When does BI make sense over an end to end ERP solution?

The objective of a BI solution is to have actionable information accessible to appropriate employees. To achieve this objective, the Top 10 things to consider are:

1) Ease of access
2) Ease of use
3) Ease of creating ad-hoc reports/graphs
4) Cost of implementing
5) Cost/Resources of maintaining integration as the applications are upgraded or modified
6) Is mobility a requirement (i.e. web accessible)?
7) Time to deployment
8) Is real-time information a requirement?
9) Seamlessness and automation of Integration
10) Security

The order of importance of the above list really needs to be determined by each company as the importance will vary according to each company's requirements. Once the above items are analyzed, the answer for your particular situation will probably flush itself out.

Having said the above, I am a firm believer in having all your key information to be housed in the application where people are able to accomplish their job most efficiently and cost effectively. Having a solid fully integrated ERP system should really be the foundation for your accounting/finance and manufacturing (MRP) needs as this is where a company can really leverage the power/efficiency of an integrated system.

Other functions and workflows such as CRM and Quality Management don't necessarily have to be accomplished in the ERP as there are companies completely focused on these applications and have extremely powerful targeted solutions. Examples are salesforce.com (CRM) and AssurX (Quality Management). However, the key is to have the ERP and these other applications to be automated and seamless integrated such that all appropriate employees have timely and accurate information in the applications where they perform their specific jobs.

Once the source and location of all the key information is determined (both short term and long term) only then can the BI strategy be finalized. As a final note, the ERP really needs to be the single source of truth for financial and manufacturing information as, by definition, it is the official book record for the company.

To me Martin nails it at the end. If the efficiency gains from an MRP does not motivate the extra costs then you would have to motivate it some other way.

Ofcourse, above the same argument could be extended and generalized: we should pick that which offers the highest efficiency gains per cost unit. And in real life that choice could be (and surprisingly often actually is) staying put and doing nothing or at least waiting.

Usually, a fully integrated solution is more efficient. I would look at moving to this kind of situation before implementing BI. Otherwise, you are going to continue entering information multiple times and trying to keep them in synch, by doing reconciliations etc. You are never really going to sure if the information is accurate. If you implement BI on top of this, it doesn't really solve the underlying problem. Of course, you would have to do a detailed comparison and look at priorities, cost etc.

I think that a BI solution is always needed, independently of what applications you're using.

Here are a few basic - practical questions I would ask myself

(1) What questions/decisions are you trying to make? If the questions/decisions are more tactical in nature then I would focus on the existing ERP solution. If the questions/decisions are more strategic (multiple period trending, corporate insight, etc) then I would focus more on a BI solution. I see BI as a maturity in enterprise reporting - if I cannot answer the basic tactical questions then BI would provide me limited success.

(2) Does the ERP package support the entire business process. If you need to grab information from multiple sources then BI may be the viable option.

(3) Most important is to have a BI strategy before even looking at possible vendors. My rule of thumb is - do tactical reporting in source system and strategic, analytical analysis in BI.

Hope it helps!
Brett

Great answers from everyone.

I have worked with an organization that had separate Finance and Manufacturing systems. There was no correlation between the Financial data and the Manufacturing data, even though there was a tremendous effort put into trying to align them. They could not rely on the inventory $ matching the inventory quantities. The best solution was to wean them off their existing separate systems and implement an integrated ERP system.

As stated above, BI systems address analytical, strategic needs. Certainly not a replacement for an integrated ERP system. While it may be tempting to try to bring disparate systems together by means of a BI system, it is very difficult to achieve. Somehow, you have to clean the data as you copy it into the BI database. Otherwise, the resulting analytical reports will be misleading and drive the wrong decisions.

Another factor to consider is that every time an upgrade is applied to either the Financial or Manufacturing system, you have to revisit the connections to the BI system to make sure you are still extracting the correct data. With 2 separate systems, you have 2 sets of connections to maintain.

My recommendation is to replace the Financial and Manufacturing systems with an integrated ERP system. If you still need the analytical functions provided by a BI system select one for that purpose, not to try to pull 2 disparate systems together.