I don t know where doritos live

Brand Strategy and Super Bowl Twitter Analytics

Image Source: Adweek.com

Twitter popularity as 2nd screen

Infographic Source: Fortune, 2015

About Twitter

3

Twitter Stats (cont’d)

4

About Super Bowl Ads

5

Motivation • Cost per 30 seconds slot – 5 mil (CNNMoney, 2017). • Brand strategy in Super Bowl ads? Enhance & reach.

• The role of IS tools in this strategy? – Analytics of audience (demographics, locations etc.). – Supplementary info for future decision making.

• What is the relationship between Social Media buzz around Super Bowl ads with ad ratings? – Twitter activities surrounding SB ads with USA Today Index ratings (USA Today)?

3 Tweet Examples Those Budweiser commercials had me balling my eyes out!! #Budweiser #SuperBowl #horsepuppy #tearjerker

Thanks #CocaCola & #Cheerios for showing U.S. multicultural families and successfully including diverse markets #adbowl #AmericaIsBeautiful

Scarlett Johansson should realize that the only real flavor of #SodaStream is oppression http://t.co/QLpDD7vkXA #superbowl

Objective • Learn to extract valuable metrics from social media data using MS Excel.

• Know what insights are valuable to brands. • Simple introduction to (review on) descriptive statistics, correlation, charts, regressions, and word clouds.

• Lab instructions become more vague as lab progresses to encourage student self learning.

8

Lab Setup • Download this presentation. • Download the SB 2014 spreadsheet ‐‐ This is a summary of actual tweets *downloaded using PHP and MySQL scripts (done prior).

• Download the doritos.txt tweet text file*. • You will need MS Excel with Data Analysis ToolPak.

• Prepare a Word doc for submission.

Introduction • Start with research question:

– Is social media (Twitter) an effective tool for measuring brand performance (SB ads)?

• How? – Find a relationship between Twitter metrics and SB performance.

– If a relationship exists then social media may be a valuable (real time, low cost) monitoring mechanism in addition to existing tools.

– In addition, social media provides rich feedback (WOM, influencer, competitors, etc.).

Analysis 1. Descriptive Statistics 2. Correlation Analysis 3. Charts (Bar chart and scatter plot) 4. 3 Sets of OLS Regressions 5. Word Cloud diagrams

SB 2014* Dataset

• Super Bowl 2014 Ads • 51 Ads • Contains social media measures and USA Today ratings for each ad.

• Measures extracted from downloading and analyzing tweets for each ad.

*Seattle Seahawks beat Denver Broncos (43‐8) 12

Descriptive Statistics 1. Open MS Excel 2. Open SB 2014 spreadsheet 3. Highlight all the number cells (including titles) but

excluding official# and brand columns. 4. Go to Data > Data Analysis

Descriptive Statistics 5. Select Descriptive Statistics. 6. Select the shown choices below.

Descriptive Statistics 7. Results appear in next Sheet. 8. Copy Result to Word doc.

Descriptive Statistics 9. Discuss result. 10.Focus on average, min, max for

– USA Index – total tweets – brand only tweets – ad only tweets

Correlation Analysis 1. Select Correlation then OK.

2. In the Correlation window click OK.

Correlation Analysis 3. Results will appear in next Sheet. 4. Discuss which variables are highly correlated. Why? 5. Copy Result to Word doc.

Charts (1) • Do a Bar chart of ‘ad tweets’ and discuss. Add the brand names.

Charts (2) • Do a Scatter Plot of ‘Ad Tweet with USA Index’ and discuss. Add the trend line. What does the trend imply?

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

0.00 2000.00 4000.00 6000.00 8000.00 10000.00 12000.00

USA Index

What is linear regression? • Linear regression is the most basic and commonly used predictive analysis. Regression estimates are used to describe data and to explain the relationship between one dependent variable (y) and one or more independent variables (x).

• At the center of the regression analysis is the task of fitting a single line through a scatter plot. The simplest form with one dependent and one independent variable is defined by the formula y = a + b*x. 21

3 Sets of Regression • 3 Sets with the following as Y variable

– USA Index – Total Tweets – Brand Tweets

• Each set should have 3 regressions (i.e. version 1,2,3).

• You should end up with a total of 9 regressions.

22

1. Select Data Analysis and run regression. 2. Select the Y (USA Index) and X (all others except

#official and brand). Then click OK.

Regression v 1.0 (USA Index)

3. Copy to Word doc. Regression v 1.0 (USA Index)

4. Discuss Results. What variables can explain Y? You are expected to use your knowledge from your stats class here.

5. What to report? – Adjusted R‐squared – Variables with p value < .10

Regression v 1.0 (USA Index)

1. Copy main sheet to another sheet. 2. Select Y (USA Index). 3. For X, select only brand, ad, total, celebrity,

and pre‐release. Delete other columns. 4. Run regression. 5. Click OK.

Regression v 2.0 (USA Index)

7. Discuss Results. Copy to Word doc. Regression v 2.0 (USA Index)

1. Copy main sheet to another sheet. 2. Select Y (USA Index). 3. For X, select ad, celebrity, and pre‐release.

Delete other columns. 4. Run regression. 5. Click OK.

Regression v 3.0 (USA Index)

7. Discuss Results. Copy to Word doc. Regression v 3.0 (USA Index)

1. Perform another set for ‘total tweets’ as the Y variable.

2. Perform another set for ‘brand tweets’ as the Y variable.

3. Report result to Word doc and discuss. 4. Remember there should be 3 sets of 3

regressions totaling 9 runs.

More Regression Runs

Word Cloud 1. Go to http://www.wordclouds.com/ 2. Upload the ‘doritos’ tweet text file. 3. Discuss word cloud. 4. Why are certain terms

more frequently used?

Summary Paragraph • Write a summary paragraph for lab. • Describe what you have learned. • What IS tools are used? • How does this lab help you to understand the relationship between organizational strategy and IS tools?

• Can this be applied in your future career?

32

Lab Submission • In a Word doc:

1. Descriptive Statistics 2. Correlation 3. 2 Charts 4. 3 Regression Result Sets (USA Index, Total tweets, and

Brand Tweets) [9 total regressions] 5. Word Cloud 6. Summary paragraph

• Organize well to help easier grading == higher scores. • Submit on canvas

END