Group06 proposal

Growth Signal

Proposal

Poster

Application

Research Paper

Problem & Motivation

Rapid advancement of technology over the past few decades have brought forth many opportunities and convenience to mankind. In the world of businesses, especially marketing, it is extremely important to use various numbers and metrics to quantify and evaluate the standpoint of the company. This is made possible with the discovery of big data, analytics, AI/ML and the likes of data visualisation.

"If you can't measure it, you can't manage it." — Peter Drucker

In this era, many companies in the marketing department rely on platforms such as Google Analytics to track their businesses. However, problem arises when they advertise across multiple channels, such as Facebook, Instagram, email, and possibilities are endless. Decentralisation is one problem. Ironically, although Google Analytics houses large amount of data, oftentimes they do not provide useful and actionable insights relevant to the company. Hence, users are overloaded with the wide variety of data available that are reflected by a vast number of charts of which some can be complex and nugatory (i.e. chartjunks). To save time (sieving through chartjunks) and boost marketer's productivity, we aim to develop a data visualisation dashboard that serves as a one-stop solution where users can derive knowledge and uncover insights required to judge the company's success, and whether they are on the right track. This will thus help them drill down to the root causes and take necessary actions.

Objectives

Target Group: eCommerce Business Owners or Digital Marketing Exectives

With abundance of data and the lack of insights, our team would create dashboards to enable eCommerce Business Owners or Digital Marketing Executives to answer the following questions:

• How is my digital business performing compared to previous period?
• Are customers stuck at a particular buying stage in their customer journey?
• How much sales are attributed to my marketing efforts?
• Which channels most profitable?
• Where should i optimise my marketing spend today?

Dataset

Data/Source	Variables/Description	Rationale & Methodology
Google Analytics Data from Client's Account (1 Jul 2016 to 1 Mar 2020) Livestream capable	Users New Users Sessions Number of Sessions Per User Page Views Pages / Session Average Session Per User Page Views Pages / Session Avg Session Duration Bounce Rate Conversion values Number of Users all at Stages of Customer Journey	This is a clickstream dataset which provides strategic data from all traffic sources. Non exhaustive list.

Background Survey

Research	Learning points	Improvement
Source: https://www.tapclicks.com/resources/blog/6-steps-marketing-dashboard/	Example of sales funnel (top right) on the to show conversion of Users to Leads and Trial. Can be broken down into timeline as well. The sales funnel would give an accurate idea of how many percentage of users are really interested in the advertisement or those that actually manage to secure leads. However, diagrams in 3D can be misleading due to overlaps and is not very neat. How many leads in the last 90 days or this month can be combined.	Can add Interaction for across a time period. Funnel can be broken down into more categories such as number of potential leads, marketing leads, sales leads, deals etc. Can give top 5 sources of traffic
Source:https://www.klipfolio.com/resources/dashboard-examples/marketing/digital-marketing-dashboard	Another dashboard example provided by kilpfolio. Useful example to show metrics such as quality of leads and conversion metrics, leads by region and top few social media channels. It can be animated or add a scale to show changes across timing However, leads by region is too small. Google web analytics Web traffic targets is not labeled and just shows the bar. Can change to label with highlight of red and green (based on whether the progress is negative or positive)	Perhaps can show the top 5 areas instead and table can be replaced by a bar graph. Key conversion metrics and quality of leads can be replaced with line graph to show changes over time.

References

• 6 Steps to Creating a Better Marketing Dashboard
• Klipfolio - Digital Marketing Dashboard
• 100 Tasks
• Web Analytics 2.0 by Avinash Kaushik

Technical Challenges

Challenge	Description	Mitigation Plan
Unfamiliarity with Tools	Most of the team members are not too familiar with R and R Shiny.	Make use of external sources for information and troubleshooting Utilise DataCamp to learn, practice and brush up skills on R and R Shiny
Lack of Marketing Domain Expertise	There may be gaps in terms of marketing jargons and concepts we are unaware of or fail to pick up in such a short time span.	Learn and apply the best practices from Web Analytics 2.0 by Avinash Kaushik Consult with client for feedback on the dashboard
Data Quality & Integrity	As Google Analytics uses clickstream data, there exist a possibility of bot traffic registering as data points. Much time is also needed to clean the data before feeding it into the visualisation.	Conduct proper EDA and data cleaning methods to minimise quality problems Apply filters and follow best practices from the industry to reduce data abnormalities

Storyboard

Dashboards	Description
Dashboard 1: Key Metrics Performance	Key metrics performance indicators would provide executives on the pulse of the business. A dropdown box at the top left would allow the selection of time period and metric the graph would represent. Instead of providing a value for the metric per time period, Sparklines and percentage changes from the previous time period would show how the metric changed.
Dashboard 2: Sales Funnel Analysis with Segementation	Sales funnel analysis would allow executives to show volume of users at each part of the customer journey. Percentage changes would show the difference between time period, this would allow executives to identify which part of the sales funnel is generating the most friction for buyers. Upon clicking or selecting from a dropdown menu, there would be options to filter via traffic source (Organic, Direct, Paid Social etc). This would allow executives to determine the best traffic source that leads to sales and optimise their marketing spend. Our team is still exploring the possibility of adding more filters such as geographic location or demographics to allow more insights.
Dashboard 3: Customer Acquisition Cost vs Customer Lifetime Value	The last dashboard would be the show the cornerstone of all businesses. Would their business survive in the long run? Are they spending too much to acquire a customer? How long would it take for their customer acquisition cost to break even? Some calculation for context: Customer Acquisition Cost (CAC): How much does it cost to acquire a customer CAC = Total Acquisition Cost / # of Customers Acquired Customer Lifetime Value (CLV): How much does a customer bring over the lifetime CLV = Total Acquisition Cost / # of Customers Acquired Filters per time period and channel would allow further segmentation.

Project Timeline & Milestones

Detailed Project Timeline from Asana (team's task tracker)

Comments

Name	Date	Comments
(Name)	(Date)	(Comment)
(Name)	(Date)	(Comment)
(Name)	(Date)	(Comment)