Other Considerations for Creating Embedded Analytics
Updated: Apr 21, 2020
It is of critical importance to set up the right resources along with processes and manages stakeholders’ expectations along the way to make it a success. Basically, the entire implementation needs an incremental approach split into the following phases:
- Discovery & Design
Below is a sample description of how the skillset composition looks to be to deliver the unique needs at each phase.
The Total Cost of Ownership of BI Implementation
This is important to understand that the entire cost of BI implementation depends on many factors viz. the Volume of data processed, the complexity of data sources, the infrastructure components taken up, the team set up, the management expectations etc. Overall, a mid-size company which is very much data-driven could see the cost of BI anywhere ranging from $20K-$250K annually. There are opportunities for many accounting firms, individual companies to implement BI at their premises.
Broadly, the total cost of ownership can be strategically put under the following buckets:
(a) Resource Cost: There could initial cost of setup the entire BI and then the maintenance, enhancement cost going forward. As described, depending on the complexity of this engagement, one may need a management consultant, project manager, BI developer, Data Analyst, QA engineer, Data engineer, Web Developer, UX/UI designer etc. (b) Infrastructure cost: Infrastructure cost include the on-premise cost or cost of various components taken up on the cloud. Please note that the cost can significantly vary depending on which technology stack is taken up – Open source technology viz. building up the data architecture using Python etc. could be highly cost-saving but needs a lot of maintenance which probably increases the resource cost to maintain or change it. However, technology stack from Microsoft viz. Azure Data Factory, Data lake store, Azure Analysis Services, Azure SQL Datawarehouse etc. would be costly as compared with Open source tools but there are drag and drop features which makes it very easy to maintain and change.
Designing Data Pipeline and Optimizing Data Mart
Once the initial setup of the underlying Data Architecture is over, there are several optimization steps that can be built-in depending on the complexity of the Data. For example, if Data needs to be refreshed on a daily basis and the volume of data is too big, it is best to do an incremental load (say, we loaded the data until 7th April and the only data pertaining to 8th April is loaded instead of doing a full time), fault tolerance to recover in case of errors etc.
Track Usage Stats
If you are running a professional services firm or a data aggregator company which shows client data to different users, it would be great to captures statistics on report usage, number of views per the report, number of users opening a particular report etc.
Smarter organizations are ramping up investments in BI to create strategic proposition and position them differently from their competitors in the extremely competitive market. The Aays Analytics Platform brings together the core capabilities of today’s most powerful BI technologies, enabling smart solutions across our offerings to augment and enhance BI development.
We’re helping our clients across many different industries to optimize outcomes by leveraging available data assets. We manage data assets every day, to help clients quickly and easily derive insights and take action by applying highly sophisticated data modelling and analytics capabilities.