Completely integrated company IT infrastructure – The future or a daydream?

The age of big mainframe computers and synthesizers was in 80’s. Many things have changed since then, yet too many companies still invest in legacy systems which are heavy, complicated and require an army of IT staff to run them.

In the last few years we have seen new innovative solutions entering to productivity and sales tool space. This should be very welcomed by sales professionals, as the case is rarely so, that a single big system, such as salesforce or AX/Dynamics, can serve efficiently everybody, even if they seriously try to convince you so.

So, how we manage to operate and sync all specific tools needed by different units within organizations. We don’t. That’s why the question is, who does these tools serve and what is their real value to user and “observer”? In fact, do we need all this data automation, at all cost? Clearly, management does not need the same tools than field force and vice versa.


The problem in a nutshell

The problem arises when companies find out that the system they use, whether fully fletched CRM system or ERP, does not answer to the need of all different areas of daily operations. Naturally, any management looking for competitive advantage should become interested in adding the new “efficient and better” tool to their IT infrastructure.

So, adding a new tool is approved. But the what happens when there are several different applications which all contain some amount of overlapping information? It is not difficult to see that maintaining all of this “up to date” separately in all different systems, is a task which is not only quite impossible, but does not make any sense when it comes to time consumption. But integrations on the other hand bring their own problems. In the best case it exists by default and even works faultlessly. In many cases it doesn´t work that well and sometimes it must be built in order to get something really useful.

Not all companies using CRM and ERP are big and can’t simply afford time and money to get this done. Sure, the optimum solution would be that we sync this overlapping information automatically, so it is always the same and updated in all of these different systems. Those who have been dealing with integrations, knows that this is not without challenges.

Let´s think of a simple example: we want to integrate our customer overall information (just name, address, contact person, sector and other criteria, without any situational information) between two systems. What does it require? Well, let´s assume first that this information is up to date in system 1 and we want to get it synced into system 2. System 1 has an interface of functions that can be used to fetch this information and then it has to be decided what field in system 2 corresponds some field in system 1. When this is clear the data can be copied from system 1 to system 2.

So what we see here is that there has to be a component which acts kind of an interpreter between the two systems. Ok, whose “job” it is to provide this component then? Of course this works so, that smaller system providers have to provide these components to connect to bigger systems, because the leading CRM companies don´t see any reason to build connections to systems which are not really globally renowned or they do not contribute to their ecosystem in positive way. On the contrary these smaller “apps” might even be taking away some revenue from the incumbents.

Still, it is clear that even covering those two big ones there is many and many more systems still out there that we are not providing connection by default. We of course can build the connection to almost whatever system but it would be a customization project then, unless there are many companies desiring the connection to that particular system.


How to validate if integration is worth doing?

Here the real headache arises. The customer may see value in tool that answers to a specific area of sales but if it doesn´t integrate with their core system, does it make any sense? What is reasonable cost to buy this connection as a customization? How can you measure and reflect the cost in productivity or in saving of time so that it is worth the investment? Hardly any of us have a direct answer to these questions.

Here we have to go back to the basics and think first how it would work. Would the connection need to be two-directional or is one-directional enough. If one of the two systems is clearly the system where the data entry happens (new data is created) it may be enough that this data is created there and then transferred to the other system where it is further processed then. If something has to go to the opposite direction, then it has to be two-directional. It can be said that creating two-directional connection is at least twice as much of work than creating one-directional, many times even way more.

Company´s internal data process flow becomes vital and it has to be clear. You have to think what particular information you have to transfer and how often is enough. And this depends really about what information we are talking about and what are the systems, there is no general answer.

And now coming back to the original question. Yes, the integration of all systems in company IT infrastructure is going to be reality in the future but when interviewing different companies now it comes apparent that currently it is still a daydream. I believe we are just taking our first steps to that direction. The integrations between systems are already a business of it´s own, but it most probably will grow exponentially.

Before we are there, where all systems talk to each other seamlessly and effortlessly, we have to still rely to our own common sense. We have to

  • first identify what is the purpose and output of different systems in our IT infrastructure.
  • Then we have to identify which is the share of the data that is common to these different systems.
  • Then we have to identify if the direct integration is absolutely crucial or is the use case and the need something you can manage by transferring data through manual operations.
  • Finally if the answer is that we really need an integration, we have to clearly and specifically be able to demonstrate how it works in reality.

As the doing of the integration comes only fifth after these steps, maybe the discussion should not start from how you can integrate it, when presenting a new solution. I think the first thing is still, what do you achieve with it.

The Writer: Tuomas Härkönen, Director of Technology, THE RUDOLF Oy
Back to main page