I recently consulted with a client who I hadn’t seen in a few years who is currently deploying a homegrown business intelligence (BI) solution. It occurred to me that when I had last talked with this client, the company was busy deploying an enterprise resource planning (ERP) system. At that time, this client was very optimistic, anticipating that the ERP system would satisfy all of the enterprise’s information and reporting requirements.
In reality, the ERP vendor’s promises of easy information access have turned out to be empty ones. My client’s organization is one among many that have not only found themselves no closer to information than they were before ERP, but who are also having to deploy a BI solution to close the gap.
In this article, I’ll examine the need for implementing a BI solution along with ERP, and I’ll discuss the pros and cons of building the BI solution either inside or outside the ERP system.
Why ERP alone doesn’t always cut it
Although it seems entirely plausible that ERP’s tight integration of the enterprise and the consolidation of its data should somehow resolve information and reporting problems, the fact is that ERP systems don’t give up information easily. They are incredibly complex transaction-processing systems. Due largely to the way their data is structured, transaction systems are great at getting data in but not so good at getting information out. (See one of my earlier articles, "Pros and cons of online transaction processing," for more information on this topic.)
In addition, the way most ERP systems are implemented forces you to rely on static reports to distribute information. You can always expect a high demand for specialized reports when static reporting is your principle mechanism for distributing information. Why? Because static reports are never quite right. Tweaks and customization are inevitable because individual information consumers require different data or a unique arrangement of data. Since you can’t drill down or drill across in a static report, you have to build another report to meet this custom requirement.
Just as in the days before ERP, all requests for special reports end up placing additional burden on your IT staff—especially since most ERP systems’ report engines require special know-how, making the generation of any special report expensive, complicated, and time consuming.
Unlocking ERP data
Amazingly, ERP vendors are now admitting to these system shortcomings, and they are telling customers that they will need to add a multidimensional BI adjunct in order for information consumers to be able to properly leverage the system’s transactional data and access information instantly and directly.
Most, if not all, ERP vendors are offering such multidimensional BI tools. As a CIO, you should know that you do have some choices when it comes time to make this decision. You can implement the BI solution within the ERP environment (which your vendor will lead you to believe is the only option), but you can also implement the solution outside, using third-party tools and technologies. Naturally, pros and cons exist for each choice, so let’s consider those more closely.
This is a possibility if your ERP vendor has either created a multidimensional BI environment within their ERP solution or has partnered with a third party to create such a solution.
- The ERP vendor knows where the ERP data is. This is no small task when considering the complexity of the data model.
- It is likely to work with the next ERP release without painful modification.
- You’ll likely have to covert any non-ERP data into ERP data to make it available to information consumers. This can be very expensive and time consuming.
- You’ll face more painful ERP system configuration and perhaps equally painful interaction with ERP implementation vendor(s). Often, these relationships are strained, owing to a difficult history.
- As many CIOs have already discovered, spending additional money on a system that was supposed to “solve everything” is an explosive topic in a lot of companies.
- The ERP vendor’s BI tools and environment may not be very useful.
- The systems are generally deployed with a proprietary/closed architecture, requiring ERP “black box” functions to get the job done. This translates to a single-vendor solution and precludes best of breed, both in terms of the server and client-side services and tools.
ERP data is stored in a database, after all, and you can extract that data for loading into independent data warehouses or marts. Organizations implementing BI in this way, using third-party tools and technologies, often feel they have the best opportunity to leverage the array of BI tools (especially client tools) to better meet their organization’s needs.
- Integration is easier when non-ERP data—like legacy data—is in the picture because you don’t have to move data into the ERP system’s environment.
- You have more options, including the ability to choose flexible, extensible “best of breed” tools and technologies.
- Certain data may be difficult to find and subsequently extract from the ERP environment, perhaps requiring your vendor to build data-extraction functionality. Some technologies—such as Acta, SAS, and Informix—purport to make this easier, however.
- You must either keep up with ERP data schema changes yourself or rely on your vendor or a third party to do so. Just because you’ve successfully mapped the ERP database, permitting you to extract data today, doesn’t mean you can extract the same data when the next ERP patch or version is installed.
It’s likely that every organization deploying ERP will at some point be compelled to implement a multidimensional BI solution. Today, ERP vendors are aggressively promoting such systems as adjunct tools, and they want you to believe that their tools are your only choice. But owing to a wide array of stout third-party products, you may stand the best chance of responding properly to your information consumers with an independent solution, implemented outside the ERP environment.
Has your enterprise deployed an ERP system? Did you find it necessary to complement it with a business intelligence system? Post a comment below and start a dialogue with your peers.