Marketing Channel Attribution

Markov Chains

A Markov model predicts the probability that a user will transition from point A to point B based on steps that a user takes through a website. The sequence of these contents are determined by the Markov order, which range from [0,4]. A Markov model representation of the customer buying or conversion process might look something like the following

The outcome of a Markov model is binary; a customer either converts or does not (NULL).

Build the simple heuristic models (First Click/first_touch, Last Click/last_touch and Linear Attribution/linear_touch):

##    channel_name first_touch_conversions first_touch_value
## 1           eta                    3164        11909.4762
## 2          iota                    4606        19597.2613
## 3         alpha                    6308        19121.2724
## 4          beta                    2831        12235.5917
## 5         theta                    1606         6652.3493
## 6        lambda                     902         3735.6022
## 7         kappa                      74          305.7432
## 8          zeta                      27          103.0040
## 9       epsilon                      99          412.3012
## 10        gamma                     165          718.9780
## 11        delta                       1            6.1190
## 12           mi                       2            5.2730
##    last_touch_conversions last_touch_value linear_touch_conversions
## 1                    4167       16754.2038              3539.951157
## 2                    3355       13487.9743              3857.096221
## 3                    8447       28414.2143              7574.718594
## 4                     989        3850.0210              2083.500145
## 5                     653        2799.0920              1022.801394
## 6                    1207        5249.9500              1035.257572
## 7                     230        1069.3843               137.964078
## 8                     107         453.2608               136.551540
## 9                     531        2202.6123               272.170438
## 10                     92         506.0140               121.041639
## 11                      5          10.9720                 1.725000
## 12                      2           5.2730                 2.222222
##    linear_touch_value
## 1        13783.497051
## 2        15988.988995
## 3        24524.709569
## 4         8954.266717
## 5         4295.743619
## 6         4430.316169
## 7          599.747786
## 8          539.528763
## 9         1106.270065
## 10         569.417358
## 11           4.404050
## 12           6.081444

##    channel_name total_conversion total_conversion_value
## 1           eta                0                      0
## 2          iota                0                      0
## 3         alpha                0                      0
## 4          beta                0                      0
## 5         theta                0                      0
## 6        lambda                0                      0
## 7         kappa                0                      0
## 8          zeta                0                      0
## 9       epsilon                0                      0
## 10        gamma                0                      0
## 11        delta                0                      0
## 12           mi                0                      0

4 types of data modelling are put together to compare and contrasts

Marketing Channel Attribution

YAP S S

12/4/2018

Markov Chains

Outcomes

The Total Conversions are plotted against the channel name

Combining Data and Final Visualizations