Business Analytics Capstone

Framework for Strategy

16/04/2019

Problem Statement
Adblocking software and evolving user behavior is disrupting
online advertising practice.
Problem Statement–
Describe the Problem Adblockers present to GYF

• Use this space for describing the problem. Be as specific as possible! You should focus on the implications of adblockers on GYF’s ad-
buying customers; in exploring
this, you might also consider the implications for GYF’s end users, operations, and/or internal organization.
In the recent years, two main phenomena have emerged – the changing behavior in our users
(spending more time on mobile than on desktop), and the proliferation of new technologies to
enhance user experience (i.e. adblocking software has become increasingly popular); these
factors are perceived to have a direct influence on GYF’s main source of revenue – selling
advertising to other companies.

End user – The end user group must be understood, so thorough analyses have to be
conducted – including trending behavior, projected demand, qualitative and quantitative
research.

Operations – Are there any problems with the current operations or optimization techniques
to increase operational effectiveness?

Internal Organization – Should resources be allocated to R&D and implementation to address

this problem, if yes, how much and what is the estimated monetary value of the anticipated
prescriptive solutions?
Problem Statement–
Application Exercise 1 – Research Methods and Tools (Optional)

• Use this space for to answer the questions set out in Application Exercise 1: 1. Given your definition of the problem faced by GYF,
what type(s) of research will you employ to learn more about the strategy the DATA Team should pursue? 2. What research tools
could you use to conduct that research?

Step 1 – Exploratory Research

Undertake exploratory research to collect data, so that the problem statement can be better defined. This would
mean to design an unobtrusive survey for the current GYF user base and/or engage third party analytics companies to
undertake research on its competitors such as Yahoo, Google & Facebook. (note: 70% user activity on mobile, hence
it would be ideal to deploy in-app or in-browser surveys with some reward incentive on completion) Depending on
the research budget, other tools like scanner data, social media data, etc. can be collected, pruned, to be included
into the research.

Step 2 – Descriptive Research

After the Exploratory Research is completed, run models to formally review the results. The results will better inform
our understanding of the problem objectively, as to avoid subjective biasness or intuition-like statements. Use models
like Moving Averages, Seasonality Studies to understand the trending behavior over time.

Step 3 – Causal Research

At this stage, we can create some hypotheses formulated from the Descriptive Research, and run experiments within a
controlled environment to ensure that causal statements are proven (i.e. if A happens, then B happens), and to avoid
the ‘correlation-causation’ problem.

Strategy
Strateg
y
Describe your proposed strategy

• Make sure your strategy is clear, well-defined, and feasible

A starting point may be to define the Approach (Traditional, Adaptive, or Visionary)

Once the Approach has been determined, and based on the Analytics undertaken, stipulate prescriptive actions to be executed Some
examples include,
- Allow only ‘delightful ads’
- Optimize for native advertising only
- Create incentives for attention from customers

Once the possible outcomes are determined, conduct surveys and experiments to accurately measure the outcomes of each iteration
before ranking the solutions. Be mindful to estimate the cost of each solution for the expected monetary return in value. Decision trees
and simulation exercises should be undertaken.

Other considerations:
Given the emerging behavior of increase mobile activity compared traditional desktop interaction, resource allocation should be skewed
in the favor of adblocking strategies on the mobile environment.

How will these strategies differ and affect all products such as GYF Search, Mail, Chat, and Digital Media?

What are other competitor companies doing?

Strateg
y
Application Exercise 2 – Hiring a Team Leader (Optional)
• Based on the information provided in Application Exercise 2, describe here why you would hire Carrie Candidate or Peggy Prospect
to be the second-in-command of the DATA Team.

Purely based on the information provided, the following decision-basis is as follows:

The most correlated factors to performance outcomes such as ‘Work Samples’, ‘Cognitive Ability Tests’, and ‘Structure Interviews’,
were weighted strongly during the review process. Based on the model below, Carrie Candidate was selected for the job, as she fared
better on a weighted average when compared against Peggy Prospect.
Effects and Measurement
Effects
Describe the anticipated effects of your strategy
• Please describe the anticipated effects of your strategy. Make sure you address the effects on customers, revenue, and the internal
organization.

1. Anticipated Effects of Strategy on CUSTOMERS

Customers should have an improved user experience, and be incentivized to increase interaction with the GYF platforms. Increased screen
time from customers is a likely outcome.

2. Anticipated Effects of Strategy on REVENUE

The business should experience an increase in REVENUE from advertising
Mindful that PROFIT may not be realized at the early stages as cost for planning, reallocation of resources, implementation may increase,
resulting in negative cashflows.

3. Anticipated Effects of Strategy on INTERNAL ORGANIZATION

The business may experience a shortage of cash due to new strategy implementation and reallocation of resources.
The staff may feel overwhelmed at the beginning as new experts/professionals are hired to implement new strategy, so some time for
readjustment and refocusing is required.
Effects
Application Exercise 3 – Designing a Deterministic Optimization Model
• Provide an explanation of the calculations you performed to build an optimization model using decision variables, constraints, and
an objective; this model could use the scenario in Application Exercise 3, or one of you own devising.

1. Building the Optimization Model – DECISION VARIABLES > SOLUTION > OBJECTIVE
2. Also define the following – OBJECTIVE FUNCTIONS & CONSTRAINTS
3. Formulate the ALGEBRAIC MODEL to undergo SOLVER to find an Optimized Solution
Measurement
Describe the anticipated effects of your strategy and how you will measure them
• Please outline your plan for measuring these effects using data. Make sure you use techniques you learned about in the courses
1. Plan for measuring effects
Collect both quantitative and qualitative data as each strategy is being rolled out
Ensure that the strategy is always in line with the original problem statement
Crucial data to be collected include customer experience measures, screen time, and ad revenue
Define linkages and verify the causality of these measures
Ensure that organizational issues are addressed during the whole process

2. Techniques applied
Regression Models and working backwards to define causality
Prune noisy data, identify persistence, draw upon large samples where possible, identify independent

signals Ensure that the goal has been defined early on, and techniques are applied directly to achieve the

goal.
Measurement
Application Exercise 4 – Identifying Key Drivers
• Apply the “causal business model” performance measurement framework to your strategy

1. Identify at least one hypothesis that is explicitly linked to your strategy in a causal way

Curating and allowing only ‘native ads that are delightful’ increases quality of ads, in turn increasing user’s experience and directly
contributes to revenues for advertising companies and GYF.

By rewarding ‘GYF Points’ as an incentive to ad-watchers, creates a loyalty bonus for user interaction with the GYF ecosystem, therefore
increasing revenues along different business streams.

2. Identify at least one key driver and explain how that driver will be measured as your strategy is being implemented

Quantify and categorize the change in the following,

- Native ads
- Native ads that are delightful
- Non-native ads
- Non-native ads that are delightful
With all other conditions maintained constant, measure the screen time against each category and corresponding ad-revenue The
next phase, would be to roll out a ‘loyalty points’ program and measure the change in screen time, as well as corresponding
adrevenue accordingly.

3. Describe how you would go about verifying the linkage between the first two steps
Conduct surveys for each implementation in order to collect qualitative data from the users – if the data collected concurs with
the hypotheses, it should verify the linkages of ‘increase user experience’ > ‘increase screen time’ > ‘increase revenue’. Ensure that
data fiefdoms and organizational intuitions are kept independent of analyses.

Conclusion
• Summarize your key points from the preceding slides

1. Problem Statement
Exploratory, Descriptive, and Causal Research to be undertaken to define a robust and deterministic Problem Statement.
Collect as much data as possible and generate models to study the problem and its peripherals in order to adequately understand the
context of the problem. The problem statement should suggest the outcome and goal of the proceeding stage.

2. Strategy
Determine the business approach before deciding on the strategy.
Outline each strategic initiative in accordance to the problem statement and all research undertaken.
Undertake a optimization modelling to maximize effectiveness of each implementation.

3. Effects & Measurement

Ensure that the effects of the implementation of strategy can be directly measured for future deployment and scaling.
The outcomes may not be favorable all the time – hence re-strategizing or re-optimization of the strategy may be required, redeployed,
and remeasured. This process may repeat until the outcomes become favorable and addresses the original problem statement directly to
achieve the overall business objective (i.e. maximize revenue, minimize expenses, outperform competitors etc.)