Wednesday, February 17, 2016

Amazon and Dimensional Modeling


Hey guys,

I am back with my new blog. This blog mainly contains a brief description about Amazon and its business. It then talks about what could be some metrics that will be useful for the CEO to track the progress of the business and how dimensional models can be used for this purpose.

Introduction
Amazon.com, Inc. or simply Amazon is an e-commerce and cloud computing company in the United States headquartered at Seattle, Washington. Amazon is the largest online retailer in the United States. Found in 1994 by Jeff Bezos, it started out as an online bookstore and later on branched out into DVDs, CDs, software, apparel, furniture, etc. It has also grown its business beyond USA and maintains separate retail websites in many countries in Europe and Asia.

Metrics
Amazon’s business has been growing and to maintain its top spot in the online retail world, it will help CEO Jeff Bezos to have some metrics to analyze the performance of the company. Following are some of the metrics:
1.       Total quantity of items sold in a given time period: This metric will give an idea about the sales side of the business. It does not tell anything about the revenue earned from the items and hence, the items could be anything like cheap, expensive, large or small items.
2.       Revenue earned from selling these items: The revenue metric can be used to gauge the economic factor of the business.
3.       Profit: This metric can give us an idea about the actual money Amazon is making. By analyzing this metric, we can determine how much to spend based on the amount of returns we are getting.
4.       Number of items returned: This is an important metric for Amazon’s business model since it will help determine the percentage of faulty items in the warehouse. If this number is more, it means the suppliers are sending defective products and they need to be communicated to rectify the problem.
5.       Revenue earned per customer: Customers are important for Amazon and to know what revenue each individual customer is generating can be of great importance. By finding patterns among the high revenue generating customers, Amazon can make special offers for them.
6.       Revenue earned per item category: The category information can be useful to judge what categories are selling more and what are selling less. This way we can focus our attention to a particular category during Sale seasons.
7.       Items sold together: We can also analyze to see which items are often purchased together and add them in suggested list of items or create special offers on these items.

Dimensional Model
The metrics can be best tracked with the help of a dimensional model. With the help of a well-defined grain, these metrics can be accurately monitored. Dimensional models can help to bring together the required data from different sources and generate a single and consistent view for the user. Dimensional modeling can reveal certain inconsistencies and also help in fixing them. Whenever there are ad hoc queries, a dimensional model can help execute them to solve business problems faster and better. Dimensional models can also help increase the flexibility and scalability.
For Amazon’s business, Periodic and Transaction snapshot both can be equally appropriate. I would, however, go for the Periodic snapshot as it will give us an idea about all the metrics discussed above from time to time on a daily, weekly, monthly or yearly basis. A transaction snapshot can give us information about each transaction but we can analyze those with our metrics. From the point of view of future growth, the Periodic snapshot suits best the needs of the organization.

A sample dimensional model can be as follows:





In conclusion, dimensional modeling can be of great help to Amazon’s CEO Jeff Bezos to help determine important factors for company’s growth and make key business decisions.

References:
2.       http://www.amazon.com


Thursday, February 4, 2016

BI Tools - Analysis and Comparison

Hello Folks,

The BI and Analytics market is currently going through a major shift. More and more number of business users want to use interactive methods for analysis. However, these users have very limited knowledge about IT or data analysis. This blog compares the various tools available in the market with the key points of comparison being ease of use and learning and dependence on data analysis skills. By the end of this blog, you will have a fair idea as to which tool best suits your skills and needs.

According to the Gartner report, following is the graphical representation of products in the Business Intelligence and Analytics domain.



As we can see, some products are leading in the ability to execute while some have great vision. Let’s select 5 products that combine the best of both parameters.

1.     Tableau: Tableau is the leading tool with regards to its ease of use and providing excellent visualizations. It has a large customer base and is growing exponentially. However, it has made limited progress as far as integration with other tools is concerned.

2.      SAS: The highlight of SAS tool is its ability to integrate with a wide range of BI tools and its integrated products like SAS Office Analytics, SAS Visual Analytics and SAS BI/Enterprise BI Server (EBI). Some concerns are its high license cost, usability and business benefits.

3.     SAP: It is the enterprise standard as far as BI is concerned because of its product capabilities. However, user experience and business understanding are some of the areas where it can improve.

4.     Qlik: This tool has a highly interactive dashboard and is one of the top tools with respect to ease of use. Qlik is comparatively a new platform and does not have any production reporting as of now.

5.     MicroStrategy: It is an end-to-end platform majorly used for large and complex systems. Another advantage is its mobile capabilities. The development environment is flexible but it lacks behind in the ease of learning.

Now that we know some strengths and weaknesses of each of the BI tools, let us compare them with each other based on the some key criteria as follows:

1.     Reports and Data Visualization: The main function of any BI tool is to provide efficient reporting through effective data visualizations that will make it easy for the end user to understand it. Data visualizations are achieved through a variety of ways including graphs, pie charts, histograms, scatter plot, and more advanced features like plotting on map and a combination of any features mentioned. Reporting involves making efficient use of the visualization techniques.

2.     Usability and Dashboards: This is one of the most important features of any tool. The ease of use is very important from the point of view of the end user, especially for those who have little or no knowledge of data analytics. One of the ways to achieve this is with the help of efficient dashboards.

3.     Cost benefit: Accept it or not, money always matters. One cannot intend to buy a product at a high cost if it does not provide comparative business advantages. Cost includes the initial cost for the license of the product and the maintenance cost.

4.     Integration with other tools: Being able to use the features of some other tools greatly impacts the reporting. Hence, integration with other tools is important as we can combine the pros of both and this also helps the reporting to be flexible.

5.     Large Scale: Whether a tool is designed for large scale use or not is also an important factor. The choice of a tool is greatly influenced by this factor as we would need to know beforehand if it is capable of processing large data sets.

Now, let’s do a weighted score analysis of all the selected tools against these features to better understand where they stand with respect to each other.

Metrics/Tools
Weight
Tableau
SAS
SAP
Qlik
MicroStrategy
Reports and Data Visualization
30%
9
9
8
9
8
Usability and Dashboards
25%
10
9
8
9
7
Cost benefit
20%
9
8
8
9
9
Integration
15%
8
10
9
7
9
Large Scale
10%
9
9
10
8
10
Total points
100%
9.1
8.95
8.35
8.6
8.3
Rank

1
2
4
3
5

The weighted scoring model tells us a lot about the efficiency of a tool with respect to any particular feature as compared to other tools. Let’s have a closer look at these tools and the metrics they were compared against and understand why a particular score is awarded to a tool.

1.     Reports and Data Visualization: Tableau, SAS and Qlik are leading in this section. As seen from the Gartner’s quadrant, Tableau has a strong lead in the Ability to Execute axis. Its reports are intuitive and easy to understand. SAS offers innovative visualizations using powerful in-memory processing capabilities. Furthermore, SAS Visual Analytics is a flagship product which delivers interactive reports.

2.     Usability and Dashboards: Tableau ranks number one in this department. It has the best and the easiest to use dashboard that can help people create effective reports. Tableau is followed by SAS and Qlik with scores of 9 each. Qlik gives the user to build their own dashboards while IT is given the ability to govern, manage, scale and embed them.

3.     Cost benefit: Tableau is ranked among the top when the cost to business benefit ratio is calculated. Qlik is comparatively new in the market and hence it is low cost but provides great business benefits. MicroStrategy is an industry benchmark for large BI deployments and hence its cost benefit ratio is quite high.

4.     Integration: SAS, SAP and MicroStrategy are the leaders here. SAS has a wide range of products within itself such as SAS Visual Analytics, SAS Office Analytics and SAS BI. SAP products like SAP Lumira can be integrated SAP BusinessObjects Enterprise. With the recent big data related enhancements, MicroStrategy has come up with Prime, a scalable multiterabyte in-memory engine developed with Facebook and direct Hadoop connectivity.

5.      Large Scale: MicroStrategy and SAP score full points here. MicroStrategy is an end-to-end BI platform that is designed to handle large and complex data sets. SAP BusinessObjects BI tool is chiefly used for large customer deployments.


Based on this analysis and the total score, Tableau comes at the top and it is safe to assume that Tableau has the best mix of features compared to the other BI tools as far as the above mentioned metrics are concerned.

References:
4.       www.sapbi.com/
6.       http://www.docurated.com/all-things-productivity/50-best-business-intelligence-tools