MSDM Culminating Experience Project (aka Capstone Project)
Tips for Maximizing Value for Your Career
Introduction
This page is intended to accumulate important information for MSDM students as they embark on the capstone project.
1 What to Consider When Choosing a Project (Early Fall)
To get the best out of MSDM Culminating Experience, consider multiple factors and make an informed decision.
1.1 Contribution of the Project (Basic research vs. Applied Research)
1.2 Characteristics of Project Sponsor (Faculty vs. Students)
1.3 Project’s Alignment with your Career Objective (Digital Marketing Strategy vs. Marketing Science)
1.4 Availability of Data
1.5 Type of Data: Primary data vs. Secondary Data
1.6 Sophistication of Data
1.7 Nature of AO: Exploratory vs. Confirmatory
1.8 Interest in the Topic
1.9 Interest in the Project Objectives
2 Recommendations Relevant for the Entire Project Journey (Fall, Spring, & Summer)
2.1 Make Yourself Vulnerable
Share your entire project cloud folder—not just a few files—even if your faculty lead has not requested it yet. This transparency helps your group regulate contributions and maintain accountability.
Create folders labeled with each team member’s name to show individual contributions clearly.
2.2 Use a Project Management Tool
To plan activities and track milestones effectively, you should use a project management tool that keeps everything in one place. If you already have one, continue using it. If not, Notion is highly recommended.
For an extensive review of the project management tools and factors to consider, click here.
3 Tips for Preparing for Successful Project Proposal (Fall Semester)
3.1 Secure Feedback From Your Faculty Project Lead
Share your written project report with your faculty lead even if it still needs improvement. They can give you preliminary feedback early on, giving you time to revise before the final submission for grading. Remember that your faculty lead’s assessment will influence the grade you receive for IBM 6200.
3.2 Project Objectives (PO) vs. Analytics Objectives (AO)
Clients often have expectations and deliverables that do not fully align with MSDM CEP requirements. For this reason, it is important to distinguish clearly between project objectives (PO) and analytics objectives (AO).
POs often include broader marketing or business outcomes—such as creating marketing content (e.g., designing a website or social media posts) or increasing conversions (e.g., customers, leads, sales). These are the goals of digital marketing activities.
AOs, in contrast, must be directly tied to data, because analytics objectives guide the analysis you will present in the report.
One way to identify whether an objective is a PO or AO is to ask:
Is there data associated with this objective?
Is the same dataset used for other analytics objectives relevant here?
If the answer is no, you are likely dealing with a project objective, not an analytics objective.
Another way is to ask yourself what method (descriptive statistics, inferential statistics, predictive analytics) you would describe in Chapter 3 (Methods).
For example: Is there any statistical method required to design a website?
- The answer is no—because designing a website is a project activity, not an analytics task.
3.3 How to Convert PO to AO
When an objective is not related to data or methods, you will have little to write about in Chapter 4 (Analysis and Results). Therefore, you must translate marketing or business objectives into analytics objectives.
A useful guiding question is:
What information do I need, and where can I find it, in order to achieve the project objectives?
Suppose the PO is to design a website to engage potential clients and increase conversion rates. For the website to be effective, it must follow UX principles, conversion-centric design, and SEO best practices.
To evaluate performance, you can collect visitor behavior metrics from Google Analytics 4 (GA4), such as:
bounce rate
session duration
conversion rate
number of visitors
These metrics allow you to identify the website’s strengths and weaknesses.
Once you know what data you need and where it comes from, you can determine which analytical methods to apply. The next question becomes:
What should I do with the data to generate insights that will help improve website performance?
You may also have expectations about relationships among metrics.
For example:
- High bounce rate on landing pages → likely low conversion rate
(a negative relationship)
You should also consider what counts as an acceptable bounce rate, session duration, or conversion rate. Comparing your metrics with industry benchmarks will help you identify areas for improvement.
Following the logic shown in the example, you can lead to the following conclusion.
PO1: Design a website to communicate results and engage potential client brands.
AO1a: Explore ways in which the website can be optimized for potential customers.
Of course, AO above could have been expressed in different ways.
AO1b: Generate insights from the GA4 and Google Ads data for website optimization
AO1c: Generate insights from the GA4 and Google Ads data to increase visitor engagement and conversions.
Consider multiple options and settle with one that your team likes the best. A few criteria to consider are the clarity and concreteness of the statement.
Using these criteria, you may choose AO1C as your formal AO1.
3.4 Data Collection for the Above AO
3.4.1 Preparation: Owned Media
To address this AO, you need to collect KPIs through GA4. If no website exists, you must:
create one,
connect it to GA4,
link Google Search Console to GA4, and
use UTM parameters to track email and social media traffic.
If this preparation has not been completed, it should be done as part of Chapter 3 (Methods). At a minimum, you should complete this as soon as possible so you have at least one month of accumulated GA4 data before the spring term begins.
Be sure to describe your digital marketing infrastructure and data collection plans in Chapter 3.
3.4.2 ETL Tools for GA4 and Google Ads Data
During the winter, you should set up the necessary own media (website, social media, email, Google Ads) and connect them to GA4.
Once GA4 and Google Ads are operational, you will have data for spring analysis.
During the spring, you can:
use GA4/Google Ads dashboards for descriptive statistics, or
import GA4/Google Ads data into Google BigQuery for greater data control, inferential analysis, and predictive modeling.
Inside BigQuery, you can also use Google Looker Studio for custom visualizations and Gemini AI for modeling. If you prefer working in R/RStudio, R can connect directly to BigQuery. Packages like bigrquery allow your dplyr syntax to be translated into SQL automatically.
bigrquery provides a database backend that dplyr can talk to using the dbplyr translation engine.
You never write SQL—dbplyr generates and sends the SQL for you.
Below is an illustration of codes.
library(bigrquery)
library(dplyr)
con <- dbConnect(
bigquery(),
project = "your-project-id",
dataset = "your_dataset"
)
tbl(con, "your_table") %>%
filter(bounce_rate > 0.5) %>%
group_by(device_category) %>%
summarize(avg_session = mean(session_duration))3.5 Presentation Rehearsal
The CCIDM advisory board has generously offered to help you with your presentation. If you would like feedback, you may record your rehearsal and send it to me, and I will forward it to the board.
The board members are very busy, but they are eager to support you—presentation skills are essential in any organization. One board member is even looking to hire interns, so this could be a valuable opportunity.