Analytics & Data Strategy 101 by Deko Dimeski

Analytics &
Data Strategy
101
Why? What? When? How?

…set of tools, practices and technologies used for
discovery, interpretation and communication of
meaningful patterns in data.
…entails applying data patterns towards effective
decision making.
Analytics, is…
Machine or Human
or integrated in
other products

Evolution Of Analytics
Data Quality Aspects:
1. Accuracy
2. Timeliness
3. Completeness
4. Uniqueness
5. Consistency
6. Validity

 Data Insights
Dashboards, Business Answers, Strategical / Tactical & Operational
Reporting, AdHoc Analysis
 Data Integrations & Intelligent Services
 Data As a Service
Prediction, Segmentation, Personalization, Scoring, DWH/Data Lake,
Data Mainlining Tools, BI Self Service
 Knowledge, Expertise & Tools
Some Of The Generic Data Products & Services

Finance is the only function that needs data!
We have our developer John Dow who knows
Python and SQL, he can create any report I want
from our database?
We have a tons of data, our devs will setup a
DWH and we can do great stuff
Wrong Assumptions
Our data are perfect!

The Quick & Dirty Poor Practices
Report with
Active Customer
 How old & relevant are this data?
 How did you calculate the Active customers?
 Is this End od Month count or monthly aggregate?
 What’s the definition of active customer?
 Is it users or customers, what is the difference?
 How can I combine this information with my GA
behavioral metrics?
 I have Customer IDs but I don’t know who are
these customers.
 Can I use this data also for my Outbound
Campaign.
 Who can tell me if this information is correct.
 The number of customers does not match the one
in our CRM.
 Something has changed with my previous record.
 I cannot get the historical data because we don’t
have logs or the customer fields are not
versionized.
“John Dow, can you Slack me report with the Active Customers?

How Complex It Can Get
Multiplied:
 Event Data
 Streaming Data
 Unstructured Data
 Different data storage technologies
 Non-matching Data Models
 Missing Records & Poor Data Quality
 Missing historical information
 Data are not modelled/designed for
analytics
Your Company

 PO
 Data Analyst / Scientist
 Data Engineer
 ML Engineer
 Software Engineer
Typical Analytics Team
This is not your SW engineer
who knows Python, Scala or
SQL! Unlike developers, these
folks know what type of
problems the DS/DA
experience.
Analysts can be specialized in
Web (Behavioral), Business,
Spatial Analysis, etc…
Customer Facing Roles

Typical BI/Data User Personas
Leadership Team
•Type: BI Consumer
 Characteristics:
I oversea broad company initiatives,
strategy and manage people
I'm frequent traveler and heavy
smartphone user
I have busy schedule and often jump
from one meeting to another
 Goals
I want to know how we perform on
our key initiatives (KPIs)
I want to have my information
summarized / visualized in one
dashboard and easy to navigate at
one-click
 Frequency:
Once per day
 Channels:
Phone
Laptop
Email / Slack / Sharepoint
 Tools: Embedded visual dashboard
Business Manager
•Type: BI Consumer
Characteristics:
I oversea the sales domain and
closely monitor operational processes
I have busy schedule and often go
form one meeting to another
Goals
I want to know how efficiently and
effective we perform within Sales
I want to have my information
summarized / visualized in one
dashboard and easy to navigate at
one-click
If I see some peaks in the trends, I
want to be able to make adhoc report,
slide/dice and drill the information to
the relevant level
I will use the BI glossary and use the
data definitions so I make my adhoc
report.
Frequency: Few times per day
Channels: Laptop, Phone
Tools: Excel / Power BI
Data Analyst / Scientist
•Type: BI Producer
Characteristics:
I'm a tech and data savvy and
working with data is my day to day job.
I have advanced data analyses and
statistics skills and subject matter
expertise for my domain.
I'm convenient working with scripting
languages (SQL, Python) and can
develop charts and visualize data.
Goals
In order to perform fast extensive
adhoc analyses I need reliable access
to the raw data sets.
I need to be able to communicate my
insights in easy and seamless manner
with my stakeholders.
In order to perform advanced
statistical analyses and ML modeling I
need reliable and performant
environment and tools.
I am continuously optimizing existing
and creating new dashboards and
reports.
Frequency: Continuously
Channels: Laptop
Tools: SQL, Python, Power BI, Excel,
Shell Programming

Strong understanding in the:
Product, business and operating
model
 Underlying IT architecture and data
flows
Data science approaches and
technologies used to solve typical
business and product problems (at
scale)
Analytics & Data PO Unicorn
 Concepts & Artefacts
 KPI Definitions & Glossary
 Data Domain Modeling
 Data Management
 Data Governance
 Data Privacy & GDPR
 Data By Design Principles
 Cloud Data Architectures

Analytics Stakeholders and Customers
Data &
Analytics Team
CxO
Marketing
Finance
Customer
Success
Data Analyst
Business
Development
Business
Development
Data Analyst
Product
Teams
Product
Management
Analytics Guild
Business Enabler &
Growth Function

Value, Outcomes & Usecases
Customer Success
• Goal: Reduce churn rates
Scenario: Targeted loyalty
campaigns for Customer with
high churn probability
Scenario 2: Account’s health
dashboard
• Goal: Improve operational
efficiency and customer
satisfaction
Scenario: Track and correlate
customer success processes
and efficiency with NPS and
CLV.
Product
• Goal: Facilitate Product
Discovery
Scenario: Collect and Analyze
CES and CSS feedback.
Scenario 2: A/B testing
• Goal: Product Backlog
Priorities
Scenario 1: OKRs - Measure
outcomes and adoption on
newly developed features.
Scenario 2: Identify the
correlation between user
actions and conversion rate
• Goal: Growth and
engagement
Scenario: Personalize user
experience based on the data
points (templates)
Scenario 2: Funnel
optimization
• Gola: Adoption and growth
Scenario: Insights and
Intelligence for the customers
of your customer
Marketing
• Goal: Improve budget
allocation,
Scenario: Optimize for
AdWords and focus on more
successful channels
• Goal: Increase conversion
rate
Scenario: Develop leads
scoring mechanism.
Observing leads who sign up
but do not subscribe
behavior and find those
patterns that signal
conversion.
Business Development
• Goal: Growth
Scenario: Focus efforts on
most valuable customers
Risk & Fraud
• Goal: Portfolio Management
Scenario: Understand which
loans are at most risk of
default
• Goal: Credit Scoring
Scenario: I want to
understand what is the risk
score of my prospect/existing
customer so I would know
what type of product I can
offer and at what price and
interest rates
• Goal: Revenue Protection
Scenario: Understand
patterns of fraudulent
behavior to protect company
revenue and customers
wallets.

Why Do I Need Analytics Strategy?
Q&A

Analytics & Data Strategy 101 by Deko Dimeski

Recommended

More Related Content

What's hot (20)

Similar to Analytics & Data Strategy 101 by Deko Dimeski (20)

Recently uploaded (20)

Analytics & Data Strategy 101 by Deko Dimeski