Shared publicly  - 
There was a small furor recently about how Target corporation tracks individuals and does efficient marketing. There are many opinions on what Target does, I respect all of them and I'll leave you alone with them.

But I was watching a online presentation by Andrew Pole from Target at a predictive analytics conference and took this screenshot from that talk. It shows the variables in play that Target collects (across online and offline touch points, mobile and desktop) and ties to Guest ID (you :) in order to do better marketing.

If you were thinking of doing this for your company, you and thank Andrew and Target for this inspirational schematic (and for sharing with you what it takes to go from wanting something to actually getting it done).

Hugo King-Fretts's profile photoRob Pettit's profile photoTibli Dickson's profile photoLiyana Sulaiman's profile photo
Great overview thanks for sharing!
Smart retailer ~ targeted and Relevant Offers ~ Satisfied and loyal customers ~ Brand Supremacy. Good for Target. Good for the customers. 
that's just crazy how much they can understand about consumers, but i would imagine that would not work in the EU due to consumer privacy laws? I would assume Walmart and other big players would have a similar model but are maybe less willing to talk about it as they don't want to show their competitors the extent that they understand consumers
Hey +Avinash Kaushik what was the name/site of the predictive analytics conference you attended?

- if you dont mind sharing.
+Jeevan Katwaru I'd watched Andrew Pole's session online. The videos from the conference were put on the web (and someone had shared the link after the New York Times furor). I'm sorry but I can't find the link. But Google it, I'm sure you'll find it.
Whats so new about it? Companies have been doing it for a long time and a good example is of Amazon that started personalized and predictive results.
Finally main street should awake to this very fact, no marketing initiative can be truly effective without such an ID.
interesting post. My client, a leading pharma retailer has a similar setup, some of my findings/ opinions about using such a setup
- There is a fuzziness about an event(say a sale, coupon redemption,etc.) being associated with a customer. Hence there exists probabilistic scores which links events to customers( 1 event to many customers). Most of the times a sale is associated with more than one customer, which makes it less accurate.
Most of the times, phone, credit/debit card#, dob,etc are used for such integration.

Say a $10 transaction is associated with customerID 1001 with probability of 75%
Same $10 transaction is associated with customerID 1565 with probability of 90%

- Being fuzzy brings complication in considering it for modeling, especially when we are modeling for a very small set of data.

I am not sure if there exists robust way to track all customer interaction points without a loyalty program which incentives customers to use card for all their interactions with the brand.

~ Deepak
Hey +Avinash Kaushik .. Most of the companies are having a similar strategy in building the GuestID. There's even more to this diagram which Andrew Pole has not shared. Anyways, thanks for sharing.
Jeff, without using PII, so what would be common thread which would help link various customer touch points? Could you pls elaborate.
Is it some kind of loyalty/RFID card trackign you are talking about?
Add a comment...