ABCtech #115 12APR18 Integrating Information Systems Always Seem to Fail

ABCtech #115 12-APR-18 Integrating Information Systems Always Seem to Fail

From time to time we receive an article that reads as an infomercial, in which case it is presented as an advertisement. However, the message and insight that Blair Kjenner of Method1 Enterprise Software has shared is important for the information technology industry and those who rely on it.

The infotech industry faces continuing criticism for budget over-runs, late delivery, and failed systems.  Why? One of the essential issues is the incompatibility of the organization’s individual information systems that are almost impossible to integrate.  Who is to blame? There are too many to mention.  The answer? Set and enforce system standards from the beginning starting with procurement. – Editor

Blair Kjenner

The approach to creating information technology systems for organizations is broken and it is deeply affecting the efficiency of organizations.

Typically organizations apply a proprietary approach where all systems have been created in silos.  When a system is required, teams design a user interface, create an application framework to support it and then automate all the business functions that are within scope.  In the process, a database is created that it unlike any database that has ever been created.

Then the application is thrown into the mix of applications used within an organization.  Organizations attempt to integrate it, only to figure out there is a significant amount of common data that is duplicated across all the systems.

The reality of updating data in one system and having it automatically update every other system is overwhelmingly complex, particularly when you consider how much common data is being duplicated.

Issues with integration are directly related to how we approach developing systems.

Systems are developed in silos and then try to integrate them!

Many organizations attempt to address integration issues by implementing an ERP solution from a single vendor.

However, implementing an ERP system has its own challenges, like:

• It can be prohibitively expensive and risky to implement
• It makes an organization overly dependent on a single provider with little ability to change
• Organizations get caught in a dilemma where if they change the system to meet their specific needs then they are precluded from future updates but it they don’t they are no better or different than their competitors

Furthermore, for large organizations it may not be able to implement a single ERP system to address all its needs.

For example, a state/provincial/federal government will not likely be able to address the needs of all its departments and related entities in a single ERP system.

These organizations end up implementing many ERP systems which results in challenges attempting to integrate common data across all departments and related entities.

So why does there seem to be no good solution?

It goes back to the original issue.  The current approach to systems is based on creating silo systems.

To put this in perspective, imagine if accountants took the same approach to tracking an organizations finances.  If that was the case, each accountant would come up with their own unique method for tracking finances.  Not only would that result in a tremendous duplication of effort, it would also make it impossible to compare finances from one organization to the next.

This is precisely the situation with creating software for organizations and until it changes, we will always struggle with creating modular software that can click together in an integrated way to address the needs of organizations.

Why integration is so difficult? 

Imagine if an Oil and Gas company took ten people from different functional areas throughout the organization and had them independently create a spreadsheet with all the wells they dealt with. 

This exercise would undoubtedly end up in ten spreadsheets that looked very different.  Each spreadsheet would record the attributes about a well that were important to a given person. 

Now imagine, setting it up so that if a well was added to one spreadsheet it would be automatically be added to all the others. 

This would be challenging because everyone has different attributes and may possibly identify their wells using different methods.  Furthermore, if multiple people made an update to the same well which update would be accepted by the other systems?

Well data is just one type of data within an organization.  Organizations can have hundreds or even thousands of types of data across their systems and as much as 40% of that data can be common across all the systems.  Attempting to synchronize all the data is an unsurmountable task so organizations resort to rekeying data between systems and heavily relying on spreadsheets.

For learning more about the challenges with how the IT Industry approaches systems, stop gap measures we are taking to address it and one perspective on a future solution, please refer to the following link which contains a full whitepaper on the subject.