predictive analytics
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
1|Page
PREDICTIVE ANALYTICS FOR RETAILERS AND MANUFACTURERS
THE ANALYTICS GROUP, LLC
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
TABLE OF CONTENTS
PG. 3
ANALYSIS GOALS DATA SET OVERVIEW & SCOPE QUESTIONS WE ASKED
PG. 4
DATA COLLECTION & EXPLORATION
PG. 5
PREDICTIVE ANALYTICS REPORTS REPORT #1 – UNIQUE PREDICTORS REPORT #2 – CUSTOMER DEMOGRAPHICS
PG. 8
TECHNOLOGY & OTHER NOTES
PG. 9
CONCLUSION & RECOMMENDATIONS
2|Page
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
3|Page
ANALYSIS GOALS SAMPLE DATASET: Our data is from a marketing campaign for a retail company selling apparel. The company plans to launch a new clothing line for men and women in their late 30’s and early 40’s. They conducted a week-long social media ad campaign to gauge market response to their new products. Using the data collected, our goals were:
Identify customer behavior that can be used as predictors for engagement. Compare the market response to the targeted customer demographics. Uncover insights to help creative team, sales and marketing to improve conversions.
ANALYTICAL PROCESS: 1 . Data Collection 2 . Data Exploration 3 . Predictive Analytics 4 . Data Visualization
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
4|Page
DATA COLLECTION & EXPLORATION DATA COLLECTION: Our analytics platform is harnessed to the company’s ad campaign and collects and loads data across many different metrics.
DATA EXPLORATION: The platform automatically aggregates and organizes the data according to common starting points. For instance: calculating the cost of acquiring a customer via social media (the relationship between cost per click and amount spent). It also identifies trends within the data and suggests questions based on its perspective of trends and variances within the data. The software can also understand and respond to natural questions, such as, “What is the breakdown of customer spending by time of day?”
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
5|Page
PREDICTIVE ANALYTICS REPORT #1 This screenshot displays a report of which metrics have the strongest correlations with customer spending. Our goal is to isolate which metrics the company can focus on to drive higher customer engagement and spending. Note that in the top left, the software assigns a numerical score to the data quality (in this use case, 74/100) in order to mitigate poor data quality and highlight potential data weaknesses.
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
6|Page
BUSINESS INTELLIGENCE INSIGHTS: 1. Unique Clicks: A unique click is a 62.2% predictor of amount spent. This means that the highest-spending customers make only one visit to the site. o Insight: This is a trend the company should reverse. The majority of highspending (profitable) customers are not returning to make multiple purchases, which means the company is failing to build customer loyalty. o Action: We can collect targeted data from future campaigns to discover where the issue lies: poor online customer experience, lack of follow-up, poor loyalty incentives, or disconnects in reaching the customer base. We can measure this through data such as website bounce rate, demographical data, and response to loyalty programs. 2. Age & Gender: The company is targeting each piece of apparel towards a particular gender and age, so we expect these numbers to correlate closely. However, age and gender are only the #4 and #5 predictors of amount spent, respectively. This suggests a possible disconnect between the company’s ad and their targeted audience o Insight: There should be a strong correlation between spending and age/gender, but there isn’t. The ad is not engaging the targeted audience. o Action: We’ll run a report to isolate potential reasons. This may connect with the negative trend isolated in the first insight.
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
7|Page
PREDICTIVE ANALYTICS REPORT #2 This screenshot displays a report of age and gender breakdown. Our company is targeting their ads and apparel towards men and women in their late 30s and early 40s, so we should expect the biggest spending to occur in that age range.
BUSINESS INTELLIGENCE INSIGHTS: 1. Age: The biggest spending for both genders occurred in the 25-34 age range (blue), whereas the targeted age range was 35-44 (green). o Insight: The customers that spent the most money were outside the targeted age range. The company is not hitting its target age range. The ad may be running for too many customers outside the target age range, or the content may not appeal to the intended demographic. o Action: In follow-up campaigns, target ads more accurately by age range. Test content marketing to boost appeal to the targeted age range. 2. Gender: Female customers spend more than male customers. o Insight: Spending trends in this campaign match overall gender spending trends. There is no problem in this aspect.
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
8|Page
TECHNOLOGY Our experts specialize in the following technologies and many more:
Databases Management Systems : Microsoft SQL Server Oracle SAP DB2 Teradata PostgreSQL ETL & OLAP : Cognos Business Objects Informatica Data Warehousing : SSIS SSAS Cassandra Vertica Netezza Reporting & Analytics: SSRS Pentaho Dundas OBIEE QlikView Tableau Watson Analytics Google Analytics Programming : .NET Java Web : AngularJS knockoutJS Bootstrap JSON CSS JSP ASP.NET WEB API HTML We also offer : SAP R3 NoSQL e-mail marketing & analytics e-commerce cloud integration digital assistance
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
9|Page
CONCLUSIONS & RECOMMENDATIONS YOUR NEXT STEPS 1. Set up a conversation with us. 2. Take a hard look at what data you have and what questions you want it to answer. 3. Think about your goals for your business intelligence, and your company as a whole. 4. Come with questions.
ABOUT US The Analytics Group provides consulting, professional services, and end-to-end solutions. We specialize in predictive analytics for retail and apparel firms. In addition to predictive analytics, we provide solutions in data management, business and data analysis, change enablement, business process improvement, and more. Our headquarters are based in New Jersey, with offices in New York City, Florida, and India. Phone:
(732) 400-5521
Email:
[email protected]
Address:
1 University Plaza Dr, Suite 624, Hackensack, NJ 07601
1|Page
PREDICTIVE ANALYTICS FOR RETAILERS AND MANUFACTURERS
THE ANALYTICS GROUP, LLC
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
TABLE OF CONTENTS
PG. 3
ANALYSIS GOALS DATA SET OVERVIEW & SCOPE QUESTIONS WE ASKED
PG. 4
DATA COLLECTION & EXPLORATION
PG. 5
PREDICTIVE ANALYTICS REPORTS REPORT #1 – UNIQUE PREDICTORS REPORT #2 – CUSTOMER DEMOGRAPHICS
PG. 8
TECHNOLOGY & OTHER NOTES
PG. 9
CONCLUSION & RECOMMENDATIONS
2|Page
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
3|Page
ANALYSIS GOALS SAMPLE DATASET: Our data is from a marketing campaign for a retail company selling apparel. The company plans to launch a new clothing line for men and women in their late 30’s and early 40’s. They conducted a week-long social media ad campaign to gauge market response to their new products. Using the data collected, our goals were:
Identify customer behavior that can be used as predictors for engagement. Compare the market response to the targeted customer demographics. Uncover insights to help creative team, sales and marketing to improve conversions.
ANALYTICAL PROCESS: 1 . Data Collection 2 . Data Exploration 3 . Predictive Analytics 4 . Data Visualization
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
4|Page
DATA COLLECTION & EXPLORATION DATA COLLECTION: Our analytics platform is harnessed to the company’s ad campaign and collects and loads data across many different metrics.
DATA EXPLORATION: The platform automatically aggregates and organizes the data according to common starting points. For instance: calculating the cost of acquiring a customer via social media (the relationship between cost per click and amount spent). It also identifies trends within the data and suggests questions based on its perspective of trends and variances within the data. The software can also understand and respond to natural questions, such as, “What is the breakdown of customer spending by time of day?”
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
5|Page
PREDICTIVE ANALYTICS REPORT #1 This screenshot displays a report of which metrics have the strongest correlations with customer spending. Our goal is to isolate which metrics the company can focus on to drive higher customer engagement and spending. Note that in the top left, the software assigns a numerical score to the data quality (in this use case, 74/100) in order to mitigate poor data quality and highlight potential data weaknesses.
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
6|Page
BUSINESS INTELLIGENCE INSIGHTS: 1. Unique Clicks: A unique click is a 62.2% predictor of amount spent. This means that the highest-spending customers make only one visit to the site. o Insight: This is a trend the company should reverse. The majority of highspending (profitable) customers are not returning to make multiple purchases, which means the company is failing to build customer loyalty. o Action: We can collect targeted data from future campaigns to discover where the issue lies: poor online customer experience, lack of follow-up, poor loyalty incentives, or disconnects in reaching the customer base. We can measure this through data such as website bounce rate, demographical data, and response to loyalty programs. 2. Age & Gender: The company is targeting each piece of apparel towards a particular gender and age, so we expect these numbers to correlate closely. However, age and gender are only the #4 and #5 predictors of amount spent, respectively. This suggests a possible disconnect between the company’s ad and their targeted audience o Insight: There should be a strong correlation between spending and age/gender, but there isn’t. The ad is not engaging the targeted audience. o Action: We’ll run a report to isolate potential reasons. This may connect with the negative trend isolated in the first insight.
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
7|Page
PREDICTIVE ANALYTICS REPORT #2 This screenshot displays a report of age and gender breakdown. Our company is targeting their ads and apparel towards men and women in their late 30s and early 40s, so we should expect the biggest spending to occur in that age range.
BUSINESS INTELLIGENCE INSIGHTS: 1. Age: The biggest spending for both genders occurred in the 25-34 age range (blue), whereas the targeted age range was 35-44 (green). o Insight: The customers that spent the most money were outside the targeted age range. The company is not hitting its target age range. The ad may be running for too many customers outside the target age range, or the content may not appeal to the intended demographic. o Action: In follow-up campaigns, target ads more accurately by age range. Test content marketing to boost appeal to the targeted age range. 2. Gender: Female customers spend more than male customers. o Insight: Spending trends in this campaign match overall gender spending trends. There is no problem in this aspect.
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
8|Page
TECHNOLOGY Our experts specialize in the following technologies and many more:
Databases Management Systems : Microsoft SQL Server Oracle SAP DB2 Teradata PostgreSQL ETL & OLAP : Cognos Business Objects Informatica Data Warehousing : SSIS SSAS Cassandra Vertica Netezza Reporting & Analytics: SSRS Pentaho Dundas OBIEE QlikView Tableau Watson Analytics Google Analytics Programming : .NET Java Web : AngularJS knockoutJS Bootstrap JSON CSS JSP ASP.NET WEB API HTML We also offer : SAP R3 NoSQL e-mail marketing & analytics e-commerce cloud integration digital assistance
PREDICTIVE ANALYTICS USE CASE: RETAIL & MANUFACTURING
9|Page
CONCLUSIONS & RECOMMENDATIONS YOUR NEXT STEPS 1. Set up a conversation with us. 2. Take a hard look at what data you have and what questions you want it to answer. 3. Think about your goals for your business intelligence, and your company as a whole. 4. Come with questions.
ABOUT US The Analytics Group provides consulting, professional services, and end-to-end solutions. We specialize in predictive analytics for retail and apparel firms. In addition to predictive analytics, we provide solutions in data management, business and data analysis, change enablement, business process improvement, and more. Our headquarters are based in New Jersey, with offices in New York City, Florida, and India. Phone:
(732) 400-5521
Email:
[email protected]
Address:
1 University Plaza Dr, Suite 624, Hackensack, NJ 07601
