W1 T1 Flashcards
Google Analytics is used
d to review/analyse click-stream online (behavioral) data on WEBSITE TRAFFIC.
resuls of google analytics are
aggregated (total) numbers based on a subset (sample) of visitors
Google Analytics (Basic Version)
Aggregate Data total number of visitor and clicks
Google Big Query (360 Suite / Corporate Version):
: individual level data, so you can see each visitors’ individual journey and visiting behavior
On Google Analytics main-dashboard you can track
You can track the general traffic to your website from different devices, locations, occasions and time-slots.
GA Easy to use / Self Explanatory: Because
GA is designed-developed for everyday use for managers at small to medium sized companies.
Using Google Analytics you can check customer journeys: 2
: attribution & paths-to-purchase for your website.
Which Analysis/Regression Type to Use ? what do we need to know=
We have to know the SCALE type of our Y variable.
nominal
eye color
ordinal:
level of angry: natural order
interval
temperature: natural order, equal interval
ratio
height: natural orde, equal interval, true zero value
To choose the correct regression type we look at 2
OUTCOME (Y) variable and its scale type.
(Y) IS METRIC/CONTINUOUS VARIABLE
MULTIPLE LINEAR REGRESSION
(Y) IS DISCRETE/NOMINAL VARIABLE
MODEL-ANALYSIS: LOGISTIC REGRESSION
When use multiple linear regression?
OV (what is the outcome?):
PV (what are the potential drivers/predictors?):
OV (what is the outcome?): Quantitative/Metric Scale
PV (what are the potential drivers/predictors?): One or more PVs, quantitative or categorical scale
Why use linear regression? 3
> So multiple predictor variables can be tested simultaneously
-> Identify the relative importance (effect) of each predictor variable (X) on outcome (Y)
-> Predict the values of the outcome variable (in our case: sales) from values of multiple predictor variable
example predictor variables
facebook, banner ads, emails
example outcome variables
total sales euro
Total Sales in Euro (€) is a quantity-metric variable here > SO we use
LINEAR REGRESSION
All ad exposure variables are also measured in numeric/metric scale: i.e total number of times # each type clicked.
This means they can directly be used in
regression
linear regression: Unstandardized Beta coefficients help us to interpret + example
absolute impact of each touchpoint on saleswe interpret them like:
Facebook: One more (increase) click on our Facebook ads is likely to increase our sales per person 17,683 € per year.
lineair regression: (standardized beta coefficients)
Comparing the effects; which X variable has a stronger impact on Y
linear regression: The Concept of INTERACTION
What is the combined effect of two different (X) variables?
logistic regression most usely used tp ana;yze
touch[oints
Logistic Regression: what shows significance
sig values
logistic regression: what shows most influencal touchpoint
Odds Ratio shows us what is the most influential touchpoint
logistic regression: Organic Search with Exp(B) Odds Ratio 365,392 means:
Customers who clicked Organic Search results are X 365 times more likely to convert (buy) in comparison to those who did not click Organic Search ResultsI know it is overly high & exaggatered (maybe not realistic) but I keep the effect HIGH for demonstration purposes – since this is an example :0)
Probabilistic Attribution:
Value of Each Touchpoint in Journey
