Re: Business Models, Profitability, and Linked Data

On 6/10/13 9:06 AM, Leigh Dodds wrote:
> Hi,
> On Mon, Jun 10, 2013 at 12:00 PM, Kingsley Idehen
> <> wrote:
>> On 6/10/13 4:18 AM, Leigh Dodds wrote:
>>> Hi,
>>> On Fri, Jun 7, 2013 at 5:52 PM, Kingsley Idehen <>
>>> wrote:
>>>> There have been a few recent threads on the LOD and Semantic
>>>> Web mailing lists that boil down to the fundamental issues of
>>>> profitability, business models, and Linked Data.
>>>> Situation Analysis
>>>> ==================
>>>> Business Model Issue
>>>> --------------------
>>>> The problem with "Data"-oriented business models is that you
>>>> ultimately have to deal with the issue of wholesale data copying
>>>> without attribution. That's the key issue; everything else is
>>>> a futile dance around this concern.
>>> Why do you think that attribution is the key issue with data oriented
>>> businesses?
>> Its the key to provenance. It's the key making all contributors to the data
>> value chain visible.
> I don't disagree that attribution and provenance are important,
> especially for Open Data, but also whenever it becomes important to
> understand sources of data.


>> As I've already stated, the big problem here is wholesale copying and
>> reproduction without attribution. Every data publisher has to deal with this
>> problem, at some point, when crafting a data oriented business model.
> Every data publisher that aggregates or collects data from other
> sources certainly needs to understand -- for their own workflow --
> where data originates.

Not just for their own workflow. They have to attribute the data sources 
in a form discernible to user agents (and other consumers) further up 
the value chain.

>>> I've spoken with a number of firms who have business models based on
>>> data supply and have never once heard attribution being mentioned as
>>> an issue for themselves or their customers. So I'm curious why you
>>> think this is a problem.
>> And are those data suppliers conforming to patterns such as those associated
>> with publicly available Linked Open Data? Can they provide open access to
>> data and actually have a functional business model based on the
>> aforementioned style of data publication?
> No they weren't using Linked Open Data. No they weren't publishing
> open data (it was commercially licensed for the most part). But they
> all had successful business models.

I can't comment on this until you provide a specific example.

> But I understood you to be making a general statement about a key
> issue that is common to all data business models, one that Linked Data
> then solves.
> I agree that every data aggregator needs to understand their workflow,
> to manage their own processes. I agree that publishing details of data
> provenance and attribution is important, particularly for Open Data.
> And absolutely agree that Linked Data can help there.
> Maybe I'm misunderstanding your point but I'm not seeing evidence that
> attribution is a key business issue that data businesses have to solve
> in order to be successful.

It's kind of like the GPL re. open source code. It triggers a 
negotiation between the license holder and those seeking to use the code 
in manners outside scope of the GPL license terms.

Let's say you produce a really cool source of data for a specific 
domain, attribution ensures your specific role in the value chain is 
clear. It means that I can go directly to you for a variety of usage 
patterns associated with the data you produce. Net effect, you can 
enforce specific terms or discover new models that encourage you to 
continue to produce and publish your data.

Ultimately, you can opt to create different usage models for different 
profiles. This is where access control lists and data access policies 
come into play. Of course, this also depends on your ability to identify 
the consumers and constrain their access to terms expressed by your data 
access policy and enforced by access control technology.

Simple real-world example:

I have a Linked Data middleware service that only allows certain 
identities exploit its Linked Data transformation services. A single 
link behaves completely differently, subject to the identity of the data 
consumer. The identity is handled via a WebID and the actual data access 
policies take the form of entity relationship semantics expressed in a 
policy graph which my service leverages.

> You said that "everything else is a futile
> dance around this concern" which I found surprising, so I'm curious
> about the evidence.

I am saying that there is a futile dance around the fear of copying. 
Data producers are ultimately afraid of data being copied wholesale 
without pathways to its origin. Basically, the original data publisher 
is dislocated from its emergent value chain.

> I'm curious about the general business drivers,
> regardless of whether the data is Linked or Open.

The general business drivers boil down to being compensated for useful 
data where the fear of copying doesn't impede the ability to build 
scalable business models.

There are have always been business models around data publication. My 
claim is that in the era of the Web they simply do not scale at all -- 
if they don't leverage webby URI based attribution and webby entity 
relationship semantics.

> Making the data Linked is a solution; making the data Open might also
> be a solution, but also presents its own challenges.

Non are insurmountable when one applies the architecture of the Web 
accordingly, as per my comments above.

> Sometimes its important to know how the sausage is made, sometimes its not.

You always need to know who made the sausage :-)

> Cheers,
> L.
> --
> Leigh Dodds
> Freelance Technologist
> Open Data, Linked Data Geek
> t: @ldodds
> w:
> e:



Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web:
Personal Weblog:
Twitter/ handle: @kidehen
Google+ Profile:
LinkedIn Profile:

Received on Monday, 10 June 2013 14:07:40 UTC