Wednesday, August 15, 2012

To Integrate or to Integrate Badly

When someone chooses to integrate there is always the wrong way and the right way to go about this. Just because you chose to integrate does not mean you did a fantastic job about it. 
One such example I came across recently was a nameless organisation decided to integrate their systems, they provided a very good level of integration and it was all working fine, then they got over zealous, a contractor they hired thought hey I can do integration also, the company not knowing any wiser said OK sure let’s do it. 

The integration was a simple one, I’d like to change my surname, and it should update all the systems, because I am married now and have taken my husband’s name by choice. 

The old way of doing this was very manual and very cumbersome and took a long time to implement.
We already provided integration of other user profile details, which were updated in the various systems in the organisation. The contractor thought hey I can do it too… 

The surname request update had a fancy front end, when the user submitted this, their request was sent as an email to someone in HR, the first problem. 
They would manually make the changes required in the system… The second problem.
What is worse, is that the email got it wrong, it said update my first name to this name, and not my surname, so the person in HR went and updated their first name to their new surname…  This person was outsourced from the organisation. The third problem.

This had flow on effects, that had not been thought of, when someone changes their name, their AD name needs to update, their email address needs to change, and a whole bunch of other systems need to be informed. 

To get around this they just deleted the user and made a new one as if they had just joined the organisation, the flow on was very very bad… Major problem goes with out saying.

All of their access gone, all of their email gone, their permissions to all the internal systems gone, this would have happened even if they had changed their surname. 

The loss of productivity for this person was about two weeks’ worth, the person was not a low level admin, they were senior and on a good rate. 

Take this against the decision to use a contractor who said yeah we can do it,  probably $1000 for the efforts, then add the down time of the one employee, $20000 in loss of productivity and in ability to work, you are not only paying for them to work but they can’t work and produce income for the organisation. 

Now multiply this by a low estimate of 5 people who may have been effected by this across an organisation of 7000 people…  

You have a net loss to the organisation of: $100,000  plus your low cost of implementing this in the first place.  

Take this and compare against doing it correctly in the first place, taking the time to understand the impact, going and updating the various systems, in near real time, and ensuring that the user was good to go within 15 minutes. 

A cost estimate for this would be somewhere between  $20,000 to $60,000. Had more than 5 people requested this change, think of the cost to the organisation.

It becomes very clear that you need to integrate, and you need to do it properly. You need to have the right people working with you who understand this and are able to see these issues well before they happen. 

Never believe someone who says yeah we can do that, no problem, it is easy, we will just email someone, or we will add a manual step, or we will just leverage the existing process or method, or we will just delete them and put them back… these things make me cringe… don’t do it… ever! 

Talk to the professionals, they have done it all before. 

No comments: