Tieme ten Veen
Business Intelligence and Analytics
Samenvatting artikelen BIA 2021/2022
Week 1
Hartmann, Zaki, Feldmann, Neely: Big data for big business? A taxonomy of data-driven
business models used by start-up firms.
Big data ensures that data driven business models (DDBM) are needed. Nine building blocks are formed
to create a business model framework:
1. value proposition
2. key processes
3. key resources
4. key partners
5. customer relationships
6. channels
7. customer segments
8. revenue streams
9. cost structure (BMC)
The Data Driven Business Model framework, consists of:
- Key resources: what kind of data is used, structure the data, consolidating different sources.
- Key activities: data-related activities, divided into descriptive, predictive and prescriptive
analytics. Activities along the virtual value chain: gathering, organising, selecting, synthesising
and distributing (Rayport & Sviokla).
- Value proposition: ‘expression of the experience that a customer will receive from a suppliers
measurably value-creating’ (Barnes et al.).
- Customer segment: B2B and B2C
- Revenue model.
- Cost structure.
For this research, clusters have been made according to a four-stepped process (Ketchen and Shook):
1. selection of clustering variables
2. choice of clustering algorithm
3. choice of number of clusters
4. validation and interpretation of clustering result
The DDBM framework provides a basis for analysis and clustering of business models.
LaValle, Hopkins, Lesser, Shockley & Kruschwiz: The New Path to Value.
Analytics driven management can rapidly create value. Analytics are a differentiator. The emerging
methodology and its five critical recommendations are set out below:
- Focus on the biggest opportunities first
- Start with questions, not with data
- Embed insights to drive action
pg. 1
,Tieme ten Veen
Business Intelligence and Analytics
- Keep existing capabilities while adding new ones (keep spreadsheets, while adding new
visualization tools). Add an ‘analytics unit’.
- Build the analytics foundation according to an information agenda (data has to be integrated,
consistent and trustworthy). Information governance policies, data architecture, data currency,
data management, analytical tool kits should be on the agenda.
In an organization there are three levels of analytics capability:
1. aspirational (assemble the best people and identify challenges)
2. experienced (move to enterprise analytics and collaborate)
3. transformed (discover and champion improvements in analytics)
Most challenges occur with understanding the data to improve the business. For a management to
successfully implement analytics driven strategy this five-point methodology could rapidly create value.
Each point meets all these critical management needs:
- Reduce time to value
- Increased likelihood of transformation that’s both significant and enduring
- Greater focus on achievable steps
Steps to take:
- Pick your spots, identify the highest priority challenge, and create a PADIE diagram.
- Prove the value, with the PADIE diagram, show the gain in revenue
- Roll it out for the long haul, challenge big, model insightful, business vision complete.
Parmar, Mackenzie, Cohn & Gann: The new patterns of innovation, Harvard business
review.
Expand the existing frameworks: How can we create value for customers using data and analytic tools
we own or could have access to? 5 Patterns:
- Using data that physical objects now generate to improve a product or service or create new
business value. (Ex: smart metering energy usage, measure how safe one drives)
- Digitizing physical assets (documents)
- Combining data within and across industries (measure home conditions such as temperature,
co2-level, water usage and form patterns and alarm when this pattern is ‘broken’)
- Trading data (data from one company can be used by the other and vice versa, ex: Vodafone
and TomTom)
- Codifying a capability (sell what you’re good at, or invented (and what’s not central to a
company’s competitive advantage))
Strong technology presence, inputs from external parties, motivated leadership, emotional commitment
are success factors ‘needed’ to implement these patterns.
Woerner & Wixom. Big data: extending the business strategy toolbox.
Big data is used by companies to optimize business processes and decision making (strategy). Business
nowadays is:
- obtaining new data (using sensors to track actual product usage).
pg. 2
,Tieme ten Veen
Business Intelligence and Analytics
- gaining new insight (new approaches and technique, like high end stats and visualizations).
- taking new actions (responding to real time customer sentiment).
Big data is used to innovate the business model and find new ways to generate revenue.
- Data monetization (exchanging information-based products and services for legal tender or
something of perceived equivalent value). Monetizing happens by selling, bartering, and
wrapping.
- Digital transformation (creates new types of data, blurs companies’ boundaries).
Big data can be used to understand the customer at its best. Furthermore, big data can be used to
extend a company’s business strategy toolbox.
Week 2
Diane M. Strong, Yang W. Lee, and Richard Y. Wang D COMMUNICATIONS OF THE ACM
May 1997, Data Quality in Context
DQ = Data-Quality
IS = Information Systems
3 roles in data manufacturing systems:
1. Data producers (people, groups, or other sources who generate data)
2. Data custodians (people who provide and manage computing resources for storing and
processing data)
3. Data consumers (people or groups who use data).
Each role is associated with a process or task:
1. Data producers are associated with data-production processes.
2. Data custodians with data storage, maintenance, and security.
3. Data consumers with data-utilization processes, which may involve additional data aggregation
and integration.
We define high-quality data as data that is fit for use by data consumers—a widely adopted criteria. This
means that usefulness and usability are important aspects of quality. Using this definition, the
characteristics of high-quality data (Table 1) consist of four categories: intrinsic, accessibility, contextual,
and representational aspects. This data consumers’ perspective is a broader conceptualization of DQ
than the conventional intrinsic view.
We define a DQ problem as any difficulty encountered along one or more quality dimensions that
renders data completely or largely unfit for use. We define a DQ project as organizational actions taken
to address a DQ problem given some recognition of poor DQ by the organization. We intentionally
include projects initiated for purposes other than improving DQ. For example, during conversion of data
to a client/server system, poor DQ may be recognized, and an improvement initiated.
pg. 3
, Tieme ten Veen
Business Intelligence and Analytics
Divide a DQ project in 3 problem-solving steps:
1. Problem finding (how the organization identified a DQ problem).
2. Problem analysis (what the organization determined the cause to be).
3. Problem resolution that includes changing processes (changing the procedures for producing,
storing, or using data)
After these three steps you change the data (updating the data value).
pg. 4