Pentaho the best BI integrator
Updated November 04, 2014

Pentaho the best BI integrator

Deepak Paramanand | TrustRadius Reviewer
Score 7 out of 10
Vetted Review
Verified User

Software Version

3.2

Modules Used

  • PDI
  • Mondrian/Schema Workbench
  • BI Server

Overall Satisfaction with Pentaho

Our product intended to analyse Media and Communications data and provide decision making capabilities to our end users. Whereas we built the workflow management, transactional capabilities, Pentaho provided the BI layer for our product. Analysts were our end users. Everything from data ingestion to decision making was the scope of our product. We read data across disparate data sources/formats and using our business expertise provided decision making capability.
  • Data Integration. Pentaho wins hands down. You can read huge data using a Hadoop process, do your encirchment, load it to a Netezza database afterwards, finally input the data to your WEKA model to predict which Customer will churn or what offer should be made to the customer so that he/she stays put.
  • BI Server. If you want to schedule a data read operation of your clickstream data, to finally burst out recommendations of next best actions to end users, Pentaho's BI Server performs this integration seamlessly.
  • Mondrian. This layer provides Cube based Hierarchical data modeling on the front end and at the back end converts this OLAP structure to a ROLAP, SQL based model. Hence any relational database becomes a ROLAP engine seamlessly.
  • Data Visualization. Provide richer library of data visualization capabilities. Ultimately the dashboards are the end result of all the hard work done at the back office. Yes, Pentaho has a lower TCO compared to other Products but richer data visualization capabilities would make it a winner! Pentaho has overcome this limitation by allowing external charting engines to be integrated with their product suite, but more needs to be done to strengthen core Pentaho Data Visualizaiton capabilities. Alternatively the external charting engine capabilities need to be documented and evangalized.
  • Alternative to Mondrian. In our case we needed to analyse data for a million subscribers over key performance areas like Churns, Activations etc. In these scenarios semi additive measures needed to be calculated and presented in a report across days to years grains of time. In such scenarios the Mondrian based ROLAP capability did not scale up to our expectation. Pentaho needs to address such issues and fast.
  • In the world of R/SAS/SPSS its hard to find use cases where WEKA was used in Production environments to solve a business problem. We would have needed some hand holding to replace our R/SPSS code to WEKA and help us build newer alogorithms on this platform ground up.
We evaluated Panorama, Cognos, MicroSrategy, Jasper Reports, Talend and homegrown solutions. Though each were awesome in their own right, none of them provided a end to end integration like we wanted. Pentaho did the job for us and more. Knowing that Pentaho was built by a team who were industry veterans made us feel comfortable that the basics were always a given and that we could always reach out to them for more. Plus Pentaho's responsiveness in providing a hands on expert to train us was a huge plus.
We have moved to homegrown ETL and have matured as a company as to what we need from a BI tool. Our focus is now rich and intuitive data visualization with blazing speed. Our customers had already used MicroStrategy/Tableau in the past so that bar was already set very high. Unfortunately Pentaho could not give us a better if not similar experience as the tools our Customers had already used.
Integration is best with Pentaho. If you want to overlay your product capability with Pentaho's BI Server, Data Integration, Reporting Engine, Workflow capabilities then Pentaho is the only answer. Other tools provide each of these very well, but if you have already built your product and dont want to buy Informatica, Cognos/Microstratey and SPSS/SAS then Pentaho is the way to go.
Pentaho is Java based, so no shortage of skilled Java resources who can help you integrate. We did not spend time building our own reporting engine, Data Visualization layer, Security layer, workflow capabilities. We simply used our proven transactional product and used Pentaho to fill our gaps.

Using Pentaho

10 - Managers, System Analysts and Business Analysts. We provide financial analytics hence the users are somewhat tech savvy, know how to use the system and get meaningful information from it to complete their day to day activities. Some users use our product for strategic planning purposes such as budgeting and marketing spend.
5 - Java is a must have. Somebody who is proficient with object oriented programming concepts, understands APIs.
ETL Developer. Must have hands on experience coding ETL using SQL and/or proprietary tools.
Report Developer. Must have experience with building reports with Excel, Business Objects et al.
Solutions Engineer. Must understand how Pentaho fits with the overall Product offering. Should have the customer in mind at all times during product integration.
  • As a replacement for our costly and technically disparate ETL and data visualization solution.
  • As a integration software for our transactional systems with our BI systems.
  • As a workflow management, business process specific application system.
  • As a dashboard/reporting solution.
  • As a end to end data integrator. Right from data ingestion to pretty pictures that provide business value and decision making capability.
  • As a workflow management solution.

Evaluating Pentaho and Competitors

  • Price
  • Product Features
  • Product Usability
Integration. That Pentaho had ETL, reporting, scheduling, workflow management included in one software plus being Java based it could be easily integrated with our existing software stack was the most important factor in choosing Pentaho.
I would place high marks on the visualization layer. If I cannot use one aspect of a Product suite it becomes difficult to evangelize the product within an organisation. Or I would set internal expectations that though the Product offers various modules, our evaluation should restrict itself to 2-3 core ideas and evaluate accordingly.

Pentaho Implementation

Get the right people in before starting implementation. Start small and build as you go approach is time consuming and involves lot of rework.
Evangalize within the organization the capabilities and limitations equally so that correct delivery expectations are set.
Set expectations with the Customer that the tool cannot replace proprietary software in terms of stability/usability and that timelines could change given the new ness of the product.
Yes - We first chose the reporting solution as a module to integrate. Via API calls this module was integrated with our existing Product for transactional reporting purposes. Next phase of our implementation was centered around integrating the ETL solution into our product as a single point of data ingestion mechanism. Once the data was brought in using Schema Workbench we built the logical layer to help publish dashboards/reports via the Pentaho BI Server.
  • Technical expertise. Finding the right people who understood Java and BI was a tough ask. So we decided to keep the teams separate and let a Solutions Engineer/Architect provide direction for the integration.
  • The team had prior experience in using stable proprietary products where documentation was plenty and hence any issue could be solved easily by tapping into the knowledge base around. Customer support interactions were infrequent. With Pentaho it was vice versa, where we depended more on Customer support than on community knowledge.

Pentaho Support

They were responsive to our questions when we raised issues.
They gave us workarounds when required.
They were quite knowledgeable when it came to issue analysis and providing fixes.
They were forthright in informing us if a bug was not due for release soon.
ProsCons
Knowledgeable team
Problems get solved
Kept well informed
Support understands my problem
Support cares about my success
Escalation required
Slow Initial Response
Yes - For any production related issues where a hotfix is the only way to go.
For any configuration related issues where best practices are known but are not working as expected.
Yes - Some bugs were resolved whereas some were slated in a much later release.

Using Pentaho

I would have liked some hands on help in trying to match what Customers were already used to. Plus having a team of Java/BI folks was difficult to assemble since what required to be done in an object oriented fashion could have easily been done via a SQL construct. This hash of technologies was difficult to manage.
ProsCons
Like to use
Technical support not required
Feel confident using
Difficult to use
Slow to learn
Lots to learn
  • End to end data ingestion to data visualization is easiest to do.
  • Dashboard/report creation, distribution is easy too.
  • Workflow integration is best used in Pentaho. You can schedule an invoice to be loaded as soon as it arrives in your ftp server, you can apply business rules, load it to your database, refresh the data warehouse with this latest invoice and finally burst an email with telling all users to go and see the report. All this can be achieved in Pentaho's Workflow integration tool.
  • You HAVE to know Java to use Pentaho to the best of its capability. For pure BI enabled users like myself, it was a steep learning curve.
  • Trying to match Customers data visualization expectations was tough, Our customers had already used Tableau and/or Microstrategy/Cognos/Business Objects, hence it was always a tough ask to match/better what they had already seen and experienced.