Erik Bower & Josip Lazarevski (Palo Alto Networks) 3/11/17

Transcription

1 Predictive Next Best Action for Marketing Demand Generation and Sales Erik Bower & Josip Lazarevski (Palo Alto Networks) 3/11/17

2 Agenda Current state of marketing Omnichannel predictive next best action marketing concept The plumbing The model and how it was made Results so far

3 First up A better name Omnichannel real time predictive next best action marketing Machine Marketing

4 Marketing nirvana! Omnichannel Right message Right person Right time 4

5 What are channels? , Palo Alto Networks, Inc. Confidential and Proprietary.

6 How to reach nirvana? Get right message? Get right timing? Target the right people? A/B testing Test send times, schedule, some have automatic throttling Segmentation and audience based Web Personalization A/B testing Depends on the prospect Segmentation and audience based Display/Programmatic Advanced optimization trained on conversion Day parting, frequrency capping Segmentation and audience based , Palo Alto Networks, Inc. Confidential and Proprietary.

7 Example Buyer s Journey , Palo Alto Networks, Inc. Confidential and Proprietary.

8 Traditional Marketing Automation Logic is static Content gap, content age Coverage is spotty Customer journeys are infinite VS. Machine Marketing Dynamic and adaptive Greater coverage of microsegments Predicts best content/offer Syncs and optimizes across channels 8

9 9 2016, Palo Alto Networks, Inc. Confidential and Proprietary. How would Machine Marketing work?

10 First you need the messages or offers 10

11 Then you need to be able to sync messages across channels Campaign Target Audience Manager Predictive Analytics 11

12 Then deploy that message on the web , Palo Alto Networks, Inc. Confidential and Proprietary.

13 .on the display channel.

14 ..and via channel , Palo Alto Networks, Inc. Confidential and Proprietary.

15 Figure out the best time to send an Account Contact DnB Product Web Visits, Form Fills Contacts we have sent at least 1 between July 2016 and December 2016 Modeling Period TRUE FALSE 6 MONTHS 1 MONTH July 2016 December 2016 Clicked on at least 1 Not clicked on any PerformanceTesting Period Contacts we have sent at least 1 on January 2017 TRUE FALSE January 2017 Clicked on at least 1 Not clicked on any 77% Accuracy 80% Accuracy , Palo Alto Networks, Inc. Confidential and Proprietary.

16 But wait there s more Identity resolution + Behaviorial data + CRM data + Other data 16

17 A bit complicated to pull off 17

18 How the model was built

19 Understand the audience and their preferences Lookalike modeling using similarity search to understand user behavior

20 B2B: Understanding account/company behavior Lookalike modeling using similarity search (similar to household in B2C)

21 Hybrid of boosted trees and ensemble trees

22 These are the people in your neighborhood All that joined together forms our neighborhoods from where we extract knowledge collaborative filtering

23 Organizing the messages with keyword mining Humans got tagging wrong 70% of the time

24 The formula + Keyword Clustering (KNN,SVM) + Term Grading = Keyword Topic Mapping

25 Language recognition We want to localize our recommendations but we don t know their language Using blend of R script and KNIME we can recognize 32 languages library("textcat"); xls<- knime.in result.detectedlanguage<-textcat(xls$dmtext) knime.out <-data.frame(xls$"recsentity.id",result.detectedlanguage);

26 Deploy to Adobe Marketing Cloud Understand the audience and their preferences Understand the content of assets Asset to person matching

27 , Palo Alto Networks, Inc. Confidential and Proprietary. Results so far

28 Homepage King of the Hill results +1300% Predicted Default

29 25. Dez 27. Dez 29. Dez 31. Dez 02. Jan 04. Jan 06. Jan 08. Jan 10. Jan 12. Jan 14. Jan 16. Jan 18. Jan 20. Jan 22. Jan 24. Jan 26. Jan 28. Jan Programmatic in Doubleclick results Dynamically Populated Headline Button Exit URL 1.200% 1.000% 0.800% 0.600% 0.400% 0.200% 0.000% Click-Thru-Rate (CTR) By Audience Ad Populates on 3 rd Party Sites 1st Party Cookies Display Maestro Display Standard Retargeting Standard Display CTR:.027% Retargeting CTR:.080% Maestro Display CTR:.351% +338% Increase in Engagement , Palo Alto Networks. Confidential and Proprietary.

30 Early signs are positive 32

31 If you are a marketing person, don t worry I m not going to take your job No really , Palo Alto Networks, Inc. Confidential and Proprietary.