dunnhumby Guest Lecture Flashcards Preview

Business Intelligence > dunnhumby Guest Lecture > Flashcards

Flashcards in dunnhumby Guest Lecture Deck (18)
Loading flashcards...
1
Q

Business Overview

A
  • customer-centric operation / customer first
  • they partner with retailers and vendors to embed the customer,
  • meet current and future needs
  • offer a mix of consulting, software-led tools and media solutions
2
Q

dunnhumby’s Data Strategy Framework

A

How data must be considered to meet business challenges and objectives

3
Q

dunnhumby’s Data Capability Ladder

A

Level 1 – Essential Data

  • data a company captures for legal reasons,
  • e.g. reporting accounts to the tax authorities

Level 2 – Advanced Data

  • the vast majority are here
  • Loyalty card data – can associate the data with an individual Clickstream data; progressing through digital

Level 3 – Joined Data

  • data can be used for different reasons - true data analytics
  • e.g., HR data can work out how many staff you have on the floor and correlate it with sales spikes

Level 4 – Enriched Data

  • enriching via data profiling
  • retailer will have a score about you – a propensity score
  • Connected Customer Data – is the customer, family, household

Level 5 – Data Partnerships

  • never enough data – find third parties with complementary data
4
Q

A Typical Data Platform

A
5
Q

Layers of the data platform

A
6
Q

Data Alligned Business

A

Focusing on data instead of gut feelings.

Understand how your customer and data strategies aligns to business requirements and how each part of the data roadmap supports and enables measurable business objectives and getting the most value of your data and technology resources.

Linking Customer and Data Strategies

  • There is a clearly defined customer strategy that enables the creation of a comprehensive data strategy through measurable business objectives.
  • The data strategy is seen as a key to competitive advantage.
  • Organization approach is shared to the industry through conferences and publications with external feedback sought.

Business Requirements and Objectives

  • All Business requirements objectives support the customer strategy.
  • KPIs measure objectives and set priorities

Data and Technology Roadmap

  • Data and technology projects are aligned to measuring business objectives that support your customer strategy.
  • The roadmap should be prioritized with key stakeholders across the business and well communicated.
  • Continual improvement and feedback loops exist.

Internal Stakeholder Alignment

  • An internal alignment of stakeholders on data and technology priorities results in a clear perception of data value in the organization.
  • Scope of the strategy and roadmap is well understood across the entire business and well documented.
  • Clear accountability of data is understood within the business.

It is very common that decision-making within companies is not based on any data sets but more on experience and other arguments. This is not sustainable.

7
Q

Connected Customer Data

A

Review what 1st, 2nd, and 3rd party data a client has at every point of engagement, such as their needs, how they behave, how they pay, and what they say about you and your products across different channels.

Defining the Customer

  • Customers can be linked to behaviors like transactions.
  • Or throgh identification like email.

Single Customer View

  • Customer identities from multiple sources are collected, linked and managed centrally.
  • There is a clear and nested customer hierarchy if multiple customer levels exist (e.g. card > person > household).

360 View of the Customer

  • Customer can be linked to a breadth of data assets covering their behaviors, motivations, thoughts and interactions.
  • Customers can be linked to 1st, 2ndand 3rdparty data assets.

Data Granularity & History

  • Key customer interactions, e.g. sales orders, or direct customer contacts are held a granular level.
  • All data pertaining to any customer facing part of the organization is collected cleanly and held for a required and reasonable period of time.
8
Q

Data Governance

A

Identify and clarify policies and processes for handing data, managing risks to legal, data security, customer trust and overall management and education surrounding ways to use the trusted data.

Data Security, Privacy & Compliance

  • Security Requirements and Governance is defined, communicated and accessible.
  • Regular database vulnerability tests are conducted with results shared and improvements planned.
  • Data Security Maintenance is well managed.
  • Data Retention rules are defined.

Master Data Management

  • System/Data of record is tracked or well documented. Access to Data Profile Metrics.
  • Established Master Data Business Rules and Governance are well documented, communicated and accessible.
  • Ownership and Stewardship of data assets is well established.

Metadata Management

  • A metadata governance process is well established to maintain its accuracy.
  • Entire organization utilizes the 3 types of metadata (business, technical and operational).
  • Data lineage is easily identified.

Data Knowledge Management & Communication

  • Documentation supporting understanding data and technology is organized, accessible and refreshed for accuracy.
  • Data and Technology roadmaps details exist with business value communicated to entire organization.

Data Quality Management

  • Data Quality is planned into all projects.
  • Post-fix checks on data are performed and results shared with appropriate stakeholders.
  • Cross-system scorecards exist, Automated KPIs exist, Quality Governance framework is well established.
9
Q

Data Architecture

A

How data is currently held, in which technologies, how data flows through different systems, how frequently these data sources are updated, and how they enable downstream use by business users.

Data Technologies & Administration

  • Data Management technologies (DBMS, File Storage Systems) are well aligned to business objectives and allow for cost effective, scalable solutions.
  • Change Management best practices are adopted and group meets regularly to discuss proposed changes to production systems.
  • Performance Metrics are feedback to stakeholders.

Data Architecture Management

  • Data models are optimized for creation and maintenance independently of data modeled for consumption needs.
  • Consolidated data operations are managed under the guidance of an established MDM program.
  • Data Architecture Implementation is well documented and understood within the organization.

Data Consumption

  • Data stores are organized and designed for varying data consumption needs.
  • BI Technologies and Tools
  • Data visualizations

Data Enrichment

  • Make data usable by application of business rules, understand the velocity of data coming into your systems, learn how to integrate new data sources and your ability to create enriched data services.
10
Q

Talents, Teams, Ways of Working

A

Assess and measure the current skills, skill levels, and support structure within teams and divisions overall and understand the methods at which business colleagues collaborate towards an aligned set of strategic priorities.

Data Leadership

  • Clear accountability
  • There is a named individual who is accountable for the data strategy across all areas of the business (i.e. “Data Champion”)

Resources, Skills & Experience

  • The right amount of resource, skills and experience to support the data and tech roadmaps exists with the organization.
  • Investment in data people resource is prioritised and clearly linked to achieving specific business objectives or use cases

Organization Structure

Ways of Working

  • optimised working
11
Q

Key benefits of a data platform

A

Innovation friendly

  • Raw source data can be directly consumed by data scientists to investigate new sources of data and new capabilities without the cost of over integration.

Organised and optimised

  • Organised data to support specific capabilities. As needed, multiple logical data stores are created to support varying consumption requirements.

Close to the source

  • Raw source data is directly loaded to a data lake without any transformations to maintain the integrity of the source data.

Intelligent enrichment

  • Enrichment can happen during traditional extract, transform, load (data) and dynamically at the point of consumption using APIs.

Focused and flexible

  • Not all data needs to be enriched, so it will only exist in the data lake. This helps us focus on high value, high frequency data, saving time and money, but retaining the flexibility to work with new sources in the lake.
12
Q

Why the cloud?

A
  • Reduce hardware and upgrade costs
  • Meet demand – ramp up and down as needed
  • Flexible plans – pay as you go
  • Improved security
  • Access to new tools and machine learning APIs
  • Reduced carbon footprint
13
Q

Technologies used

A
14
Q

Role of Data Partnerships

A

Data partnerships can accelerate delivery of data strategy

15
Q

Key factor in privacy

A

informed consent

16
Q

GDPR Basic Principles

A
17
Q

Micro-locationing

A

Using phones, people can be tracked inside stores and the company can predict where they will go next in the store.

  • location-specific adds
  • measure effectiveness of in-store promotions & media
18
Q

define clickstream analysis

A

capturing online behaviour by recording the clicks