Text Mining Mainstream?

Are words data points? I am asking this question because I have seen a broadening applicability of text mining approaches and methodologies for our marketing field. Trying to understand how we data-driven marketers can combine data and text mining into a more holistic understanding of our consumer, I have not seen too many interesting solutions or even attempts to combine both. There are still two much siloed “Mining” worlds, one focused on behavioral or attitudinal data, the other one on texts.

There are some rare exceptions like MotiveQuest which mines online blogs, and chat rooms to understand how consumers discuss brands in an “unsupervised” and “natural” environment online. Or SPSS which just released the first text mining application as part of their suite. But if we all agree that words are data points, then basic text mining capabilities should be part of any state of the art analytical marketing approach and team, not just a peripheral story outside of the mainstream.

How can we have real and meaningful applications of text mining methodologies as part of any innovative marketing science discipline? I see mainly three areas of applicability:

• Understand how all the different communication messages of a particular brand are disseminated throughout every channel. Most marketers of a global multi-channel brand don’t really have a good understanding of how they are messaging with consumers. A traditional brand audit does not fully capture all brand voices
• Understand how a competitor’s brand talks in every channel. Comparing one’s own brand voice with key competitor’s brand voice will reveal key differences of word clustering and might function as an early warning system if a competitor changes its brand position and language
• Understand any spoken consumer word about a company’s brand or phenomena in any channel. This could be either on the Web, in Call-Centers, in email communications, etc. It’s not just important what a brand says but also how consumers interact with a brand

The accelerating applicability of text mining will push us marketers to expand our horizon when we talk about understanding our consumers. Soon this will not be just a nice addition or gimmicky approach but an essential part of any consumer research.

Clausewitz and Uncertainty

A couple of days ago a good friend of mine and brilliant mind Michael Lang, Professor for History at the University of Maine, helped me to rediscover Carl von Clausewitz. Clausewitz, a Prussian 19th century general, is famous for his strategic writings about war. Michael believes strongly that Clausewitz’ argument that “no amount of analytical insights can ever fully count for the uncertainty and unpredictability of war” continues to have strong implications for business strategies today. After rereading parts of Clausewitz writings, I could not agree more.

Clausewitz writes about the nature of war but most of his thinking and conclusions are very applicable and relevant to today’s marketing challenges. His main themes are threefold:

• The uncertainty and unpredictability of warlike situations can never be mastered,
• We have to accept and even embrace the nonlinearity of warlike event (brilliantly described in Alan D. Beyerchen’s article “Clausewitz, Nonlinearity and the Unpredictability of War”),
• Successful strategists need to search for concepts of how to deal with this unpredictability in nonlinear event situations.

Don’t these three thoughts sound exactly as a good description of today’s marketer challenges? The marketer’s job of creating relevant interaction between brands and consumers face all three strategic concepts: Unpredictability, Nonlinearity, and the need to deal systematically with both challenges without the hubris of ever mastering it. Today’s marketing issues are all about the loss of control and the uncertainty created by the non-linear interaction between brand and consumer.

What can we learn from Clausewitz’ struggle with these phenomena. I suggest three elements:

• While we marketers have to strive for more certainty, we also have to embrace uncertainty: It’s not about eliminating every single aspect of uncertainty but to approach a higher accuracy level of prediction without obtaining unjustified arrogance about upcoming events that are all just beyond secure prediction
• Understand the nonlinear make-up of today’s communication world: Analyze diligently every single aspect of the non-linear characteristics of the brand-consumer interaction space
• Design tools, and mental frameworks to deal with uncertainty: It’s not about increasing the quality of prediction by 1% but having tools and a mental framework in place to deal with uncertainty.

Clausewitz might have written over 150 years ago about the intricacies of 19th century wars but he was more a strategic mind than a warrior. Sometimes it’s more insightful to understand basic strategic issues with writings from great minds, independent of the age of their thoughts.

Declining Marketing Impact?

Over the last months a multitude of publications (e.g AdAge, Business Week) talked about the declining effectiveness of not just traditional broadcast marketing vehicles but marketing in general. But no one tries to dissect in any more details if marketing investments overall, independent of channel, create today less impact than before. Most articles focus on the efficiency decline of TV spots but expand the impact decline to other marketing vehicles, too

Looking at the growth of overall marketing spend in comparison to GDP, it does not seem that there is a big disconnect, since both numbers are rising steadily in roughly the same proportions. Therefore the marketing dollars invested per 100$ revenue generated across Fortune 2000 companies seem to remain consistent. Does this mean that we just need to reallocate marketing investments against different marketing channels, mainly away from traditional broadcast media to interactive and non-traditional vehicles?

Talking with leading marketers at Fortune 2000 firms I get the impression that the cost of acquiring and retaining customers continues to grow. If this is true, then we would need to see a significant increase in marketing dollars spend for generated revenue. But as mentioned above, we don’t see this pattern. Surprisingly for the size and the acclaimed sophistication of our marketing industry it is very difficult to get reliable data points to verify either my impression of consistent relation between marketing dollar and revenue impacted and generated or to support most marketer’s perception of increasing cost of acquisition and retention per customer.

Does anyone know about recent research about this challenging question beyond very channel specific analysis like the recent McKinsey study about the declining effectiveness but rising cost of TV spots? This is not a minor issue but goes to the fundamental issue of shifting investment strategies within a Fortune 2000 company (either away from marketing to services or product development or vice versa) and within the marketing budget itself (from one channel to another). There is almost too much focus on tactical channel optimization instead of asking the bigger question of marketing effectiveness overall. Thoughts?

The "Long tail" of brand design

Last Sunday Rob Walker wrote in the NY Times magazine about the emerging underground brands like aNYthing”, “Barking Irons”, or “The Hundreds” and how these brands are changing the whole brand development process. These new brands are built not around products or even ideas but through the urge to combine a particular self expression that materializes itself with utter commoditization. It’s not even about a particular self image or any kind of subversive desire but about distributing the expression across as many sellable media as possible, from t-shirts to videos, from coffee mugs to paintings.

Walker does not expand on the paradigm shift in brand development that he seems to observe. Traditionally brands have been developed from products or services that had a benefit advantage that consumers could perceive. It all started in the late 19th century when packaged good companies needed to identify their products across a wide geographic area without any local connection. Then over the last 20 or 30 years, brands expressed stronger and stronger a particular life style that is expressed by the offered product or service. It was a less an objective (almost material) benefit but an emotional aura that the brand was able to communicate. Now brands like aNYthing or the Barking Irons are fully dissociated from any kind of product or even idea. The lifestyle (how nebulous it might be) is the centerpiece of the brand definition.

This emerging paradigm shift of how brands are developed can move into two directions:

• It will be a short lived phenomena since no brand can last too long without any benefit oriented structure, either physical or emotional. The barriers of entries are just too low; today’s consumers are just too fickle to develop any long term loyalty to these kinds of brands.
• This paradigm shift will just get stronger. And here enters Chris Anderson with his “Long Tail” philosophy, where the number of hit brands will decrease while the numbers of smaller brands and potentially short lived brands like “The Hundreds” will take over a bigger portion of consumer’s mind set.

What can a mature brand do to be prepared for both possible outcomes? On the one hand it must further strengthen what the brand really stands for in a long lasting differentiated manner, on the other hand it can empower the consumer to build their “long-tailed” version of expressing the particular brand in multiple ways. I am not able to predict where this paradigm of brand development is going but we marketer have to be prepared either way.