Chief Marketing Technologist

Data as a new marketing channel

Marketers love data. Web analytics, behavioral targeting, social media monitoring, audience measurement, and ever-expanding CRMs are all testament to the demand for more data about prospects and customers. But that kind of data flows one-way: from the outside world into marketing.

However, a new way of marketing with data is emerging that heralds a more significant shift for marketing — and for the web itself. It's about reversing the flow, having marketing publish data to the outside world, as a way to increase visibility, grow relationships, and build a brand. Call it data web marketing.

Data web marketing is the product of several diverse forces: search engine optimization (SEO), web service APIs and mash-ups, the semantic web and linked data, and a new generation of data sharing and visualization sites. The common thread among all of them is data — particularly structured data — in formats that are easy for software programs to consume. Let's take a closer look at how these different data initiatives fit together in this new kind of marketing.

Metadata for Finding and Sharing

Much of digital marketing has been centered around the production of compelling content and the promotion of such content through advertising and social media. This content — text, images, video, audio — is designed for people to consume directly: reading, watching, listening. But as search engines and other agent-like services know well, processing that kind of freeform content with software into other useful applications can be challenging.

To overcome that challenge, web publishers — a label which includes most digital marketers — are increasingly encouraged to tag their content with metadata, describing it with machine-friendly RDFa or microformat vocabularies. For instance, Facebook's Open Graph Protocol leverages such metadata to facilitate sharing among friends and like-interested communities. Google Rich Snippets and Yahoo SearchMonkey leverage such structured data to improve search results for locations, people, reviews, events, and more.

Marketers who tag their content with such metadata are rewarded with better visibility in search results and social media streams.

This is the first step towards data web marketing. It's an incentive for marketers to start to think about deploying structured data on the web in a consistent and logical fashion. These skills are currently lumped into SEO and social media optimization (SMO). But as the metadata vocabularies used get more sophisticated — like the GoodRelations vocabulary adopted by several major online retailers such as Best Buy — the art of organizing and maintaining this metadata becomes a more significant mission unto itself.

Data APIs for Apps and Mash-ups

The next incarnation of data web marketing is evident in the proliferation of web APIs. For years, such APIs have been provided by major web players such as Google, Amazon, Facebook, and Twitter as a way to strengthen their third-party ecosystems. Enabling mash-up developers to remix key pieces of their content has helped those companies build their brands, drive affiliate revenue, and spur innovation around their core services. In the case of Twitter, these APIs have arguably been paramount to their success, as most people tweet via one or more third-party apps.

But now, organizations as diverse as The New York Times, Netflix, Compete, and Wine.com are offering APIs too. The API directory ProgrammableWeb lists over 2,000 APIs from Adgooroo to Zillow. There's even a thriving market of service providers — including Mashery, 3scale, and Sonoa Systems — to help companies launch and maintain these APIs. What's driving this?

Although APIs are a technical endeavor, the reasons to offer them — and the decisions about what to provide and under what terms — are rooted firmly in marketing. Generally, the motivation is one or more of these objectives:

• generate traffic back to the organization's web site
• build brand authority by being the "reference source" in a particular area
• drive affiliate revenue with embedded links for e-commerce transactions
• strengthen relationships with power users and their respective communities
• accelerate feature innovation beyond internal development resources

The explosion of apps for the iPhone, iPad, and Android devices has further increased interest in APIs. Developers crave new sources of data they can mash up in cool, innovative ways, and publishers (i.e., marketers) are eager to extend their reach in this burgeoning app-sphere.

Here is where the shift towards outward-flowing data clearly manifests itself. Via APIs, businesses are publishing data — sometimes metadata about content, but in many cases the data itself as the content — as a marketing tactic. The quality, reliability, and usefulness of that data impacts the brand of the company that is publishing it. Therefore marketers should take a proactive role in the design and management of such services.

It's worth noting that the distribution of data this way is a kind of "channel" to prospects and customers. Often the people who will actually buy a company's products and services are not the people who will directly consume these APIs. Rather, the target audience is the people who will use the widgets and apps built with these APIs. The developers leveraging these APIs are, in a manner of speaking, distributors for the electronic essence of the publisher's brand. This raises many of the classic challenges of channel marketing in a new context — balancing the needs of two very different yet interdependent constituencies.

Data as Social Media

But the migration of data from a back-office workhorse to a front-office ambassador is perhaps most apparent in the rise of data sharing and visualization sites such as Factual, Infochimps, Freebase (which was recently acquired by Google), Many Eyes, and Swivel. These sites turn data into social media. What YouTube does for videos and Slideshare does for presentations, these sites do for data sets — let people share, search, and associate them in the social sphere. But data, unlike videos or slide decks, also lends itself to community contributions in assembling and editing data sets, enabling crowdsourced data management.

The people consuming these data sets include some of the same developers using web APIs. However, these data sets are also being downloaded by non-programmers, regular desktop users, who use them to feed spreadsheets and other business analysis applications. This broader audience demonstrates the real value data on the web can provide. Many of us crave data in our work — it helps us make decisions and persuade others. But to date, it's been very hard to find such data in its raw form. Being able to quickly search social repositories for relevant data is a significant step forward.

Of course, the moment you have an engaged audience consuming something — in this case, data — marketers will have an interest in piggybacking on that channel to reach them. This can be a good thing, if marketers are incentivized to bring something of value to the channel. What value could marketers bring to a data web? I'd propose: good data that their audience would find useful.

As with content marketing, the strategy of such data web marketing has to be publishing things that are genuinely valuable to the intended audience — otherwise, no one's going to bother to consume it. Such data might include industry research that the company has sponsored, aggregate performance data that customers can use for benchmarking, knowledge bases to support specific applications, and so on. Like web APIs, the benefits to marketers for producing and distributing such data are brand building, relationship building, and generating pull-through for their products and services.

Converging in the Semantic Web

So where does the semantic web fit in data web marketing? It could be argued that the semantic web — specifications such as RDF and OWL, SPARQL, and the vision of linked data — has been an elegant solution waiting for a mission. Outside the walls of enterprise IT and certain scientific communities, there just hasn't been enough concrete demand for directly sharing data on the web. Yet.

But as publishing data becomes more popular in the contexts discussed above — SEO, web APIs, and data as social media — the advantages of adopting semantic web standards become more apparent as well. The problem with web APIs is that they're different for each publisher, requiring developers to explicitly build and maintain their software for each one. Similarly, the data sets shared on sites like Factual and Infochimps are currently independent of each other — each its own little silo — that makes it difficult to discover and leverage relationships between them.

The semantic web was designed to address those problems. As momentum in data publishing grows, the attractiveness of its standards will grow as well, making it easier to publish and consume new data sets. By providing a common language for data, the semantic web also makes it easier for different data sets to be related to each other and to enable those relationships to be programmatically discovered by software. This is why the semantic web movement is also called "linked data."

Perhaps most exciting, however, is that the semantic web enables all of the above data publishing scenarios to converge. SEO and SMO are already leaning towards the semantic web standard of RDFa over ad hoc microformats for encoding metadata on web pages. In addition to its web API, The New York Times now offers its data in RDF. At least one social data site, Freebase, offers an RDF interface as well. Leveraging the same formats across these different scenarios makes it easier to reuse data and software across all of them — instead of many separate pockets of online data, we can begin to view a more cohesive web of data.

In conjunction with independent community initiatives such as DBpedia and open government data programs in the U.S. and the U.K., a critical mass of such data is emerging. Like Metcalfe's Law for networks, the more data there is in a common format — particularly one that facilitates interrelationships — the more valuable this web of data becomes. That encourages more developers to create more applications, which in turn inspires more publishers to participate by sharing more data, and a virtuous cycle can ensue.

The New Wave of Data Web Marketing

Many aspects of digital marketing today are tactical in nature, with relatively technical details driving their implementation. For instance, most CMOs and VPs of marketing don't think about the techniques involved in good SEO. But they do care about ranking favorably on Google and Bing for topics that are core to their brands.

The fledgling scenarios of data web marketing described above, when examined individually, are largely tactical too. But when viewed collectively, they reveal a tectonic shift towards data as a new wave of web content — one that is powering a new generation of smart applications on the web and mobile devices. At that level, there are strategic opportunities that deserve executive consideration.

Just as companies who were quick to embrace the web, search marketing, and social media used that as a competitive advantage, a similar dynamic is likely to exist with the web of data. Whose data will be the authoritative source in each market out there? Whose data will be embedded in the most exciting third-party applications? Whose industry-specific vocabularies will become de facto standards? Whose open data will be most linked to within other data sets, generating a positive feedback loop around a data-level brand? Those are strategic questions that marketers should begin thinking about.

The shift to data web marketing will take a little getting used too — we've had 30 years of clinging to data as something to be squirreled away. Gaining organizational experience in the web-facing data tactics above is a good start, but more forward-thinking leaders in marketing should encourage their teams to connect the dots between these initiatives. They should contemplate what Tim Berners-Lee said in an interview with Talis, advising company leaders on what the semantic web can really mean for them:

You should take an inventory of what you have got in the way of data and you should think about how valuable each piece of data in the company would be if it were available to other people across the company, or if it were available publicly, and if it were available to your partners.

That is the essence of data web marketing.

Scaling the new marketing organization

What's the best way to scale marketing?

When marketing grows beyond one person, how do you divide up the work? This is one of the key strategic decisions the CMO makes, as the architecture of the marketing organization shapes its strengths and weaknesses. As my operations professor at MIT often said, "structure dictates behavior."

The most common way of organizing marketing has been a hybrid of three kinds of divide-and-conquer approaches:

• by product (and overall corporate marketing)
• by tactic, such as advertising, PR, search marketing, the web site, etc.
• by region, usually large ones such as North America, Asia Pacific, etc.

Front-line marketers specialize in a particular product, tactic, and/or region — developing experience and expertise in that area. As the company grows, the size of each of these silos grows, independently of each other.

The challenges, however, are communicating and coordinating among these different groups. Any time you cross an organizational boundary, issues arise with priorities, resources, attribution, costs, metrics, etc. Sometimes these issues are small, sometimes not so small. But they're always a drag on the organization's speed and agility — and frequently a source of continuity problems for customers.

After all, customers experience their interactions with you as a continuous flow through their eyes. They don't care about your organizational structure. They don't see the dividing lines between, say, the search marketing team and the email marketing team. All they know is when you make sense — when each interaction with you builds upon the previous one for how you will help them specifically — or not.

Unfortunately, far too often, such continuity is woefully lacking.

This problem is exacerbated by the number of new marketing tactics coming into the mix, especially with the explosion of social media marketing and marketing automation tools. There are ever more moving pieces under marketing's umbrella, but they can end up like a 100-piece orchestra with each instrumentalist in a separate cubicle, only getting cues from the conductor via weekly email memos.

The wonderful divisibility of audience segments

There is another way to divide-and-conquer marketing: by audience segment.

An audience segment is a subset of your customers who are similar to each other and have a shared context in how they buy and use your products or services. Such segmentation might be by industry, or job function, or size of the customer's organization. Clay Christensen might characterize a segment as what people "hire" your product or service to do — different customers hire you for different reasons, and a cluster with the same reasons is often a good segment.

You can get as granular with segmentation as you want. You can often take a given segment and break it into smaller, more specific subsegments. How far you should subdivide is a function of economics and common sense.

For instance, my company, ion interactive, sells a post-click marketing SaaS. We have a segment of customers who use our product for lead generation. Within that segment, we have a segment of high-tech companies generating leads. Within that segment, we have a further subsegment just for software companies generating leads. And so on.

Two rules with such segmentation:

1. Every time you break out a new segment, you have to dedicate resources to it.

2. The more specific a segment is, the more specific your marketing to it can be — and, accordingly, the better you'll perform within it.

So you want to segment as much as you can to get the advantage of rule #2, while balancing the constraints of your available resources limited by rule #1.

For most of marketing's history, the overhead associated with managing a new segment was significant. You still had big silos of tactics and products and regions, and marketers focused on a segment were either restricted within a particular silo (i.e., different advertising to reach different segments) or had to negotiate across the boundaries of multiple silo juggernauts. Since there were very few mechanisms for coordination between silos, this was really painful. Or your segment had to be big enough that it could have its own silos — in which case, I'd argue it was more of a market than a segment.

So as a matter of practicality, most real segment-specific interactions have been left in the hands of the sales department.

"Tiger teams" for audience segments

But the rapid evolution of new marketing — with software driving everything from advertising and search marketing, through to landing pages and website optimization, followed up by marketing automation and email marketing, all tied together by analytics and a CRM — is changing the mechanics and economics of marketing operations.

With the right software and support services, it's now possible for a very small team — even a single individual — to run a full-lifecycle marketing program for a specific segment. Such a "tiger team" can directly manage everything for a particular segment:

• segment-focused social media marketing
• targeted search marketing and advertising
• targeted landing pages with deep, relevant SEO content
• follow-up marketing automation and email marketing
• personalized hand-offs to appropriate sales channel
• ongoing loyalty and relationship marketing

As long as they have the tools at their disposal (access to PPC advertising management, landing page production, marketing automation, marketing analytics, etc.) and a few key services (most importantly great graphic design) they can produce everything needed to serve their particular subset of customers.

The benefits of this segment team approach include:

• more focus for messaging and engagement with customers, since the segment team is authentically living and breathing the perspective of their narrow audience
• more continuity for customers, since all their touchpoints with the company — search, social, email, web — maintain that targeted perspective in harmony
• more agility as the team managing the segment is small and has end-to-end control over their strategy and production — minimal coordination overhead
• more experimentation as segment teams are encouraged to tailor ideas to their audience and can take risks that are relatively small in scale
• more learning for segment team members, who get visibility to the entire lifecycle of customer marketing from day one, constantly innovating at each point

These segment tiger teams can move fast. They're entrepreneurial, even if they live within the context of a gigantic enterprise. All of these benefits combine to enable much higher performance.

Can you imagine a more siloed, mass-marketing-style competitor coming up against a laser-focused segment tiger team in the battle to win a specific customer's heart and mind? The safe money is definitely on the tiger.

The segment-centric marketing organization

Once you acknowledge the feasibility — and tremendous power — of segment tiger teams, then you can imagine a marketing organization where a multitude of these segment teams thrive together at the very core of marketing. Rather than marketing being primarily defined by tactical silos, it is driven by truly customer-focused segments:

How does such a segment-centric marketing organization scale? As it grows, it subdivides its segments further and further. It can also seek out new niches that would have been impractical for large teams to address, but for which agile segment teams are perfect. This enables even big enterprises to continue to grow organically, as they have an organizational architecture that empowers them to effectively pursue a Long Tail strategy.

In such an organization, segment teams — and customers — are front-and-center. They truly deliver on the promise of integrated marketing for their audience.

The performance metric for segment teams is simple and clear: number of customers acquired — and value of those customers — relative to the cost to win those customers. All other metrics, such as CTR, CPC, social engagement points, conversion rate, fans and followers, etc., are intermediate metrics that a segment team uses to guide its efforts.

Surrounding these segment teams are a set of groups to support them:

• technology services with marketing technologists run by a marketing CTO, who provide the tools and technical glue to help segment marketers execute their ideas
• creative services that provide top-notch graphic design that adheres to common brand standards
• product management that intertwines marketing and product development by synthesizing product needs from across multiple segments
• strategic & brand oversight that provides cohesion to segment-centric marketing — but as a common substrate more than a domineering dictatorship

This strategic layer is responsible for launching new segments, sunsetting old ones, cross-pollinating best practices, and resolving contention between segments. A relatively small team provides necessary corporate marketing that spans all segments — such as a corporate web site and brand standards. It also influences unity among the segment teams via the shared technology services and creative services teams.

But it's a flat, bottom-up organization, not a top-down one. These other teams are measured by the aggregate performance of all the segment teams they support.

To be sure, there are challenges with this organizational strategy, even while it sidesteps many of the known problems with traditional marketing organizations. But if we want to really take advantage of the new marketing — its new capabilities and dynamics — then we have to do more than incrementally tweak current structures. It's time to envision a new marketing organization that leverages new marketing as a native.

The cloud threatens creative destruction for IT

The winds of change are blowing through IT, and it's not a gentle summer breeze. Earlier this month, Hewlett-Packard announced they were cutting 9,000 IT jobs. They're consolidating their data centers and making them far more automated, as more advanced software and hardware eliminates the need for so many hands-on system and network administrators.

This isn't a case of HP shutting down a business. They're still going strong as a provider of IT infrastructure services to thousands of customers. Along with the 9,000 job cuts in IT operations, they also announced 6,000 new hires with customer-facing sales and delivery expertise. According to the Computerworld article, HP officials have repeatedly said that automation, not lower labor costs, are the key to increasing margins. This certainly isn't a perspective that is unique to HP either.

Ironically, IT is becoming a victim of technological progress.

There have been two main bastions of IT value: infrastructure and applications. As demonstrated with HP, infrastructure jobs are being squeezed by consolidated service providers and ever more intelligent and automated offerings from the likes of Cisco, IBM, and Juniper Networks.

Meanwhile, the growing threat to IT-managed applications is cloud computing. Increasingly, pesky end-users with a credit card can now subscribe on-demand to just about any application they want, access it just about anywhere via their web browser, and never have to bother with IT red tape.

Gasp!

An article last month on ReadWriteWeb asked, Is IT Showing Insecurities in the Distrust of End Users? Discussing a CA report about IT's views on cloud computing and security — in a nutshell, their views could be summed up as, "run! run for your lives! end-users messing with cloud applications will bring about the end of civilization! ayyyyyiee!" — writer Alex Williams commented, "Is it us or does IT seem a bit threatened by the overwhelming interest in cloud computing?"

Alex's conclusion:

What's really at risk is IT as we know it. The IT department is in danger of becoming irrelevant. And this trend will continue if the issue is always about the dangers of the cloud.

Instead, the focus should be on learning and trust. Cloud computing points to an era of sophisticated IT networks, managed by smart, open people. If such an intelligent environment is not fostered, then IT only has itself to blame if it becomes marginalized and relegated to a back room.

Meanwhile, the rest of the world seems to accept that moving to the cloud is inevitable. A recent report by the Pew Research Center on the future of cloud computing found that, among a broader set of technology stakeholders and critics, 71% agreed with the statement: "By 2020, most people won't do their work with software running on a general-purpose PC. Instead, they will work in Internet-based applications such as Google Docs, and in applications run from smartphones."

Nicholas Carr, author of Does IT Matter? and The Big Switch — two great books for people thinking strategically about the tectonic shifts underway in IT — commented:

We don't have to wait until 2020 for this shift. It's already happened. The browser (a cloud interface) is already by far the most possible PC application, and cloud services like Facebook are the most popular computing services, whether accessed via PCs, netbooks, or smartphones. For consumers, the cloud revolution has already happened.

If you work in IT, the handwriting is on the wall: the prospects of that career path are dwindling.

But where are all the tech jobs going?

"But how can that be?!" an IT professional might protest. Isn't the world becoming more technologically driven, not less?

Yes, it is. But that technology is being delivered by outside providers, not the IT department. And most of them are not providing "IT services" per se, so much as they're offering ready-to-consume end-user applications. These companies are growing like weeds, with new ones being launched every day. I've written recently about how the technology and economics of cloud computing have already created a perfect storm for marketing technology, but the same is happening throughout other areas of the organization as well.

Now, there will still be some IT jobs out there for a while. Certainly there are a lot of specialized and custom legacy applications installed out there that aren't going away quickly. Many companies will still have some level of internal IT coordination and data management, even as their own IT infrastructure investments shift to more cost-effective cloud-based service providers.

And, hey, there will be jobs at those service providers. For instance, HP will still need IT experts on staff — but in highly optimized roles, with a much higher system-to-administrator ratio. To have one of those jobs, you'd better be one of the best of the best, and really up to speed with state-of-the-art IT automation technologies.

However, if you reposition your career — and your thinking — as a business technologist rather than an IT person, your future looks much brighter. There's growing demand for software engineers and service delivery architects at all of these entrepreneurial cloud applications ventures. And, if you're willing to embrace the art of wielding technology in a specific business domain — such as marketing, with the rise of the marketing technologist — you can turn all these cloud-based offerings into your toolbox rather than your competition.

There's no way around it: we're starting a decade of massive upheaval for IT. And, sorry, but almost everyone who doesn't work in IT is thrilled about this. (As Mel Gibson cried in Braveheart, "FREEEEEEEEEDOM!!!") But if you view this change as an opportunity for your skills and talents beyond the old definitions of IT, it can be a prosperous transformation for you too.

For better or worse, welcome to the storm of creative destruction.

When marketing optimization feels like extortion

Forgive me for this rant, but I hope this will be helpful as an example of how short-sighted, overly aggressive marketing optimization techniques can backfire with long-term brand damage.

I've been a register.com customer for about 10 years now. I know, some of you will instantly point out that they're one of the most expensive domain registrars out there. I've known that in the back of my head for a while. But I don't register many domains, the ones I have at register.com run fine, and — I admit it — I've fallen into relatively price-elastic inertia. It was just always easier to keep renewing domains there, and occasionally register new ones, for the convenience of having them all in one place.

That's not exactly a ringing endorsement of register.com, but as far as them getting my money was concerned, I fit the profile of a loyal customer.

Until today.

I had a few domains that I registered last year for a one-year term each, just to kick around a few ideas. I wasn't planning on using them anytime soon, but the option of being able to use them someday was worth a small investment. I promptly forgot about them, until I received an email from register.com today informing me that they had expired:

Okay, I figured, I'd probably be willing to re-register them for another year. So I logged into my register.com account, clicked the "renew" check box for the four domains I wanted, passed through a harmless, intermediate screen offering me web sites and email accounts, and ended up in this shopping cart:

It took a moment, but the exclamation popped out of my mouth, "$432.00?! Are you kidding me?!" (I may have used a different word than "kidding.") Expensive is one thing, but for basic domain registration this seemed like blatant gouging. So instead of going through the checkout on auto-pilot — what I was originally intending to do, and clearly what register.com was counting on — I stopped to take a good, hard look at how they came to that outrageous sum.

First, they defaulted all renewals for 2 years. That would have been fine, but when I switched the drop-down to 1 year, the price was exactly one half of the 2 year price: $35 vs. $70. Since register.com really pushes you to sign up for automatic renewal — more on that in a moment — why shouldn't customers just do one year at a time? What was the rationale to have people pay for 2 years instead of 1? Other than they were hoping I wouldn't notice?

Second, they turned their "auto renew" service on for all the domains — even though though I had it turned off for them before. While I might have appreciated the offer, having it turned on by default felt like they were sneaking in the right to ding my credit card at will without my express agreement. The button to turn it off is actually located separate from the rest of the charges, up in the right corner of the screen, so it's not obvious that it's part of your purchase. The only reason I noticed it was because I was trying to figure out the justification for the 2-year default.

Third, they added "private domain registration" to all of my domains — even though three of the four didn't originally have it. It's one thing to offer it to me, but to default to adding it into my cart for the domains that it wasn't on before seems like it crosses a line. (I never wanted it on any of my domains, so the one that had it must have been a case where their sly shopping cart practices successfully picked my pocket before.)

Fourth — and this is what pushed me over the edge — they included a $25 "reinstatement fee" on each domain because it had expired. These were not popular domains. Had they just expired and been returned to the pool, I would have been fine re-registering them from scratch, with some tiny risk that someone else might have got them. But now register.com took them over, making them theirs instead of mine (or releasing them to the world), and was holding them hostage for $100.

Fifth, they flaunted a "enter a promotion code" at the bottom, near the exorbitant $432 total, which added insult to injury. Other customers weren't being forced to pay these sky-high fees was what it said to me.

If it smells like extortion...

I certainly wasn't going to renew with register.com now and decided that I'd let them play their game until they gave up — in 30 days? 60 days? — and released them. I'd then pick them up with a different registrar.

But then I took a closer look at the email they had sent. The phrase "if you do not renew at this time... it will be made available for sale or auction to third parties." Was this implying that register.com would not simply release the domain, but instead try to sell it or auction it to some domainer?

Before I registered these names, register.com had no idea that they would be valuable to anyone. I trusted them implicitly by choosing them as my registrar, saying privately to them with my action that this domain might have value someday. And now they seem to be leveraging that private information I shared with them against me. Because I trusted them — and paid a premium with that trust in mind — they were now in a position to extort money from me. $100 today, some threateningly unknown amount tomorrow?

Take the domains, register.com. But I'm pretty sure whatever money you can squeeze out of them on whatever black market you have in mind is going to pale in comparison to the yearly recurring revenue you've lost by losing my business. We can quantify that, and my LTV to you was easily thousands of dollars. What is less quantifiable, but probably more damaging, is the bad word-of-mouth I'll spread about this experience whenever someone asks me about domain registration. (We won't count this post, as I'm writing it primarily so other companies can avoid ticking off their customers this way.)

Marketing optimization lessons to learn

There's a possibility that register.com is genuinely evil. But there's also the possibility that this outcome was the inadvertent result of a series of marketing optimization initiatives that, collectively, lacked governance on behalf of the brand and overall customer relationships.

It's all too easy to fall into what conversion-oriented usability guru Sandra Niehaus calls evil conversion optimization.

Most conversion optimization projects are short-term sprints to increase the conversion rate and the average order value. So you A/B test different versions of emails — one with a more ominous set of consequences if the customer doesn't act now — and if the more threatening one pulls a high conversion rate, it wins adoption. You A/B test defaulting more stuff into the shopping cart to measure the trade-off between average order value and cart abandonment. Mathematically, you can calculate the optimum.

But it's a dangerous, short-term optimum.

What you can miss in such short-sighted optimization is how the people who abandon feel. Did they abandon because they were unqualified or not ready to purchase? Or did they abandon because they were livid, seething about your practices that they felt were unconscionable? Those two outcomes can appear identical as far as short-term conversion metrics are concerned, but the larger impact on your brand in those two cases is night and day.

Even for people who do convert, it makes a big difference if they're converting because they are happy with the purchase, or if they're converting because they feel they're being mugged at gunpoint. Such resentment builds up over time, and by the time someone reaches the breaking point — swearing off your service and loudly denouncing you to other prospects and customers — it can be hard for the company, in aggregate, to attribute the failure to a specific sequence of conversion optimization tests.

Now, I'm all for marketing optimization — as long as the means are watched as closely as the ends. I think in every conversion optimization project, at least one person should play the role of the customer advocate — or, if you prefer, the brand advocate — to make sure the methods being used to increase your conversion metrics don't accidentally fall to the dark side.

It's hard for marketing optimization projects to directly measure lifetime value (LTV), because often they iterate on such a short cycle. But common sense, brand wisdom, and good governance can balance that objective with short-term tactics.

For businesses that hope to be around a while, LTV and ongoing customer goodwill are the real goals. Don't do to your customers what register.com just did to me.

3 nimble trends changing content and marketing

Razorfish just released a great new report, Nimble: Publishing in the Digital Age, written by Rachel Lovinger. The target audience is clearly large publishers and mega media companies, but there are important trends here that are relevant to everyone who thinks of distributing content on the web — e.g., anyone in marketing.

The three key points that pop out to me — particularly because I keep seeing them in a lot of posts and discussions these days:

1. The concept of "the web site" is becoming less important than the underlying content and its social propagation.

People now consume online content on a variety of devices, from their computers, their iPhones and Android phones, and new tablets like the iPad. They consume content not necessarily from the site of the content producer, but from social media hubs such as YouTube, social networks such as Facebook, aggregators such as Alltop, and a wave of social sharing services such as Twitter, Buzz, LinkedIn from which it is reblogged and reposted (or reshared as the term du jour).

You need to design your content — and the process and metrics by which you generate and maintain that content — to be effortlessly absorbed into this malleable and organic distribution engine in the cloud. It's a different paradigm than the closed walls of the 1990's web site.

Quoted in the report is Nic Newman of the BBC:

You can't afford to [create] a piece of content for any one platform. Instead of crafting a website, you have to put more effort into crafting the description of an asset and the different bits of an asset, so they can be reused more effectively, so they can deliver more value.

2. Structured data and metadata around content are becoming increasingly important.

As Facebook's Open Graph protocol is demonstrating, even just a little bit of semantic markup associated with your content can go a long way in making your content more friendly to find, share, and curate. This makes it much easier for you — and others — to adapt your content in a variety of containers, beyond the static information architecture of your web site.

Forget the fancy names of "semantic web" or "linked data." Associating structured data with your content assets lets you take advantage of Open Graph, Google RichSnippets, Yahoo Search Monkey, and a new generation of agents such as Siri. Disseminating your content with metadata through APIs enables developers to spread the seeds of your brand in a variety of mash-ups and apps. Sharing your data sets in collaborative venues such as Factual and Infochimps helps build relationships with the world of analytic power users, improve your data quality, and turn those dusty data silos into tools for advocating ideas and brands.

And with the coming data explosion (already 281 exabytes of online data in 2009!), this type of self-guiding roadmap within and around your assets will soon be essential for even you to find and use the right content in the right place.

3. Content producers must become much more nimble (i.e., agile).

When many people come to the same conclusion from independent streams of research, you know a major movement is in progress. Rachel's overarching advice to publishers is to push to be more nimble — or as I'd call it, more agile. And, as you've probably noticed, a lot of people have been talking about business agility this year.

I've been banging the drum for more agility in marketing — such as a call for an agile marketing manifesto — because in a world where everything is in constant flux, your only reliable source of competitive advantage is your ability to adapt. Fast. That's easy to say, but it requires cultural and managerial shifts that are still quite foreign to most organizations. It's a change in metabolism that needs to be taken seriously and mandated from the highest executive levels.

As Rachel writes in her report:

To succeed as a digital content producer you need to be nimble. Many companies will say that they're nimble, but very few actually are. Being nimble is about the ability to adapt quickly to the new challenges and opportunities in today's media ecosystem; things like the explosion of new media devices, the world-domination of social networks, and consumers' growing expectation of first-class digital experiences.

Being nimble is not just about an organization. It's about the industry's business models. It's about production processes. It's about the content itself.

Nimble is a thought-provoking report, well-worth reading and considering how it applies in your domain.

(It also includes a reference to my post on business models for linked data and Web 3.0, which was inspired by discussions and feedback from Rachel as well.)

The Marketing Technology Manifesto

I continue to receive terrific feedback about my presentation and essay on the rise of the marketing technologist — thank you to everyone who has contributed comments, anecdotes, and ideas.

The pent up desire for marketing to take control of its technological destiny is clearly bursting at the seams.

You've inspired me to take that 3,000 word essay and 54 slide presentation and try to boil it down to its essence, in 153 words. In the spirit of revolution, I call it a manifesto:

The Marketing Technology Manifesto

Modern marketing is powered by software.
Software to plan and design new programs.
Software to execute and deliver those programs.
Software to measure and optimize everything.
This is more than information technology.
It is creative, strategic, and brand-defining.

Software is redefining marketing's capabilities.
Marketing is now faster and more granular.
Marketing is more connected and always on.
Marketing is more accountable and efficient.
Marketing can even be more useful and engaging.

Marketing must master this software.
Marketing must build and buy it.
Marketing must mix and mash it.
Marketing must lead and leverage it.
Such leadership cannot be relegated to IT.
Such leadership cannot be outsourced.

Marketing must control its technological destiny.
Marketing must employ its own technologists.
Marketing must have a "marketing CTO" leader.
Marketing must tap software and data beyond IT.
Marketing must entwine technology in its DNA.

The new culture of marketing is a mash-up:
Strategy, creative, technology, and innovation.

Trust, brands, and the marketing dashboard of the future

This week, I ran a pilot of my latest study for improving trust in online advertising — the first time I've run an experiment with human subjects physically in the lab.

Two thoughts I wanted to share from the experience:

First, don't ever underestimate the power of brands. I'll share more about that once we finish the study. But search marketers in particular, I cannot stress this enough: brands are immensely influential in the context of search.

Do everything in your power to build and protect your brand — and the trust entwined within it — at every touchpoint in your marketing universe and product/service delivery.

Second, wow, the setup at the Harvard Decision Science Laboratory for running such experiments is wicked cool (as they say here in Boston). Subjects participate at individual workstations with video monitoring, two-way audio, physiological monitoring (heart rate, skin temperature, sweat monitor), options for group interactions — and it's all controlled from a Star Trek like touchscreen console.

As the researcher, you can zoom in on any participant, watch their computer interactions, observe their video stream, take over their keyboard to assist, initiate private or group audio dialogs, control environmental factors such as lighting, etc.

It inspired me to imagine what a marketing dashboard of the (not-too-distant?) future might look like...

7 Capabilities for the Marketing Dashboard of the Future

Too much of web analytics is dry, dull, and disconnected from the real-time experience of actual prospects and customers. Social media monitoring is starting to explore this frontier with sentiment analysis and visualization of more in-the-moment activity. But we still have a long way to go.

Here are 7 things that I'd like the marketing dashboard of the future to do:

1. Show me aggregate, high-level information — as it does today — but in real-time.

2. Sentiment analysis is great and getting better everyday — check out Google's new Prediction API — but I'd like to monitor a slightly different metric with trust analysis. What are people doing that signals the strengthening or weakening of trust in my brand?

3. Let me zoom in, ideally down to the individual experience, to "see" users follow a click path in real-time, observing their typing and mouse movements (via Javascript), to get an authentic feel of what people are actually experiencing. At one level up, let me see the threads of atomic marketing.

4. Much, much better visualizations — there's way more to life than line graphs and pie charts. For example, the visualizations produced on sites such as Many Eyes and the latest work by Martin Wattenberg of Flowing Media.

5. The ability to quickly curate and share insightful snippets of such data, possibly through services such as Infochimps, Swivel, and Factual.

6. Enable and promote test-driven marketing by incorporating A/B tests and multivariate testing (MVT) into the display of everything — encourage comparisons between alternatives and spark new test ideas. Give the power of testing executive-level visibility.

7. The capability to react immediately by changing relevant advertising, offers, landing pages, etc., interact with social media threads, engage staff — springboarding directly from my dashboard. This isn't just agile marketing. This is super-agile marketing.

What would you like the marketing dashboard of the future to do?

How soon until we have it?

3 big picture marketing themes from #ConvCon

Last week was the inaugural Conversion Conference (#ConvCon), a new event organized by landing page optimization expert Tim Ash and co-located with the eMetrics marketing optimization summit. While there were plenty of terrific sessions on landing pages and conversion optimization — my post-click marketing craft and trade at ion — three larger marketing themes emerged that I want to share with you.

Theme #1: Agile Marketing

In conversion optimization — what marketers do to increase lead generation or e-commerce sales on the web — there are two perennial secrets to success.

The first is the ability to test, for instance, to try two different landing pages in an A/B split test. You send half the respondents who click on an advertisement to page A and the other half to page B, and then measure which of those two groups has the highest conversion rate.

The second is the ability to deploy specific landing pages to match particular advertisements, emails, or social media promotions. So if I click on an ad for a romantic getaway in Tuscany, it works much better if I land on a page that talks specifically about a romantic getaway — with content and images and special offers tailored to that idea — rather than the generic home page of a hotel chain in Italy. (I've written before about the proliferation of such micromarketing in my post on landing pages as atomic marketing.)

Both depend on how quick and easy it is for you to deploy new tests and new pages. As a result, many speakers emphasized agile marketing practices — if not with that label, certainly with its essence.

The message: conversion optimization is all about agility.

As conversion optimization gains more executive recognition — which it is quickly doing, due to the ROI it delivers in the age of marketing accountability — it will help motivate broader adoption of agile methodologies in the marketing department. Agile-marketers-to-be: take advantage of this opportunity to start changing your organization's marketing metabolism.

Theme #2: Test-Driven Marketing

However, just because you can run tests doesn't make you test-driven.

Probably the biggest theme at the conference was developing a culture of testing in the marketing department. Marketers, perhaps more than anyone, have passionate and creative opinions about what will work in the market. In old-school marketing, the highest-paid person's opinion — the infamous HiPPO that Avinash Kaushik rails against — would take precedence. The boss would pick the winner.

(Imagining that bosses are inherently good at picking winners is a classic example of survivorship bias and being Fooled by Randomness.)

In a test-driven marketing environment though, such creative ideas are only the start of the decision-making process. Instead of one idea being picked by the loudest or highest ranked advocate, many ideas are tested to learn which one customers like best — and that's the empirical winner. Executives embrace the process, rather than the picking, as the mantle of their leadership.

But test-driven marketing goes further than merely arbitrating internal debate.

Test-driven marketing is about continuous improvement, iterating through experiments as a way of life, rather than a one-shot project. Chris Goward of WiderFunnel gave a fantastic presentation at the conference on applying the Kaizen method (popularized by the Toyota Production System) to conversion optimization. I think this is a brilliant cross-pollination of ideas from engineering and operations into digital marketing.

Chris and the team at WiderFunnel have also developed their own LIFT model for landing page improvement that demonstrates the power of a domain-specific framework for managing the direction of such continuous improvement.

Brooks Bell, the founder and president of Brooks Bell Interactive, presented on the power of A/B split testing and recounted her company's start with optimizing banner advertisements for AOL. She described how AOL was obsessed with testing — a company worldview that in no small part contributed to their one-time domination of the market through massive customer acquisition. AOL may have eventually been eclipsed by The Innovator's Dilemma, but their culture of testing was pioneering.

One of the anecdotes that Brooks relayed was how AOL built a technology team directly in the marketing department as a way to accelerate testing and push the boundaries of new ideas. Is it possible that they were the first company to prefigure the rise of the marketing technologist?

Theme #3: Insight vs. Automation

One of the most interesting debates in this next decade of new marketing will be the tension between human insight and software automation. On one extreme will be fully "black box" computational marketing, where computer algorithms take over the entire decision process of what marketing to present prospects and customers at every contextual opportunity. The other extreme, which will no doubt be driven by a blacklash to problems with pure algorithmic marketing, will keep people directly in charge of most marketing execution, relying on their judgement and creativity, albeit at the cost of more limited scale.

Of course, the most compelling solutions will be those that thrive in between the two extremes, seamlessly entwining the capabilities of people and machines. Think of the bionic man more than The Terminator.

Conversion Conference offered a sneak preview of that debate. Some vendors who have specialized in highly technical fractional factorial multivariate testing (MVT) continued to emphasize their ability to test thousands and thousands of possible page variations of a web page to (theoretically) find the most optional combination.

But most other speakers, who had real-world experience implementing such approaches, championed simpler but more meaningful tests. They argued passionately for intuitive and well-hypothesized A/B split tests, maybe followed by smaller full factorial MVT experiments that properly identified interaction effects.

Bryan Eisenberg, the wonderfully opinionated author of Waiting for Your Cat to Bark, showed outright disdain for the notion of throwing millions of combinations against the wall, hoping for a miracle somewhere in the jumble. He called it slice and dice optimization and claimed that all it did was burn people out on testing without achieving real results.

Lance Loveday, CEO of Closed Loop Marketing and co-author of Web Design for ROI, made a strong case for not just testing, but testing among good options (TAGO). What is "good?" That's where human insight and creativity come into play.

Neither Bryan nor Lance are Luddites by any stretch of the imagination — they're all for using every tool they can find to help with their conversion optimization mission. But they're fervent advocates for leveraging such tools with human intelligence, not artificial intelligence.

The battle lines in this debate are in flux though, so it will be exciting to see how it evolves.

Intrigued? There will be another Conversion Conference in Washington, D.C. in early October. Hope to see you there!

15 sessions on marketing and the semantic web

Looking for a conference that really pushes the envelope of digital marketing?

Last year, I gave a talk at the Semantic Technology Conference on marketing in the semantic web. At the time, the topic was in its infancy. But as predicted, a lot happens in one year. This June, at the 2010 Semantic Technology Conference in San Francisco, there is now a whole track dedicated to this subject, Marketing in the Web 3.0 World. And, I humbly say as the track's chair, it's going to rock.

With Facebook, Google, Yahoo, and Bing now all leveraging semantic web data — Facebook's announcement of Open Graph being the latest major development — the timing couldn't be better.

While the overall conference — which runs a full week — dives deep into the underpinnings of the semantic web, this track will focus on the higher-level applications of this technology in marketing and advertising. It's still not for the faint of heart, but it will be the "real deal." If you want to be on the cutting edge of leveraging linked data and the semantic web in your digital marketing, this is for you.

Here are a few highlights of the 15 sessions in this marketing track:

David Siegel, author of the terrific book Pull: The Power of the Semantic Web to Transform Your Business, will give the keynote. David has been one of the leading advocates for the potential of the semantic web to change business — tapping into the accelerating shift from push to pull marketing models.

Kavi Goel from Google will give an hour-long presentation on Google Rich Snippets — what promises to be one of the most thorough examinations yet of how marketers and publishers can embed structured data into their web pages to make them stand out in Google search results. Tactically, I expect this session to be worth its weight in gold.

Jay Myers of BestBuy.com — whose e-commerce success with linked data has been widely hailed — will join Barbara Starr and Paul Bruemmer who worked on Overstock.com for a session on leveraging RDF and the GoodRelations vocabulary in e-commerce SEO. These presenters have tremendous real-world experience with the state-of-the-art in semantic web and search engine optimization.

Gil Elbaz of Factual and Dhruv Bansal of Infochimps will join me for a session that steps back from the semantic web and looks more broadly at using the web to share data — and collaborate on it — with customers and partners. This is Data as the New Marketing Channel, and when you see the examples these presenters have, I know you'll be as inspired as I am.

David Bean of Attensity, Shion Deysarkar of Extractiv, and Dmitri Soubbotin of Semantic Engines will give a powerhouse presentation on sentiment analysis and marketing intelligence powered by semantic technologies. These are three brilliant minds on the leading edge of social media monitoring, who will show you what's possible today... and in the near future.

And there's so much more:

• Ian Davis and Michael F. Uschold on linked data business models
• Brooke Aker, Amiad Solomon, and Ian Saunders on semantic advertising
• Chris C. Arning on semantic technologies for market research
• Jamie Taylor on the social fabric of semantics
• Mark Leyden on the legal implications of linked data
• Lars Hard and Brooke Aker on mobile semantics

If I could bring together two dozen of the most fascinating figures in the semantic web and its intersection with marketing, this would be them. I'm thrilled that they will all be there in San Francisco for some terrific talks and discussions.

If you'd like to join us, you can use the special code ST10SPKR to save $200 off registration.

P.S. If this is all new to you, you might want to start by taking a look at the presentation I gave earlier this year at the Web 3.0 conference on Data Marketing for Web 3.0. There's also a video interview available as well.

Crowd sourced ads in Google -- outside Google's control

What if you could enhance your ads in Google with collective intelligence — without going through Google?

Earlier this month, browser bookmark extension company Xmarks unveiled their plans to monetize their business with a service called SearchBoost. They take the idea of Google including ratings in ads to a whole new a level.

Here's how it works:

Xmarks is a free browser add-on for Firefox, IE, Safari, and Chrome. It was originally designed to synchronize browser bookmarks across different computers — so bookmarks on your desktop, laptop, home machine, etc., are always in sync and backed up in the cloud.

As browser add-ons go, it's pretty popular. They have over 4 million active users, and Xmarks claims to manage over 1 billion bookmarks. (As an interesting note, the Xmarks' chairman and co-founder is none other than Mitch Kapor.)

But what makes Xmarks cool was their insight that the bookmarks of all their users could be treated as "votes" for what people liked — sharing the collective intelligence of good sites. Initially, Xmarks used this information to enhance the organic results in search engines and provide additional information about a site when you were on it.

But with the announcement of SearchBoost, Xmarks now lets advertisers pay Xmarks to include these crowdsourced ratings in their ads on Google and Bing.

Three aspects of this that I find fascinating:

1. Xmarks is changing Google and Bing advertisements without permission of the engines. Although add-ons could always do this in theory, this is the first time that I know of a company that intends to do so for financial gain. In an era where iPhone and Android apps are increasingly serving as customized intermediaries to content on the web, this is an interesting precedent.

2. Advertisers don't fully control their ads anymore. An advertiser controls turning SearchBoost on or off, but they don't have any influence over the crowdsourced ratings themselves. Obviously, only advertisers with good ratings will want to turn this feature on — but since such ratings are dynamic, advertisers are leaving their fate in the hands of the community. But in the era of social media, I think this is a great way to increase the trustworthiness of advertising via social proof.

3. Xmarks provides a third-party audit of Google and Bing SERPs and their ads. At least for the sample of users who have Xmarks installed, advertisers can receive end-to-end analysis of impressions and click-throughs — for both their ads and organic listings — and their competitors. Since neither Google nor Bing provide that level of transparency, this is valuable intelligence for marketers.

Now, of course, all this only works for users who have installed Xmarks. But according to the company, their 4+ million users conduct about 500 million searches per month — which for ads targeting the demographics of people who tend to install add-ons (technically-oriented users) is not an insignificant audience.

Xmarks claims their enhanced listings result in a 15% increase in CTR.

I'll be curious to see how this works for Xmarks — and how those three aspects of SearchBoost appear and evolve elsewhere in the larger advertising ecosystem.