The numbers for mobile access to the internet are huge; even the most conservative estimates put the number of users in the billions, which means a vast number of consumers are and will continue to look for content that needs to be packaged and delivered in bite-sized chunks. Most content that is consumed on-line is not consumed in its entirety; people are generally looking for specific data points (I may only need a paragraph of information that is embedded in an 800 page document), which means that content management standards such as DITA have been well ahead of the consumption curve for some time. Other technologies such as SMS were created with a haiku delivery format to begin with (a 140 character limitation forces pithiness), which has clearly resonated with the younger end of the consumer market (hence the success of Twitter, which is basically an SMS overlay).

DITA has always been about minimalism and brevity, as has Twitter. The interesting aspect of this is that DITA applied to complex content management has worked extremely well. The question becomes, what happens when the 140 character-set world intersects with the structured content (DITA) world? Complex content has always run in parallel to SMS/Twitter, but that also assumed two separate networks, one network for content heavy access from a fixed point device, the other network focused on mobile access where information is tightly parameterized. With the increasing prevalence of smart devices that whole model is being flipped, and at this point most content being run across Twitter and other SMS based networks does not have any standard for referential integrity the way DITA-ized content does. While this may sound wonky, it’s worth considering that many media companies and consumer-facing companies are becoming increasingly focused on Twitter as a content delivery mechanism without the benefit of content management standards. Or, in other words, the mobile internet runs the risk of becoming an information morass, similar to the way the internet looked in the mid and late 90s.

The core driver of a relationship with a consumer is the product or service purchased (and I define this as the entirety of the relationship, the product includes advertising, the product or service itself, any supporting documentation, the website, customer service, basically any interaction between the consumer and the provider is part of the product). In the context of this trend, it refers specifically to publisher content; the parameters and content that defines my search for a specific piece of information or publication speaks volumes about who I am, and is one of the most valuable elements of consumer data available. The consumer’s interaction with the publisher’s content is effectively handing the key to the kingdom to the publisher. The publishers then, oddly enough, immediately toss this data to the ad networks, where it immediately becomes diluted, taken out of context, and resold over and over.

There is a hierarchy of intimacy in any consumer transaction; the most direct connection is when it is just the buyer and seller, the least direct is when multi-channel or advertising networks enter the equation. The more entities involved in the transaction, the more diluted the relationship becomes, because more people are getting their little piece of the consumer. Because publishers are the source, they are in the perfect position to establish pompetus (my favorite term, explained on my post for 1.24.09). Depending on what the publisher is offering (and we’ll skip the most obvious example), they have an unprecedented opportunity to genuinely connect with the customer. Why in the world would a publisher give something that valuable to an ad network? Money? They can make way more money by hanging onto the information and cultivating a profitable long-term relationship on their own. The ad networks are already picking up vast quantities of information as consumer troll the web, the additional data the publishers provide is a relative drop in the bucket to the ad network, but a relative gold mine to the publisher. It looks like publishers are finally starting to figure this out, and given the overall trends in publishing, not a moment too soon.

The fact that information resources can be meta-tagged and categorized via some form of vertical ontology (that is, tagged and bagged), and have this done independent of the media type, means that when a field worker looks for information, the content management platform becomes a mobile enabler. It doesn’t (and should not) matter what the media type is; video can be tagged and bagged, just like text documents, graphics, .wav files, essentially any resource that has relevant information can be organized and stored at the component level, and assembled on the fly in response to a query (e.g. “what is the proper procedure for replacing the armature platter on an MRI Scanner?”). Having a medical/technology ontology that categorizes tagged and bagged information resources means the field worker receives a full blown rich media response to this query: this can include a text description of the procedure, a video tutorial, graphics that can be exploded and rotated, a voice walk through, and so on.

In this context, a component content management system become a mobilization platform for rich media enterprise applications, and this can also be expanded to include supply chain partners that may feed information resources into an assembled product field guide. The combination of mobility requirements and the drive towards rich media is breathing new life into the content management domain across a broad range of vertical markets.

The other advantage of course is ease of integration; once information about consumers and the choices they make are kept in a cloud framework rather than a “rack mounted, behind the firewall, supported by IT” model, integration across disparate touchpoints becomes much more straightforward, and presumably provides retailers with a better understanding of consumer habits and requirements, which in turn would lead to a better customer experience.

In most companies all these elements are treated separately from a process and personnel point of view. Customer support rarely interacts with product management, product marketing, documentation groups, billing, or sales, yet all these groups revolve around the same core: the customer. Sales is overtly focused on customers, product marketing focuses on customers, but normally on an aggregate level, product management interacts with select customers to define requirements for product roadmaps (unless they’ve implemented crowd-sourcing, which most companies have not), billing interacts with customers on a monthly basis, documentation rarely interacts directly with the consumer (in spite of the importance of the deliverable), and customer support deals with customers on an exception basis, when something isn’t working the way it’s supposed to.

Every one of these interactions provides product or service vendors with an opportunity to get a step ahead of the customer and anticipate requirements (and in all fairness, most people do take their customers seriously), the problem is that all these groups are probably dealing with the exact same customer, gaining multiple perspectives on incredibly valuable information, yet a holistic and integrated view that factors in all organizational facets in the context of the customer is rarely available. The best example of this is probably the customer support center. Why? Most people are calling because they’re having a problem; the product doesn’t work, the documentation isn’t clear, something is not working, etc. This is an ideal opportunity to gain meaningful insight because the customer is dealing with the vendor on an emotional level; as a vendor you’re actually much more likely to get unambiguous feedback from someone who’s upset, and of all the organizational groups, Customer Support is most likely to take the hit.

In most instances we find the core problem is ambiguity in documentation; complex products that are poorly documented can be a nightmare for the customer, which draws a direct correlation between the documentation and customer support groups, who rarely, if ever, interact. This is another push for a fully integrated rich media deliverable, pictures speak louder than words, moving pictures speak even louder, and providing a direct feedback loop to customer support through the documentation deliverable (which is quite possible in an on-line model) closes the focal gap that keeps most companies guessing about their customer intentions.

Let’s not kid ourselves; consumers are NOT going to make a decision that reduces their privacy across the board. Why not? Because the only sound they hear is the shrill scream of the privatistas who clamor that under no circumstances should privacy be broached. The only ones who can influence consumers to allow better access to targeting data are the content and service providers, and their approach has been to put their company privacy policies in disclaimers (“hey, the information is there, they just have to read it”). Again, yeah, right.

Do the benefits outweigh the risks? Ask a well-targeted consumer (if you can find one), and you’ll be surprised by the answer. There are people out there who go on-line (often to the same destination over and over), almost immediately find what they’re looking for, and as a result continue a relationship that is satisfying to them and profitable for the vendor. This type of dynamic tends to be site or service specific, and is the result of a long, positive history of interaction that creates trust. This dynamic is what needs to lead the charge on the privacy/targeting trade-off conundrum; vendors need to take their head out of the ground and realize no one will believe them when they claim to have the consumers best interest at heart. A genuine, grass-roots movement addressing the benefits of targeting, lead by consumers, needs to be developed to counter-balance the privacy-at- all-costs movement. The beneficiaries of this effort (the vendors) need to move on this quickly; they should all know which of their customers are likely to step up. This battle needs to be taken out of the hands of lawyers, lobbyists, and geeks, and be driven by the people who are actually affected.

The vendor has a specific, encapsulated objective in targeting me; get me to buy their product or service. This is not normally much of an issue; the more accurate the targeting mechanism is, the more likely I am to be interested, the less accurate is it, the more likely I am to delete the vendor with a fast click. Presumably a good targeting algorithm will note that I delete the vendors efforts (or don’t respond), and pretty soon they stop pestering me, because why bother with the expense and effort of pursuing someone who isn’t interested? Keep in mind that any vendor who uses a sophisticated behavioral targeting application is probably trying to develop a sustained relationship, not a one night stand, and it’s in their interest to have a happy and responsive customer. The more accurately I’m targeted, the more I am likely to be interested in a continuing relationship.

Social targeting has a completely different dynamic and is driven and defined by context. The whole notion of privacy in a social network is defined by a group dynamic; nearly everyone on Facebook, MySpace, Hi5, etc. is part of a group, but even here there are levels of segmentation. Family and friends are one thing, but what happens when you start to move into a setting with commercial aspects to it? If I am part of my alumni network on Facebook, does that give my alma mater the right to contact me through the FB site for donations? I would say yes, because the line of commercial validation has been crossed in that context, either now, or in the past. If I join an interest group focused on, for example, analytics (which is a business, not social context), I would expect (and frankly would want) to be contacted by peer group members, many of who are consultants.

The sticking point seems to be breaching the privacy firewall in a strictly social (not commercial) setting. I would not want my friends marketed to in my name unless I have specifically authorized it. Since this is not likely to happen (I don’t have a commercial relationship with my friends, that’s what makes them my friends and not my clients), no one else should on my behalf either. Integrity of context does in fact make a significant difference, and hopefully most vendors will continue to respect that distinction. The problem is that the non-commercial and commercial sectors are starting to overrun each other, and you can’t assume people will behave professionally when it’s not business-centric. The bottom line should be if money has changed hands, the context and therefore the rules of engagement have been defined. If it has not, the same rule applies in reverse, context is defined, and a distinct and tighter set of privacy rules should be in place, something the social networking sites are still struggling to understand.

The direct marketing and advertising industries are operating at World War II levels of targeting efficiency; think about the unsolicited mail you receive every day (both hard copy and electronic) that isn’t even remotely relevant to your interests. Companies carpet bomb an entire neighborhood (for example, demographic segmentation by zip code) hoping to get that one hit. This is a staggering nuisance for consumers, and an expensive and inefficient process for marketers. Why would consumers be interested in something irrelevant? They’re not. Why do direct marketers carpet bomb? Mostly because they don’t have a choice. As the underlying technologies that identify consumer interests improve (cookies, embedded java scripts, advanced analytics, etc.), there has been a backlash by privacy consortiums acting in the “best interest” of the consumer to ensure than nothing relevant or personalized is delivered unless the end-user has jumped through both legal and technical hoops. The result is that when companies develop consumer-targeting capabilities that would actually create serendipitous moments, they’re restrained by concern of regulatory oversight driven by privacy fear-mongers. The mongers wildly overstate the potential for privacy abuses; this is similar to saying people could drive recklessly, therefore no one should be allowed to drive, or consumers using a credit card are at risk of identity theft, so credit cards should be restricted.

The sad part is the privacy advocates appear to be winning this battle, the IAB and other advertising groups are clearly on the defensive, and have allow the privatistas to set the agenda. It is clearly in the interest of both the consumer and the vendor to reduce the amount of spam (which is defined as irrelevant, as opposed to unsolicited advertising—something can be unsolicited and still be highly relevant). Focusing advertising on what is meaningful to the consumer and therefore profitable to the vendor creates a win/win scenario, but it requires an analysis, understanding, and prediction of the consumer’s most likely interest, which is delivered by targeting technologies. The advertising and direct marketing industries have got to stop backing up, take control of the agenda, and show consumers the benefits of accurate targeting, while addressing privacy issues in terms that are in a non-technical, non-legalese, and non-hysterical context.

There are a number of trends that will be affected by this continued shift; social networks (including business social networks) are and will continue to have a huge effect on how content is packaged and delivered, consumers who are now used to a very user friendly mobile experience will expect the same type of easy interface while access mobile applications in a corporate setting (since presumably most of them have jobs), which has a completely different dynamic from pure-play consumer interactions. Security is a huge issue (as is privacy on the consumer side), virtualization and synchronization are a baseline requirement for mobile content, and require precise collaboration between mobile content enablers and mobile infrastructure enablers, most of whom are dealing with their own particular, separate ecosystems. Continued focus on content standards such as DITA need to expand to encompass the entire range of media types (video, graphics, speech, text, etc.) while also expanding to include the inherent security concerns associated with a mobile framework.

The problem is complex, there are a lot of players with competing agendas (all self-serving), and the primary target for all this (the consuming public) seems to be blissfully unaware of the debate on any meaningful level. There have been lots of surveys taken by the anti-BT side that indicate consumers hate being targeted; “we surveyed 800 people, and 40% said they were less than comfortable being targeted”—so let’s see, that’s 320 people out of a market of 300 million plus, but that’s enough to draw a conclusion that gets taken to regulators. The annoying part is this is not a pro-consumer approach. It is, in fact, just the opposite.

What happens when BT is completely reigned in, and no one can run an advertisement without a double opt-in after a consumer has slowly and carefully read a privacy policy? For starters, almost no one gets targeted. This doesn’t mean they won’t receive advertising, in fact, what is likely to happen is the level of untargeted advertising will skyrocket (since targeting is restricted). This means massive amount of spam, because the bottom line for consumers is they are going to receive ads whether they like it or not, the question is, do you receive ads that are meaningful and timely, or are you carpet-bombed with everyone else in your zip code?

How can the pro-BT side claim the moral high ground in this debate? Claiming “transparency” won’t do it; stating your policy in dense legalese is nonsense, no one in their right mind reads disclaimers, and even if they do, most people won’t understand them. Telling people how they’re tracked (e.g. java script, cookies, etc.) is the technical flip side to legal babble, and equally pointless. Any marketer knows you don’t lead your pitch with features, you lead with benefits; the conversation isn’t about the vendor, it’s about the customer and why this technology is good for them. The whole debate assumes consumers hate being targeted (they don’t, people like a relevant experience), and that absolute privacy is paramount (it isn’t, otherwise sites like Facebook would not exist). The argument needs to be redefined, this isn’t about policies and technology, it’s about how BT benefits the consumer. Well-crafted targeting leads to serendipity and a compelling experience; this is the model the pro-BT space needs to focus on, with the consumer leading the charge.

Knowing what your customers truly want also implies knowing what they don’t want, so you don’t waste your time or theirs in unwanted solicitations. It also means understanding temporal drivers; some people like to shop all the time, others much less frequently, but everyone has an internal clock that that be set off by the right triggers, if you know which button to push and when.

So why is this level of targeting accuracy so hard for analytic vendors? The short version is insufficient data; as an example, demographic analysis (the most common type of consumer analytics) is driven by broad, arbitrary separation strokes such as age or zip code, that have nothing to do with an individuals interest or motivation in purchasing any product or service. Web analytic approaches such as clickstream analysis focus on a single dimension of consumer interaction, even thought the relationship between a supplier and their customers is always multi-dimensional. This isn’t just a matter of depth, it has to be depth and breadth, it has to be real-time, and it has to anticipate why?

Consumers leave a vast data wake as they move through a network or even a website, the problem is not lack of information, it’s lack of correlation. Knowing how to take the myriad of details available on consumers and combine them into something truly actionable has been the holy grail of consumer analytics, and one that is about to undergo a fundamental shift as we develop new analytic models that are designed to leverage the vast increase in available data while addressing core, fundamental motivations.

Here’s the issue. Why are these companies interested in tracking or profiling consumers? So they can do a better job of serving the consumers need for information, products, or services. Do you hate getting good service? Does it bother you to walk into a retailer, be recognized, and have people falling over themselves to give you what you want? This type of scenario apparently bothers the privacy advocates, and they have decided they know what’s best for you.

The most compelling aspect of being on-line is having access to unlimited information, goods, and services. However, the infrastructure that delivers this unlimited choice needs to be paid for, most often by advertising revenues. The advertisement is only cost-effective if site visitors click on the banner, or link, or whatever has caught the consumer’s attention. Knowing what type of advertisement a consumer is interested in is likely to increase the response to that ad; the consumer gets what they want, the vendor sells more product, the economy grows, and everyone is happy (except the privacy advocates).

So lets’ say the privacy advocates are correct, and the legislature rolls over and give them what they want. What are the likely effects? First off, spam on an unprecedented level. Why? Because vendors will have no way to accurately target consumers, and will be forced to a one size fits all model. There will also be a noticeable decline in the delivery of on-line services, because the model that pays for them will be hamstrung. On-line companies will be forced to cut back, the technology vendors that support them will be forced to cut back as well, and sectors of the economy that are normally on the leading edge of customer service will be hit very hard. You can also kiss personalized service good-bye, vendors will be forced to treat all customers as unidentifiable drones, so again, if you hate getting good service, this is your lucky day.

The behavioral targeting space has done a miserable job of convincing consumers that it is in their interest to be targeted, as a result, the privacy advocates have claimed the moral high ground, and are yelling at the top of their lungs. To makes matters worse, the only ones who are really listening are the legislators (most of whom have a thin grasp of business, and an even thinner grasp of technology). I believe there is an elegant way through this, where nearly everyone comes out ahead, and this is something I will be addressing in upcoming posts.

Who are they? This is demographic information, which is normally gathered at the point of website registration or product/service purchase. This is the best (and possibly only) opportunity to gather as much personal information as possible, and if the request is phrased properly, most people will not have a problem disclosing details such as contact preferences, etc.

What have they done? This includes both absolute and relative information. Absolute information is gained by transactional analysis of the customer’s buying behavior; the further back the transactional analysis can go, the better position the vendor will be in to derive long term changes in customer behavior and anticipate future changes. This also includes clickstream analysis of the customers website visits (browse vs. search, etc.) The relative information is derived from collaborative filtering (people like you also bought things like this).

When did they do it? This includes a temporal overlay tied to transactional analysis (including clickstream behavior). Temporal patterns can often be anticipated, since most people are creatures of routine and habit. Integrating temporal analysis to purchase behavior can indicate the best time to send a message to a customer, based on their past behavior.

How did they do it? There are two aspects to this; where did they come from, and how did they get to you? Search engines have long been the gatekeeper that brings consumers to e-commerce and entertainment websites, however social networks are quickly becoming a driving force, and the rise of micro-blogging applications like Twitter offer a potential gold mine of time-sensitive opportunities, once people figure out how to work the system. The how did they get to you part has extensive implications (which will be addressed in detail in a separate white paper), but the short version is that there are two primary means for consumers to come in contact with your business; either on-line through a fixed access point, or on-line through a dynamic (mobile) access point. The mobile access point also has implications for website design; looking at a website through a 21 inch monitor is very different from looking at it through a 2 inch smart phone screen, and it’s safe to assume that over the long run, the default access device is likely to be portable. If your website is not optimized for mobile access, this is something that needs to be addressed sooner rather than later.

Where were they when they did it? This aspect correlates closely with mobile access information, and is the province of geo-spatial analytics. What you’re looking for are behavioral patterns that are tied to specific locations; again, most people tend to do the same thing at the same place every day. Knowing where they are when they access your site can provide a treasure trove of actionable information.

Why did the do what they just did? Of all the consumer modeling technologies, this is probably the most controversial, and by far the most effective. The why of consumer behavior is derived from psychometric profiling; people are motivated by different things at different stages of their lives (or even by time of day, time of year, etc.). Psychometric profiling can be used to address specific behavioral attributes (is your customer impulsive, an early adopter, price sensitive, technophobic, etc.). All of these behaviors can be captured and quantified through profiling questionnaires; the questions need to be applied to a statistically significant sample of your customers (2%-3%), and can then be correlated and extrapolated to the remainder of your customers. These types of extrapolations tend to run at approximately 85% + accuracy during the initial iteration, then subsequently improve as the system gathers more learned behavior.

The who, what, when, where, how, and why dimensions of customers can be combined using advanced multi-variate analysis techniques to create an N-dimensional model of the consumer (referred to as a hypercube), that provides a a unique profile of what your customer is truly like, and therefore can be used to define the optimal marketing approach. While this may sound like an elegant solution, it is an elegant solution for a single customer, which makes it interesting, but ultimately useless. In order to make this a truly useful marketing tool, this profiling methodology needs to be tied to an intuitive visualization tool that allows manipulation of data at an aggregated level, but provides capabilities to execute in a highly segmented manner.

An example of this would be a marketing specialist who wants to know which customers are most likely to be interested in a new product, and what is the most effective way to reach them. Assuming that behavioral data on customers (who, what, when, etc.) can be integrated with existing product schemas tied to profitability models, it becomes relatively straightforward to determine (in rank order), which customers are most likely to be interested in product X vs. product Y. If contact preferences have been captured as part of their demographic profile, then you also have the optimal approach vector (given to you from the source of truth, your customer). Combining product profitability attributes with channel preferences delivered in the context of an N-dimensional hypercube model allows the marketer to finely slice and dice their customer base, while still dealing with a critical mass that is large enough to justify the associated marketing expense. This not only provides an integrated, holistic view of the customer, it provides linkage between back-end operational systems and outbound marketing systems that can be used across divisional lines, allowing companies to fully leverage their systems to deliver a quality experience to their customers.

If you look at the claims made by analytics vendors, it would appear that their corporate customers have a complete and fully integrated view of their consumers; everything is working perfectly, the results are spectacular, everybody is happy, etc.

The reality is that most companies have a very thin, rear-view mirror perspective of their customers; there is a one size fits all approach to marketing, segmentation is at best a nascent science (the best example of which is the direct marketing industry trying to claw it’s way back up to a 2% response rate—otherwise know as a 98% failure rate), there is little or no integration across operating units that touch the same customer (for example, sales, customer support, billing, etc.), and there is absolutely no integration of the vast amounts of transactional data into a modeling framework that provides a genuine, holistic, and actionable view of the customer, and puts that information in the hands of the people who are in the best position to drive both top and bottom-line revenue.

This situation is further exacerbated by a genuine lack of understanding of consumer motivation. Companies know who their customers are, what they’ve done, when they’ve done it, how they did it, but the key piece they lack, the one element that holds it all together, is why they did it? What drove a consumer to site A vs. site B? What drove the consumer to choose product C vs. product D? The why of consumer interaction is a huge gap in understanding the dynamic of customer relationships, and is one that does not appear to be moving towards any sort of resolution.

The net result of all this is consumers who are frustrated, clearly misunderstood, and able to switch vendors with the click of a mouse, coupled with vendors who are struggling wildly to hang onto their customers, about whom they have almost no understanding, even on a historical level, much less a real time view of what is going on and why. And just to keep things even more interesting, the underlying technology infrastructure and it’s usage are moving at a far faster pace than even the most nimble companies are able to manage; two year ago no one had heard of Twitter, now they dominate the consumer technology landscape, and in spite of this if you ask your average B2C business what their Twitter strategy is you’ll get a blank look. And Twitter is just one small example, there is the whole rich media landscape that needs to be navigated, the stunning rise and dominance of social networks, and all of this occurring in the context of a mobile operating framework which is very different from the old school internet.

There have been a lot of business models developed and deployed over the years to try and commercialize the vast amounts of on-line content that is continuously generated by professional authors, the news media, and consumers. Most commercialization models are either ad hoc pricing (popular with the analyst communities), subscription pricing (popular with the news media), advertising based content delivery (also media as well as User Generated Content), as well as, of course, free content (like this blog).

One of the core challenges content creators face is driven by the need for utility, uniqueness, and monetization. If I’m a market research professional with a large organization, and I need a report on a very specific topic in a hurry, I can trot over to an analyst site and cough up $3000 for a copy of exactly what I need. The information has high utility, it is unique, and I have the means to purchase it. However, if I’m doing research for myself, then that $3000 is coming out of my own pocket, and given the opportunity cost of $3000 (for example, a 65 inch LCD monitor), I am much more inclined to take my time and try to find a free version of the same material. If I’m lucky enough to find the free version, it doesn’t matter if it’s supported by advertising; clicking on the ads are optional, the important thing is that I’ve found what I want at a much lower price point.

This utility/uniqueness model works when the information is a reflection of extensive research and analysis done by experienced professionals. The problems facing the news media is that most of what they cover is current events, with limited analysis. If you want a quick and dirty overview of swine flu, you can find unlimited sources of information without having to pay a penny; therefore anyone trying to charge for it is one click away from being out of luck. They have utility, but they lack uniqueness.

User generated content faces an even higher hurdle; the barriers to entry are essentially flat, the rate of content generation is staggering, and as I’ve mentioned in prior blogs, very little of the information begin created is organized or tagged, and is therefore going to be very difficult to find or syndicate. This ecosystem is then further muddied by folks who are trying to create device specific readers such as Kindle, or the recent announcement by News Corp that they are looking to create a delivery device specifically for their own content. The default access device for any network based content is going to be a smart phone; you can build devices like the Kindle, but if you can get the same content on an iPhone, why bother with a new device that only serves a single purpose?

The only area that is a solid bet for direct monetization is high value information written by experienced professionals (utility and uniqueness). User generated content and news can generate revenue through advertising, but it s a secondary effect (that is, people are not paying for the content directly). Content creators in this group will still make money, but it requires a much broader reach since the click through rates are a small percentage of people who view the content. The more interesting challenge is how to apply monetization schemes to the 7th mass media channel (see the post from 3.3.09 for more detail). I will address that in my next post.