Professional Documents
Culture Documents
5.1. Key Performance Indicators
5.1. Key Performance Indicators
KPI
is a visual cue that communicates the amount of progress made towards a goal
An organization must establish its goals and then choose the KPIs which best reflect those goals.
For example, if a company's goal is to have the fastest growth in its industry, its main KPI may be
the measure of revenue growth year-on-year.
Number of shares: Content that gets shared by your audience is one of the fastest ways to grow
relevant follower bases in social networks
Link Click Through: is a free and easy way to gauge the effectiveness of your brand messaging.
If no one clicks your links, you probably aren‟t presenting relevant and valuable information (or
your follower base is not the right audience)
Sales
Sales Growth: Analyse the pace at which your organization's sales revenue is growing and use
that information in strategic decision-making
Product performance: Measure and rank products based on revenue performance, to inform your
sales team as to which products are selling well and which products need special attention
Sales per representative: Measure the ability of each of your sales representatives to generate
revenue for your organization and foster a healthy, productive level of competition amongst team
members
Content Marketing
Unique Visits: Are the most standard measure of how many different individuals have viewed
your content within a given time frame (typically a 30-day window)
Mobile Readerships: Understanding trends in how your content is being delivered to different
devices is key to determining how to optimize your content and its design (i.e., responsive
design) for future publications
Page views: is a good measure of which content is being viewed the most. They help you
understand how to develop future content for your audience
SMART Criteria
KPI should be well defined and should follow the SMART criteria.
S-Specific
The metrics should be specific and targeted to the area you are measuring. It should be clear
and focused to avoid misinterpretation.
For example, if you are measuring customer satisfaction, a good metric would be direct feedback
from customers on how they have rated your service or product. A poor metric would be simply
the „number of returned products‟ or „number of customer complaints‟ (without correlating with
the number of products being sold). A product that has very few consumers will have very less
customer complaints or number of returned items; while a product that has high demand might
have relatively higher number of customer complaints.
M-Measurable
The measure should be quantified so that it can be compared to other data. It should allow for
meaningful statistical analysis
A-Achievable
The metric should be attainable, reasonable, and credible under conditions expected.
R-Relevant
This means that the metric shouldn't measure things that are not important.
T-Timely
Timely metrics are those for which we can get the data when we need it.
Enterprise Reports
Enterprise Reports (or Management reporting) is the regular provision of information to decision-
makers within an organization to support them in their work. These static reports can take the
form of graphs, text and tables and, typically, are disseminated through an intranet, as a set of
regularly updated web pages (or "enterprise portal"). Alternatively, they may be emailed directly
to users or simply printed out and handed around, in the time-honoured fashion. The emphasis
here is on the effective delivery of information to managers. The below reports are examples of
Enterprise reports.
Operational reports
Production reports
Sales reports
Statutory Reports
Invoices
Account Statements
Employee Reports
OLAP Reports
OLAP Reports are dynamic reports in which data manipulation is possible using *OLAP features
like Slicing/Dicing, Pivoting/Rotation, Drill Up/Down, Drill Through, and Trend Analysis.
OLAP Features
Rotate / Pivot
This operation rotates the data axes in view, in order to provide an alternate presentation of data.
The slice operation performs a selection on one dimension of the given data. Slice is the act of
picking a subset of the data by choosing a single value for one of its dimensions.
The dice operation performs a selection on two or more dimensions allowing the user to get an
exact view of the data required. The dice operation produces a subset of the data by picking
specific values of multiple dimensions.
Roll up and Drill down
Moving up the hierarchy of a dimension is called „Roll up‟ (also called as Drill up in some
reporting tools), and moving down a dimension‟s hierarchy is called „Drill down‟. In drill down, we
navigate from less detailed data to more detailed data; while in roll up, we navigate from more
detailed data to less detailed data.
In the below figure, the first chart shows the yearly sales in Delhi and Chennai. Then a drill down
operation is done on the year 2013. The next level in the Time dimension hierarchy is Quarter.
So when we drill down on year 2013, we get the quarterly sales in Delhi and Chennai for that
year. When we drill down on the third quarter, we navigate to the next level in the time dimension
i.e. the months.
In roll up we move up the different levels of hierarchy in a dimension.
Ad hoc Reports are interactive, customized reports generated by business users/analysts on-the-
fly depending on the specific analyses they want to perform. These kinds of reports are typically
undertaken once to deal with a specific initiative, and then never revisited. They often involve
building a model in a spreadsheet to allow exploration of "what-if" scenarios.
Consider a retail store chain. Let us assume that during quarterly analysis, the management
finds that the revenue from one store has dipped a lot during the current quarter. In this case, the
manager might ask the IT team for an ad hoc report that will help him/her to find out why that
particular store performed badly in that quarter. In this case the creation of that report is a one-
time activity and it will not be used later.
Dashboards and Scorecards are “visually intuitive” reports that provide “at a glance”
information about business performance, to top executive management. Dashboards can have
Text, Reports, Charts, GIS (Geographical Information Systems), Dials, Graphical indicators, or a
combination of all.
Scorecards track business critical KPIs and also show their current value, target value,
variance/trend. It's common to have traffic lights defined for performance (red, orange, green) to
draw management attention to particular areas.
5.3. Multidimensional Database – MDDB
MDDB is a specific type of database that has been optimized for data warehousing and OLAP
(online analytical processing). A MDDB inherently uses multidimensional structures (commonly
known as cubes) to organize and store data. Note: The term „Star schema‟ is typically used in
Relational databases context, and „cubes‟ in the MDDB context.
The relational database model uses a two-dimensional structure of rows and columns to store
data
Here for storing the same amount of data, the relational table took 9*3 =27 cells (see relational
table above), while MDDB took only 3*3 = 9 cells.
Each axis in an array is called a dimension (one of our data perspectives). Each element within a
dimension is called a Member. In Sales volumes example, model and color were the dimensions.
Model dimension had 3 members (Minivan, Sports coupe and Sedan), and Color dimension had
3 members (Blue, Red and White).
Benefits of MDDB
In MDDB, data is stored in the same way it is viewed. The user is able to view data along
dimensions with data arranged in an inherently more organized, and accessible fashion than the
one offered by the relational table. A great deal of information is gleaned immediately upon direct
inspection of the array. Structure of the array tells the dimensions & possible values of each
dimension.
Storage Space
Very low space consumption compared to a Relational DB. Multidimensional array is a more
efficient & effective means of organizing data.
Performance
MDDB gives much better performance than relational databases. Relational DB may give
comparable results only through database tuning (indexing, keys etc.), which may not be
possible for ad hoc queries.
Ease of Maintenance
No maintenance overhead as data is stored in the same way it is viewed. In Relational DB,
indexes, sophisticated joins etc. are used which require considerable storage and maintenance
Disadvantages of MDDB
Sparsity
Sparsity (the property of lacking denseness) occurs when all the members of different
dimensions do not interact.
For example in the previous example let us assume that the sedan is available only in blue color,
coupe is available only in red color and minivan in only blue and white. Then the relational table
will look like this.
In relational database if data does not exist for a combination of dimension, then that row does
not exist. However, in MDDB, in such a situation, a blank cell is left behind and that space cannot
be utilized.
Data Explosion
Data Explosion is a phenomenon of multidimensional databases which means that data size
multiplies when it is loaded from source into MDDB. When the database “explodes”, the size of
the database becomes much larger than what it should be. It is caused due to the data sparsity
and pre-aggregation of data.
For MDDBs with dimensions having 50% members in summary levels, Data explosion is found to
be four times the detailed data for two dimensions, and eight times the detailed data for three
dimensions.
Performance
MDDB does not perform better than RDBMS at high data volumes (>20-30 GB).
In MOLAP, data is stored in a multidimensional cube. The data is not stored in a relational
database. MOLAP is more appropriate for cubes with frequent use and for rapid query response.
Advantages:
High performance: MOLAP cubes are built for fast data retrieval, and are optimal for
slicing and dicing operations.
Can perform complex calculations: All calculations have been pre-generated when the
cube is created. Hence, complex calculations are not only doable, but also fetch the
results quickly.
Disadvantages:
Limited in the amount of data it can handle: Because all calculations are performed when
the cube is built, it is not possible to include a large amount of data in the cube itself. This
is not to say that the data in the cube cannot be derived from a large amount of data,
which is very much possible. When large amount of data have been used to derive the
cube, only summary-level information is typically included in the cube.
Requires additional investment. Cube technology is often proprietary and does not
already exist in most organizations. Therefore, to adopt MOLAP technology, additional
investments in terms of human and capital resources are often needed.
Advantages:
Can handle large amounts of data: The data size limitation of ROLAP technology is the
limitation on data size of the underlying relational database. In other words, ROLAP itself
places no limitation on data size.
Can leverage functionalities inherent in the relational database. Relational databases
already come with a host of functionalities. ROLAP technologies, since they sit on top of
the relational database, can therefore leverage these functionalities.
Disadvantages:
Performance can be slow: Because each ROLAP report is essentially a SQL query (or
multiple SQL queries) in the relational database, the query time can be long if the
underlying data size is large.