Data Web Marketing

Marketing in the Semantic Web(“Semantic Marketing” / “Data Web Marketing”)

Scott BrinkerMarketing Technologist

Email: [email protected]: @chiefmartecBlog: http://www.chiefmartec.com

Semantic Technology Conference

June 16, 2009

Deafening silence

Sweet sorrow

Controlled chaos

Organized mess

Open secret

Same difference

Civil war

Forward retreat

Living dead

Semantic marketing

What will marketing be likein the semantic web? *

* Depends on your definition of “marketing” and “semantic web”.

Official definition of marketingfrom the American Marketing Association web site

Marketing is the activity, set of institutions, and processes for creating, communicating, delivering, and exchanging offers that have value for customers, clients, partners, and society at large.

(Approved October 2007)

Informal definition of marketing

Marketing is what you do to find and win new customers, grow your relationships with existing customers, differentiate yourself from the competition, and build a “brand” that helps achieve those goals.

from the top of my head

Peter Drucker on marketing

Because the purpose of business is to create a customer, the business enterprise has two — and only two — basic functions:

marketing and innovation.

the father of modern management *

* Drucker argued in a 1984 essay that CEO compensation should be no more than 20 times what the rank and file make — especially at companies

where thousands of employees are being laid off. “This is morally and socially unforgivable,” he wrote, “and we will pay a heavy price for it.”

Marketing is continually evolving.

In recent years, that evolution has been accelerating — with more changes ahead.

Marketing as a missionspans the specific tacticsby which it is executed.

Marketing tactics,circa 1900

Marketing tactics, circa 2000

100 years of progress?

Emerging marketing tactics, circa 2009

$1.4 billion of SEO in 2008

1,254 APIs and 3,852 mashups

Search engine optimization (SEO) and web APIs for mashups are a

qualitatively different kind of marketing tactic.

Even across shifts in medium…

…previous marketing

tacticswere

crafteddirectly

forhuman

consumption.

1st order:Directly crafted for computer consumption.

2nd order:Indirectly crafted for human consumption.

These new tactics are different.

This opens the door to

data web marketing.

What is the semantic web?

The Semantic Web is a web of data.

The Semantic Web is about two things.It is about common formats forintegration and combination

of data drawn from diversesources… It is also about

language for recordinghow data relates toreal world objects.

from the W3C web site

iro•ny, noun, \ˈī-rə-nē\:Debating the meaning of “semantic web”.

What is the semantic web?3 broad spheres

Document Disambiguation

• Semantic technology that doesn’t necessarily require publisher cooperation

• Advances in text analysis for context and sentiment

• “Semantic advertising”(popular interpretation)

• Usually invisible to end-user

• Top-down semantic web

• Here today

…no fundamental change to marketing behavior, however.

LinkedData

Structured Data

Semantic marketing is about data—and the spread of that data.

Semantic Marketing

=

Data Web Marketing

* Is this a better name for it?

Marketing loves data.

…but it has only flowed in.

• CRM• Point-of-Sale• Market research• Web analytics

…really flowed in.

Semantic technology can help organize this data… but again, no fundamental change to marketing.

Data web marketing is about data flowing out.

A simple yetrevolutionary change in perspective.

Instead of data from the channel…

Data is the channel.

Marketingwill makedata sexy.

Why should marketing lead the charge for data web adoption?

Somebody needs to consistently beat the drum for data web initiatives.

Somebody needs to fund data web support as an ongoingcommitment.

Who has the incentive?

If the data web can be used to:

• help connect to new customers

• strengthen relationships

• differentiate from the competition

• build reputation and brand

Who does that align with?

Informal definition of marketing

Marketing is what you do to find and win new customers, grow your relationships with existing customers, differentiate yourself from the competition, and build a “brand” that helps achieve those goals.

from the top of my head

Marketing is willing to experiment to achieve its goals.

If marketing can tie the cause-and-effect of data web initiatives to the achievement of its objectives…

…there’s sustainable sponsorship.

Marketing success will cause business to embrace the data web.

So what exactly should marketing do in the data web?

Missions for marketing in the data web.

1. Champion production of data for external consumption.

2. Drive semantic/data branding across the organization.

3. Distribute and promote your data. (SEO++)

4. Convert data web initiatives into business relationships.

5. Track and attribute semantic/data web initiatives.

6. Make your own data mash-ups.

7. Control data quality and protect data/brand standards.

7 missions of data web marketing:

#1.Champion productionof data for external

consumption.

Not just “brochure data.”

Not (primarily) pricing data.

Semantic “bargain hunter” agents are not attractive to most marketers.

That vision of the semantic web is dystopian to marketers.

Marketing is all about avoiding commoditization.

• Price isn’t everything.

• Only one lowest-cost provider.

• Race to the bottom dynamics.

• Specs aren’t everything.

• Relationships have value.

• Quality matters.

• Context matters.

• Service matters.

• Trust matters a lot.

Pricing and product specifications don’t do justice to the potential of sharing data in the semantic web.

Discover data that is valuable in:

• Domain of your expertise

• Domain of your partners’ expertise

• Domain of your customers’ expertise

• Application of your product/service

• Integration of your product/service

• Benchmarking related results

“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.”

Tim Berners-Lee in Talis 2008 interview, answering the question from a CIO, “what does it mean, what should we do?”

—

Content Data is King

Ways to produce valuable data:

• Generate it internally

• Collect it from customers

• Collect it from partners

• License it externally

Thinking about this kind of data is hard — because it’s not been done before.

But that’s the opportunity.

Hypothetical example:

Major chain of nurseries producing the leading reference of plant properties (climate, growth, soil, water, feeding, compatibility, etc.) — maybe specialized for a particular region.

Hypothetical example:

Marketing software company aggregates performance data across customers to offerreal-time industrybenchmarks.

(With permissions of participants, of course.)

Look to existing “mash-up” APIs as inspiration for data web ideas.

Goals:

• Become the authoritative source.

• Popularize canonical references to your products, categories, competitive dimensions.

• Build reputation, goodwill, brand.

#2.Drive

semantic/data branding

across the organization.

Framing data with the right metadata — your data brand standards:

• Establish canonical URIs for products, properties.

• Establish the organizing ontologies.

• Determine the ideal granularity of data structures.

• Embrace and extend existing external standards.

• Encourage data linking in the organization (DRY).

• Lobby for standards beyond your organization.

• Maintain and evolve this architecture.

Building for linked data is an art.

Linked data success depends on:

• Consistency

• Logical organization

• Stability

• Trust

Data consumers must be able to rely upon your data to use it as a foundation for their own applications.

Ontology as

Strategy

Goal:Shaping the conversation in your market space.

Semantic spin

Framing references with properties and nomenclature that are advantageous to your brand.

Do the names of things really matter that much?

Ask these brands:

Our company is redefining the market.

A more literal interpretation of an old hype line.

First mover advantage.

Authoritative references generate a

positive feedback loop, a virtuous cycle.

Gender bias in the semantic web?

A kind of semantic branding.

http://www.readwriteweb.com/archives/will_the_semantic_web_have_a_g.php

There are fascinating parallels between the concept of brands and the semantic web… but that’s a story for another day.

http://www.chiefmartec.com/2008/03/brand-and-the-s.html

#3. Distribute and promote your data. (SEO++)

Just as SEO was about visibility (and ranking and authority) in the document web…

…there will be an analogous need in the data web.

SEO++because it’s an incremental evolution of SEO practices now focused on data objects.

Alternatives:Semantic Web Optimization (SWO)Data Web Optimization (DWO)

Spreading your data:

• Build external links to your data

• Link reciprocally to other data

• Increase the findability of your data

• Optimize the format of your data

• Shape and adopt standards

• Promote your data in other channels

As with SEO, this mission will require continual nurturing.

You want to join the data graph.

You want your data to be utilized.

You want your data to help others find you.

For example, embedding data into your primary web site, such as for Google Rich Snippets and Yahoo! Search Monkey.

And that’s going to get harder as more data comes online.

Surface web:167 terabytes

Deep web:91,000 terabytes

545-to-one

Bridge your data with others in semantic communities (data networks).

Networks:

• Global

• Vertical

• Private

Provide the glue to connect to these different networks.

This will probably be a little messy for a while.

A new kind of semantic advertising

Paid inclusion in other authoritative data networks.

A.K.A.data advertising

Hypothetical example semantic advertising (data advertising)

Sponsored data

I just made this idea up, so whether or not Calais actually does something like this is purely coincidental.

semantic advertising (paid inclusion)

Goals:

• Include your data in more places

• Get links to your data in more places

• Win more overall visibility/authority

Extend your reach.

Data web marketing services are a logical evolution for

search agencies.

#4.Convert data

web initiatives into business relationships.

Information asymmetry will still be alive and well.

Data web marketing must strategically decide how much to share, when, and with whom.

Some data may be better harnessed as an incentive for other business goals:

• Become a subscriber

• Become a lead

• Become a partner

• Become a customer

• Become a data buyer

availability

valu

e

access if you pay

access if you qualify

access if you sign-up

free access

A continuum of data access choices.

Data for nothing and links for free.(the SEO++ approach)

Capturing value via visibility and authority.

“open data”

Restricted “members only” data.

Exchange of value in permission marketing, or an added benefit to customers and partners.

Capturing value via leadgeneration and customeracquisition/retention.

Data as a direct revenue source.

• Data more pragmatic in standardized format.

• Paid data access as a stand-alone business.

• Paid data access as a “add-on” business.

Capturing value the old-fashioned way: people pay for it.

Balancing the trade-offs in data value capture is a marketing decision.

Multiple “data packaging” options.

• Granularity of data.

• Depth of data.

• Breadth of data.

• “Freshness” of data.

The same underlying data may be packaged differently depending on access level:

Goal:Harness the value in your data.

#5.Track and attribute

semantic/data web initiatives.

“data web analytics”

How do you measure the success of semantic web/data web initiatives?

What are the right metrics?

Different than web analytics because…

…clients are not necessarily browsers.

Cookies—a staple of web analytics—may not be as prevalent in data access tracking.

Referrer—a staple of web analytics—may not be as prevalent in data access tracking.

Tracking is also going to be hard due to mashing and caching.

How is data used and redistributed once someone gets it from you?

Count subscribers to data feeds or visits to URIs.

• Measures 1st-order reach.

• Measures frequency of access.

• Measures new vs. repeat access.

• But maybe limited to IP address.

Time-sensitive or frequently updated data is one way to encourage more visits to

gauge usage.

Count inbound links to your URIs.

• Measure authority (DataRank).

• Measure findability.

• Limited by who indexes you.

New methods of tracking and attribution?

Particularly among cooperative parties?

Inspiration from Daniel Weitzner (MIT) and the policy-aware web (PAW).

Goals:• Discovering what data is popular.

• Discovering who your data audience is.

• Discovering how your audience uses that data.

• Keeping track of competitors and comparable benchmarks.

#6.Make

your own data

mashups.

Data web fluency is something you learn by doing.

You must use data to understand how to use data.

Applications and mash-ups are where data surfaces into the visible web.

Leveraging data in your own value-add mash-ups.

mashable marketing

• For prospects

• For customers

• For partners

• For internal use

Mash-ups for prospects and customers.

To assist, educate, entertain, orinform.

Mash-ups beyond Google Maps.

Mash-ups for internal use.

• Market research

• Customer monitoring

• Marketing operations

Competitive intelligence mash-ups.

Uncovering the pros and cons of data web marketing.

Opportunities for “joint venture” data web initiatives — your chocolate with someone else’s peanut butter (exclusively?).

Goals:

• Cool applications for your customers.

• Advanced your own internal operations.

• By doing the above, better understand how data is consumed to be better at producing it.

#7.Control data quality and

protectdata/brand standards.

“semantic police”

Data web marketing won’t be magic.

• Coordination challenges with distributed data management.

• Rules about what can be shared, when and with whom.

• Maintaining the accuracy of data (i.e., data entropy).

• Refereeing conflicting data silos coming together.

• Enforcing data brand standards.

Legal questions:

Do we have the rightto share certain data?

What are the liabilitiesfrom sharing data?

Does sharing certain data constitute a risk to our intellectual property?

Determining how much data to share…

…or not to share.

“Legal”

“Marketing”

Agreement on data standards may be contentious among stakeholders.

When data was in silos, inherently fewer conflicts. As data web marketing grows, this will become a larger issue.

Remove bad or expired data.

It’s much more unattractive when the public has come in for U-pick-it data.

Detect misuse and data theft.

Goal:

Achieve balance between openness vs. protection, distributed vs. controlled, standardized vs. loosely-coupled data relationships.

1. Champion production of data for external consumption.

2. Drive semantic/data branding across the organization.

3. Distribute and promote your data. (SEO++)

4. Convert data web initiatives into business relationships.

5. Track and attribute semantic/data web initiatives.

6. Make your own data mash-ups.

7. Control data quality and protect data/brand standards.

7 missions of data web marketing:

Web 3.0 = Data Web 1.0

A more technical future for marketing?

A role for “marketing technologists” in the organizational DNA of the marketing department & agencies.

The new leaders of data web marketing?

Thank you for running this marathon presentation with me.

Scott BrinkerMarketing Technologist

Email: [email protected]: @chiefmartecBlog: http://www.chiefmartec.com