Marketing Attribution -- assigns value to each channel

In our course we learnt about many online and offline channels to market. Number of channels keeps growing and marketing is becoming more and more complex. When a conversion happens which marketing channel gets the credit? For example, a consumer who is in the market to buy a camera, s/he viewed few TV commercials of Nikon, viewed few times display ads while reading Yahoo email, searched Google to research for Nikon models and clicked on few links and advertisements and discussed with friends on Facebook and saw advertisements on Facebook and then bought it on Amazon. For this sale, how to attribute which channel(s) influenced consumer sale? There is no accurate way to come up with an answer. But when we look at these data points across more customers, we can derive patterns of influence and attribute value to the channels. In the industry different attribution models and techniques are evolving to solve this complex problem. Using market attribution techniques marketers can optimize usage of their budget for greater ROI. I will give an example just to give a flavor of how a simple marketing attribution model works. In this simple example, say we have only 5 customers who converted and got exposed to four different channels that the product got advertized. In this table value 1 indicates customer has seen the advertisement in that channel and value 0 indicates customer hasn’t exposed in the channel. Here by looking at this table, which channel you give most credit? Most likely we say Google as most converted customers exposed to Google.

	TV	Google	Facebook	Display Ad
Customer 1	1	0	0	1
Customer 2	0	1	0	1
Customer 3	0	1	1	0
Customer 4	0	1	1	0
Customer 5	0	1	1	0
Total	1	4	3	2

Table-1: Converted Customers only data points

In Table-1 we just looked at converted customers data points. In table-2 we combine all people who are exposed to the advertisements. Just to simplify say we have only 10 people who are exposed to the advertisements and 5 of them converted customers that we already looked in Table-1 and 5 non-converted. When you see both the records, now which channel advertisement you think has more influence in buying our product?

	TV	Google	Facebook	Display Ad
Customer 1	1	0	0	1
Customer 2	0	1	0	1
Customer 3	0	1	1	0
Customer 4	0	1	1	0
Customer 5	0	1	1	0
Not Converted	1	0	0	0
Not Converted	0	1	0	0
Not Converted	0	1	0	0
Not Converted	0	1	1	1
Not Converted	1	1	0	0
Total exposed to Customers	1	4	3	2
Total exposed to non-customers	2	4	1	1

Table-2: Converted and non-converted customers data points

Answer is Facebook, because even though only 4 customers got exposed to Facebook ad, out of 4 three of them got converted. So facebook advertisement seems to have more influencing in converting the customers than other channels based on these data points. So we attribute more value to Facebook channel than other channels.  

In real world the marketing attribution problem is much more complex and data intensive and there are products specialized in solving complex attribution problems. These products use complex modeling techniques to calculate attribution to different channels and helping marketers optimize their usage of budget and gain higher ROI. 

For more details you could refer to

http://www.inc.com/casey-szulc/why-you-should-care-about-marketing-attribution.html

http://www.visualiq.com/wave-report-pdf-download