman Reiss writes:
April 15th, 2008 at 6:43 pm

Interesting perspective on a difficult question. My current organization has struggled with integration for quite a while, yet the criteria ‘how well does this product integrate with our other products’ never seems to be part of how we make purchase decisions. Maybe it is worthwhile to accept that sometimes it isn’t worth it to try to have one database.

Leave a Reply

»  Substance: WordPress   »  Style: Ahren Ahimsa