As part of AirFrance's marketing team, we had to develop an SEM marketing campaign in order to maximize ROA and increase ticket sales. After collecting data from “Double-Click” on previous marketing campaigns and assessing key KPIs, we analyzed the different search engines that AirFrance worked on within July 2007. Then, we used R Studio to execute some descriptive and predictive analysis and come up with useful business insights and recommendations.
- R MARKDOWN -
#DESCRIPTIVE ANALYSIS
From the #Total Income Pie Chart, which is Total Income for Advertising - Google US has the highest value with 35.6%, Google global has 19%, and third is Yahoo. MSN is the least.
From the #Total Cost Pie Chart, which is Total Cost for Advertising - Google US has the highest cost with 46.6%, followed by Google Global, and Overture. MSN has ranked with the lowest cost
Hey there, this is the default text for a new paragraph. Feel free to edit this paragraph by clicking on the yellow edit icon. After you are done just click on the yellow checkmark button on the top right. Have Fun!
From the #Total Clicks Pie Chart, we can see that Google Us has the highest clicks, followed by Overture US and Google Global. MSN has the lowest total of clicks.
From the #Average Cost per Click Bar Chart, Google US has the highest. Yahoo-US and MSN-US have the lowest.
From the #Average cost per booking Bar Chart, Overture-US and Google -US has the highest. Kayak and Yahoo have the lowest.
From the #Average Profit per Booking Bar Chart, Yahoo-US and MSN-US have the highest. Overture US has the lowest.
#PREDICTIVE ANALYSIS
Analyzing Search Engines - In order to find the engines that delivered the highest value, we looked at KPIs like Revenue, ROA, and Probability of Booking. We defined our business success and business failure by comparing all values with their respective variable mean and assign 1 to every value that was above the mean and 0 for those who were not.
From the graphs above, Kayak results to be the most effective search engine with the highest ROA and probability of booking. Although we have limited data, we can observe that Kayak is an upcoming platform with lots of opportunities and room for investment.
#GLM & GINI TREE
A creative solution that will agile the process in future opportunities - Once we got the results, we compared 2 different logical regression functions (glm and Gini trees) to define the best one for our purpose. Finally, we came out with a process that can lead us to test if our designed campaigns are suitable and have a high probability of being successful. This process can optimize also future campaigns and will solve possible future time inconveniences.
GLM function on the Revenue variable.
GLM function on the ROA variable.
GLM function on the Probability of Booking variable.
#BUSINESS INSIGHTS & RECCOMENDATIONS
Selected Works
DonenziLiveStartup Founder
AirFrance SEM Campaign OptimizationData Science Team Project - R Studio
#A.C. Milan Tweets AnalysisText Analytics & NLP - R Studio