|
|
| |
To answer those questions, more brand marketers are turning to predictive analysis – and these days, what’s being accomplished with advanced analysis and modeling is downright mind-boggling.
Brandchannel had the exclusive opportunity to discuss the topic with Dr. Michael Haydock, global analytics leader for IBM's Business Analytics & Optimization Practice. Dr. Haydock holds a U.S. patent in “Efficient Customer Contact Strategies” and helps large clients make smarter marketing decisions using advanced analysis techniques.
Interestingly, says Haydock, even chief marketing officers responsible for budgets of $1 billion or more typically use only spreadsheets to analyze their expenditures. That’s fine for looking at past programs and what’s happening now – but it gets far more difficult when these executives must make smarter decisions; specifically, determining how to make their marketing investment work harder going forward.
That takes analysis, and a lot of it. Haydock’s fundamental approach is to break a client’s customer file into clusters based on what he calls “feature vectors” that go beyond the typical “recency, frequency, monetary value” attributes. These feature vectors are used to describe customer behavior and predict what customers might do next. Haydock characterizes the process as a kind of “database marketing on steroids,” as each customer is scored for his or her propensity to buy, ranked from highest to lowest.
Haydock cites the example of two clients using this type of analysis, one a retailer with 80 million customers, and the other a manufacturer with 450 million customers. Both companies sell primarily to consumers rather than businesses. Each brand, says Haydock, is interested in reaching the right customers with the right information at the right time. Advanced analysis is helping these clients, he says, direct only relevant marketing messages to customers. They are rebalancing their marketing budgets for maximum efficiency and investing less money in customers who are at the lower end of the buying scale.
When it comes to social media, predictive analysis can be used to assess how customers respond via social media and model what behavior might occur. While most brand marketing organizations still view social media and traditional media separately, Haydock says predictive analysis can draw a connection between the two. For example, if a customer sees a television ad and then uses a smartphone to visit a website, it’s important to look at that customer contact cycle and understand the ROI of the total program.
|
|
| |
A particularly intriguing application of predictive analysis is something IBM calls the “Next Best Action Program.” Haydock says the program offers marketers the ability to analyze the customer touches made through any channel and then establish a relevant “conversation” with the customer, even when there might be millions of customers.
Customers’ past purchases and interactions are converted to behaviors that can be classified into “clusters” or audience segments. The program then determines where the customer is situated in a particular relationship with the company. Finally, the program guides the marketer in taking the appropriate action. In essence, the program literally determines the next best action to take with every customer on an individualized basis.
For example, suppose three customers are in the midst of a product ordering cycle. Based on analysis, the appropriate action to take for one customer might be to simply acknowledge the order by saying thank you. For another customer, it might be appropriate to suggest purchasing another item, based on the product being ordered. For the third customer, it might be advisable to re-route a delivery truck in real-time to accommodate a changed shipping address.
Haydock says the Next Best Action Program answers the need to respond rapidly to customer demands, especially in an era when customers are connected via mobile devices. That means marketers have to make customer service decisions in real time. Still, says Haydock, a program of this magnitude requires a significant investment in infrastructure and database analysis tools.
Predictive analysis is useful to brand marketers in a number of other ways. In the retail world, for example, it can be used to simulate how a marketplace might react, even eighteen months out. If the analysis shows a spike in sales activity in a particular region during a particular month, the retailer can plan to add more personnel to specific stores in that region, have enough inventory available, and plan marketing support with ample lead time.
Haydock points out that successfully implementing predictive analysis techniques takes a combination of the right hardware, software, and business analysts. He says of the 3,000 employees in IBM’s Research Division, about 125 of them are dedicated to business analysis problems of the type discussed here.
Ultimately, the value of predictive analysis to marketers is getting an intimate understanding of each customer’s needs, and delivering more relevant promotions that are better targeted. The end result: a superior ROI on the company’s marketing investment.
|
| |
|
| |
|
| |
Barry Silverstein has been a frequent brandchannel contributor since 2007. He has thirty years of advertising and marketing experience and is currently a freelance writer and marketing consultant. He founded and ran his own direct marketing agency and held executive positions with Epsilon, a leading database marketing firm and Arnold, a major ad agency. Silverstein is the author of three marketing books, including the McGraw-Hill book, The Breakaway Brand, which he co-authored with Arnold CEO Fran Kelly.
|
| |
|
|
|
|
 |
|
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
Jul 1, 2010
|
Brands at Sea: The Ups and Downs of the Cruise Industry
|
|
|
Cruise brands are struggling. In the choppy sea of brand differentiation, it is getting more difficult to tell Carnival from Royal Caribbean from Norwegian. The problem: similar routes, ships, and offerings.
|
|
|
|
|
| |
|
Jun 25, 2010
|
On Demand: Digital Video Creates New Players
|
|
|
The digital delivery business has done something else besides revolutionize the way viewers receive video: It has spawned new brand names and re-shaped the way many firms in the business operate.
|
|
|
|
|
| |
|
| |
|
Jun 11, 2010
|
One Year Later, What Does General Motors Mean to America? -- Dale Buss
|
|
|
The crash in the U.S. auto market, the historic political calculations, the wrenching financial pain for everyone from executives to bondholders to car dealers to employees – all of those epochal developments are beginning to recede, each at its own pace, into the 102 years of history of the General Motors Corporation.
|
|
|
|
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
Copyright © 2001-2013 brandchannel. All rights reserved.
|
|
commenting closed
bookmark
print
suggest topic
recommend
email
Barry - Great ideas and an interesting read. The business case for predictive analytics seems to be robust from an ROI perspective - as has also been demonstrative with SAS and SPSS's tools in the markets for ~40 years.From what I understand there's a raging debate regarding which predictive analytics platform - SAS, SPSS, or R - should be used moving forward. The latter is an open-source platform that has the power to leverage people worldwide to make predictive analytics that much more accessible, useful, and accurate. R is also headed by the person that invented SPSS.Additional thoughts?John Stonewww.leadingthebrand.com 
| Privacy Policy | About Us | Contact Us | FAQ