Linking Knowledge in a networked society
I recently attended the annual ODI summit at the British Film Institute in London. There were some 700 people talking about how their organisations have embraced the open data concept and were figuring out new ways to solve problems. We saw many examples of what leading individuals and organisations are doing. From the Internet of things (IoT) openly publishing noise and air quality from Heathrow Airport, to charities using data to solve problems and (social) entrepreneurs and big businesses alike getting their heads around the opportunity and moving forward. Government too (MP Matt Hancock) leading the charge on the importance of sharing data responsibly to enable a more empowered and democratic digital society.
I was wearing two hats. EthosVO and Linked Data specialist Seme4. Ethos has recently signed a strategic partnership with Seme4 and I have joined their board. One of my first tasks for Seme4 hat has been to listen, ask questions and and co-create a sensible strategy for moving forward. I’m ready to start exploring that next piece….
A couple of things struck me from this week’s ODI Summit. Firstly, the appreciation of the importance of a common culture and ecosystem approach in getting open data and collaboration to work. Second, just how many organisations and individuals are experimenting in this (new and as yet unmastered) area. Each year the number of people involved seems to (at least) double (Moore’s law). If what holds true for transistors also holds true for people then in less than 15 years 60M people in the UK alone (i.e pretty much everyone!) will be thinking this way and just 12 months before that point only 30M and then 15M and so on. Such is the power of networks… Where next? What exactly is the Seme4 technology stack? What are the USP’s? Where is the LD market now and where is it heading. How best to move forward?
I woke this morning with yet another epiphany on how the pieces of the puzzle fit together. If ever there was a danger in experiencing “epiphany fatigue syndrome” I’m first in line. But it shows no signs (yet) of doing me any damage.
For me, over the past 15 years and more it has made sense to share, to collaborate and such like. My knowledge combined with your knowledge must be worth more than our knowledge independently. The value of the first fax machine was rather limited. Give everyone a fax machine and it becomes an essential utility or infrastructure.
Making sense of this value in a networked society is a rather different exercise than in an industrial one. When the world was about manufacturing and efficiency we used a Taylorist approach. Look at the wiring diagram and we can project the value of each (single) enterprise by aggregating the functions (the value) up to the top. Working out which siloes to kill and which to promote in some neo-Darwinist kind of survival of the fittest game could be ‘objective’ (according at least to some economists, shareholders and CEO’s) and informed by learned management accountants and economists.
The network changes all of this. Open data enables outside-in thinking for freely open data (there is already lots!). Linked data has the potential to change this to include network views of “valuable data” between silos.
So far the majority of value has been derived by either getting the network to add value to free data (published mostly by Government) or from running projects within organisations to make one’s own proprietary data more valuable by trying to make sense of it. Over the years this has been called data-warehousing, big-data, natural language processing and Knowledge Management (e.g. Autonomy was acquired by HP for $10.3bn in 2011).
But unless our culture fundamentally changes we are still using our old “operating system”. Working from our ‘siloes’ outwards. We are in danger of “solving the wrong problem really well”. By this, I mean solving the efficiency problem not the effectiveness problem. A phrase I coined years ago was “running against the same brick wall even more quickly in attempt to remove the obstacle”.
To solve the effectiveness problem we have to look at things the other way round. Outside-in. Not inside out. From the network to the institutions. Not from institutions out to a mass market.
Looking from the networks view of things, 99% of the most valuable data lies hidden away in the depths of institutions behind (sometimes) well protected firewalls.
People are most valuable because our neuro-associative memories are actually rather good at “joining the dots”. But we can only do that with the information that we actually remember. And we forget much more than we remember. Sometimes (well, me certainly) having to be told something perhaps 3 or 4 times before we can commit the fact to memory.
What if computers could help us navigate this knowledge graph?
At Ethos we’ve been building collaborations around complex problems such as parking, smart cities, skills planning, and health. They all start with getting the culture right to collaborate, they all provide a way to use technology to enable the collaboration (of course the data is key here) and finally they all define multi-sided business models that enable sustainable value exchange between individuals, institutions and various organisations. Trust and moderation are key to the whole thing of course.
But as Volker Buscher of Arup said (“we are dogfooding” aka “eating our own dog food” or as Ethos prefers to say, “drinking our own champagne ”) we have developed our practice internally and externally and have built our own platform (The Ethos Platform) using linked-data to help us understand our own knowledge graph. If one were to look at Ethos from the outside, the extent to which our knowledge graph is shared today is rather limited. Here’s a snapshot of some of it….
Fig 1: Part of the Ethos Dashboard showing some of the things happening right now : Who is working more closely with who. Focus on Rob Pye : Where Ethos Partners are right now…
During the ODI Annual Summit, we saw the Linked Data world from the inside of Thomson Reuters and it all looked rather familiar!
However, both of us are guilty of an inside-out approach. It is still “my knowledge” or “your knowledge”. I could even “buy” Ethos knowledge or Reuters knowledge but I would have to go back to the old world of brochures, pdf’s, salespeople and web pages… Sales. Marketing.
We know this works for commodities. But it seems wrong for knowledge. Knowledge-workers are not really substitutable. Many of us are participating in a “knowledge society” (Jeremy Rifkin coined the phrase near zero marginal cost society) where products and services have been commoditised and all that is left is knowledge and creativity. As Seth Godin says we all “have to make our own art”. This is the pressure for most organisations. It is real, immediate and highly threatening. The manufacturing is increasingly getting done by robots.
Funnily enough, the three entities in blue at the centre of the Thomson Reuters graph (See Fig 2: Person, Company and Product) are these are the three that sit at the heart of our own meta-model. Only we call them something different (people, organisations and problem).
Corelating these two concepts that are in fact fundamentally the same is in fact quite an easy problem that we have solved already with the sameas.org service. We link together URI endpoints so that my Sony plc and your Sony plc can mean the same thing in the language of Linked Data.
Fig 2: The Thomson Reuters Knowledge Graph
IF one were able to look at the World’s knowledge graph from the outside perspective not the inside perspective, humans (and by extension organisations) could truly leverage their potential.It’s not the free and totally open data where most of the value lies. Its mashing this up with all the juicy stuff that’s inside organisations.
This is not a new concept per-se (Tim Berners-Lee’s the Semantic Web) but what is (relatively) new is the emerging culture that realises this imperative. Unfortunately, it’s easier to see the criticality of this mindset shift from the outside of institutions not from within them.
We need to solve this data integration problem outside-in. Not inside out.
Linked data provides a way to describe data that works but it’s quite complex. This is why Seme4 has been developing a data-integration platform (and associated services) that provide a data integration approach that works outside-in as well as inside-out.
A Worked Example
An example might help to illustrate the potential value of this approach. This is a future scenario not possible today but perhaps not so far away…
Healthcare-Megacorp (MegaCorp) thinks there may well be a shareholder enhancing proposition if it purchases and integrates One-Man-Pharma Limited (OneMan) with its core business. OneMan has some Phd’s and Patents and some clients and history of when they fell out with half their original founders who left to start Two-Man-Pharma Ltd and so on and so on. It’s a complex world we live in.
MegaCorp management, their shareholders, staff and competitors would rather like to form their own opinions on this hypothesis too. It’s not a secret that they are thinking of an acquisition and in any case an offer would have to go through a public bid process. Everyone knows the dubious track-record of mergers in value creating shareholder value.
MegaGorp initially and some other interested parties initiate some independent searches online and via brokers asking the question “what the likely value?”. One search respondent is McNewBie Strategy Partners (Newbie) who have the world’s best M&A insights database (linked data) and people. Newbie benefits from an impartial web service that aggregates what stories, experience, parallel etc are known on MegaCorp and OneMan. It aggregates third parties, people, case studies and much more. All around the specific scenario asked. It doesn’t disclose all this information to MegaCorp, Newbie or the others as some of the disclosures come with rules such as here is the evidence that I have completed an independent valuation of Newbie 6 months ago and you can buy this service for £20k but I do not want to disclose my company name or the individuals who would do the work at this point. The search agent is secure and the aggregation information can be destroyed at the end of the transaction/search. It has all been implemented in Blockchain and bespoke rules regarding the life of the data can be programmed for volatility. Rather like Snapchat has been designed for “read then delete” messaging.
Fig 3: Sketches to Inside out thinking (left frame) Verses outside-in thinking (right frame). In a Networked approach a number of queries are aggregate knowledge from all sources on behalf of different stakeholders (Yellow middle). Outputs (RHS) are then aggregated from the inputs to show different contracting possibilities drawing on information from the whole network.
Each endpoint decides what to share (dynamically) and what (if any) atomised pricing would be required to release this information, service or knowledge to Megacorp should they be commissioned.
Newbie then automatically produces a quotation on the best value-add combination of prime and subcontract services to satisfy Megacorps requirements. Of course Megacorp request instant quotes from a dozen of the best M&A consultants who produce very different bids and prices.
In commissioning Newbie, Megcorp can see not only the value that they will bring from their experts and specific knowledge from previous work but also what McNewbie will bring to the table from the world’s knowledge graph and of course that would include other value added services that are brought to the table.
The value proposition is truly outside-in. What can the network bring to the problem? who has exposed their knowledge graphs and prices sufficiently openly and dynamically to this particular problem to produce the best possible value proposition for this particular context. The figures below show a rough sketch of this “inside-out” old system of thinking versus an “outside-in” networked thinking approach.
A networked society cannot achieve its full potential through data (or organisational) centralisation. The real potential applies in linking valuable knowledge across institutions. This is about culture first. The first mover advantages are significant. There are new (multi-sided) business models to uncover and exploit.