The Big Bang BI Project – Is it necessary?

So your organization has seen spectacular reviews and demos of Tableau Software and are ready to go for a Business Intelligence (BI) project.  I write this article in order to suggest a way on how to approach implementing Tableau if your organization is implementing BI for the first time.

First, I want to describe what many consulting firms or IT teams will traditionally propose in a BI project.  IT teams or consulting firms will recommend to implement BI systems in its most complete sense.  This type of BI project typically involves the business gathering all the current and potential report requirements to be part of their BI project scope… Next, the customer realizes that they must take into account a certain data volume, usually based on x # of years with x # of Terabytes of data, coming from different source systems.  Now knowing that ultimately, users will want to consume this data in their Tableau Desktop on their personal device, an organization with big data will propose a middle layer that will automatically extract, house, consolidate, transform, and aggregate… this data from your various source systems.  This middle layer is commonly an independent data warehouse or database and is usually implemented before reaping the benefits of a self-service BI tool like Tableau.

It is this middle layer that usually demands anywhere from 1 to 6 or more months (depending on volume and complexity of data and reports) of IT related work. Typically in this BI approach, clients want the reports delivered by the IT consultants/staff and in the meantime continue on with their manual report generation.  Therefore, Tableau visualizations and reports are only utilized towards the end of the project – That is, to use Tableau between months 2-7 following the completion of the middle layer.

Well, I recommend a different strategy.  Instead of implementing the middle layer first, I recommend to use Tableau first before implementing a middle layer.  Allow Tableau’s natural functions to connect to sample data (Excel files, CSV files) or directly to source systems (Enterprise applications, databases) and start creating self-service BI reports.  Tableau users then, also have the benefit of learning Tableau earlier gaining valuable experience, rather than waiting for the middle layer to be finished.

This strategy is born out of Tableau’s confidence and ability in its own product to connect to your data and give you results right away without needing to start implementing this middle layer just yet.  Better yet, if you want to play it safe before laying out large investment costs (additional hardware costs, IT Staff/ IT consultants) in a middle layer, I recommend the following:

  • First and foremost, call your local Tableau provider to demonstrate the Tableau Technology for you … You will see how fast and easy it is to connect to data and start visualizing/analyzing your data with Tableau on your own
  • Next prototype Tableau Software on existing data sources in your organization before or during the beginning stages of your datawarehouse/database layer part of your project. By doing this, you can reap the benefits of your BI initiatives from day 1 without needing to put so much up-front money.

To further justify this strategy:

  1. By using Tableau immediately on your existing systems and data files, I have seen time and again, business users are already able to connect to 3 to 5 years of line item data using their PC alone! Tableau has high performance compression and in-memory technology that allow you consume large amounts of data. Gigs of data are compressed to MB size files.  Check out Tableau’s new Hyper technology!
  2. Tableau can handle simple to intermediate ETL (Extract, Transform, Load functions that are natively part of datawarehouse functions) , aggregation, and consolidation functions…that may be sufficient to take care of 80% of your reporting needs, while waiting for the datawarehouse to be implemented. It may be the case that a datawarehouse, may even not be needed at all or delayed for longer time, saving your organization a lot of money.
  3. Using Tableau Software to connect to data directly without the middle layer can also expose data issues which would give to your datawarehouse team advanced notice in order to help prevent or reduce rework efforts when building this middle layer.

 

 

In Part 2, we shall lay out a recommended a step by step Timeline strategy for your journey into BI

Dustin Guiyab, Subject Matter Expert in Business Intelligence and Analytics

Leave a Reply

three × five =

Cye Dela Cruz
Landline
0917 578 99 83
(02) 370 11 44
cye.delacruz@kaisa.com
COURSE REGISTRATION
Your Information will never be shared with any third party.
Course Interest
Thank you for Subscribing!
You will receive an email within 24 hours.
REGISTER FOR C4C TRIAL
Your Information will never be shared with any third party.
Contact
Your Information will never be shared with any third party.
Additional Information
COURSE REGISTRATION
Your Information will never be shared with any third party.
Course Interest
Inquire
Your Information will never be shared with any third party.
Message
Contact
Your Information will never be shared with any third party.
Additional Information
PHZpZGVvIHdpZHRoPSI3MjAiIGhlaWdodD0iNDgwIiBjb250cm9scz4NCiAgPHNvdXJjZSBzcmM9Imh0dHA6Ly93d3cua2Fpc2EuY29tL3dwLWNvbnRlbnQvdXBsb2Fkcy8yMDE4LzAxL1doYXQtaXMtU0FQLUJ1c2luZXNzLUJ5LURlc2lnbi5tcDQiIHR5cGU9InZpZGVvL21wNCI+DQpZb3VyIGJyb3dzZXIgZG9lcyBub3Qgc3VwcG9ydCB0aGUgdmlkZW8gdGFnLg0KPC92aWRlbz4=
Contact
Your Information will never be shared with any third party.
Additional Information
Contact
Your Information will never be shared with any third party.
Additional Information
Contact
Your Information will never be shared with any third party.
Additional Information
Jennie Magos
Landline
(02) 370 11 46
jennie.magos@gmail.com