My Vision for a Real Workable Web3.0 World
Since writing earlier thoughts on the future of the web back in about 2006 (please read first), I've formulated a more consistent, wide-reaching, and - I hope - workable plan for how the web could make the world an even better place to live and work.
At the core of this concept is a "reputation engine": a layer that combines the essential elements of identity and relationship into a common, simple framework that's accessible to all services.
Whereas I thought that Yahoo! was best placed to ride the wave a few years ago, my money's now on Microsoft. For more insights, read on.
I'm just going to splurge my notes, written over several weeks, rather than spend days crafting them into a polished presentation. So I apologise in advance for the fragmentation, but please do read through, as the ideas really come together as the notes evolve!
I hope this will catch your imagination. And if you work for Microsoft, or someone who'll have difficulties if Microsoft does turn on to this vision, please feel free to call or write :-)
Although it’s still very early days, we’re using more and more cloud technology to improve business productivity. I use fewer desktop apps these days. Now, most of my work is done through Google or WordPress online.
This article looks a few years ahead and presents a concept that could help bring around a “Web3.0″ where all applications are related.
Introducing My Reputation Engine
The core of this concept is a Reputation Engine, which records identities, and the links between identities, on an open (i.e. open-source) platform that’s accessible to all and that does as much as possible to ensure an ID is who it says it is.
There are different types of relationship and reputation, so this may need a simple type of reputation ontology.
I may trust ABC to provide news, but not as a custodian of my data, or as a product retailer. So is the type of reputation implicit in what I permit ABC to do??
Q: Can this be distributed and open-source without being open to abuse? (Read on)
If everyone is connected to a central hub (i.e. Identity, in the way that everyone with a GMail account is related to Google), that’s one reliable way to manage certain stuff centrally, e.g. spam. It also means that no one is more than 2 hops away.
To do a FOAF (Friend Of A Friend) poll (for example, “Should I read or bin this email from an unknown source?”), I’d just have to ask our N most trusted relationships, then maybe go down the scale of less trusted links if required (i.e. if null responses), until I’ve got a representative sample.
Who’d Have an Interest?
The whole world stands to gain from a common reputation engine:
Ebay, Ciao, Amazon, Wikipedia, Equifax, Experian, All brands (“how well are we doing?”), Anyone interested in Internet safety (Google, Norton, law enforcement), All lenders, Publishers, Advertisers, and people being advertised to!
Q: Can anyone (any ID) poll their network for another ID’s trustworthiness?
- I poll “Who knows Bob100?” (who has just sent me an email).
- If I get an “OK” back from Jeff77 and I find out the message is spam, and click the spam button, my trust in Jeff77 should take a hit.
- So all IDs are responsible for their own assessment and recommendations. So we all need to be careful to represent our own relationships honestly. If we can’t really trust someone, we shouldn’t say you can, as that may impact your trust in us, which may impact others’ trust in us!
- I can see people guarding and managing their own reputations as carefully as ebay traders do today.
- Is Nora123 good for a loan?
- Is Fred88 trusted to babysit kids??
Maybe various IDs can have special authority on certain matters (such as childcare).
Maybe Polling isn’t Actually Necessary
For an email even to get through to me, maybe it would first need to have a connection to me from the originator’s ID.
For me to ask about a babysitter, flatmate, car share, product supplier, movie review, healer, blogger, priest… I first have to have some channel to them, through a relevant ID, which can then provide the context (ontology) of the trust/link?
In the case of spam/phishing, this might be via the central ID (e.g. data from Norton), so email and web could be covered this way.
Any ID May Publish Services
Services are trusted, as are the ID. In fact – any service should be an identity in its own right, as is any product!
- IDPlus can be trusted.
- IDPlus’s spamcheck service can be trusted.
- TechCrunch’s news feed
- National Paedo Watch Service
- Reverend X’s Healing Pyramid
- Herbalife’s Member Programme
- PartyPoker’s poker room…
Any individual or organisation (company, group, political party, charity etc.) has its own ID, just as in real life.
And any output of any individual or group (product, service, media, story etc.) also has an identity of its own, to which other identities can relate.
So TechCrunch’s news feed can rate my blog post, as identity-to-identity.
Does a Service Need to be Keyed, or Permissions Fixed, to Prevent it Going Bad?
Absolutely! I think that permission may be the key. 2-way trust is implicit.
If I trust PartyPoker or ebay to pay me money, and they trust me to pay them, there’s a 2-way relationship.
So it’s possible for people to poll the world for a quick answer on how trustworthy a particular poker site or trader is!
Some Examples of Trusting Channels with Data/Services
- My shopping > Shopping Channel
- My personal profile > Friends & Family (F&F), Facebook etc.
- My CV > Monster Job Board, and people trusted by them
- My Calendar > F&F, work colleagues
- My Phone number > F&F, work, clients, prospects
- My Current Location > F&F, work
- My Bank Balance > My bank
- My Income > Shopping Channel, Credit Reference Service
- My Debts > My bank, Family
- My Photos & Videos > Friends & Family, Flickr, Facebook, YouTube
- My Documents > Work colleagues, various others I nominate
- My Wallet!! > Shopping Channel, Party Poker, ebay
- My Name > Not everyone… Will depend on trust
In many of these instances, trust could be implicitly extended to highly trusted IDs related to the IDs I trust. The onus is on those direct-contact IDs to be diligent in their own relationships, to avoid spoiling my trust in them.
At its Core, What Does this System Need?
- A centralised system for recording an ID, and information related to that ID in the cloud (similar to OpenID today)
- The ability for IDs to handshake with other IDs, and establish with a high degree of confidence that they’re both who they were last time.
- Record of 1-way (and maybe 2-way) trust / rating between IDs.
- A way of categorising that trust (tagging?)
- A way of identifying an ID as a product, service, media asset, or software client (e.g. I trust my iPhone to do x, y, z…)
- A common language for permissions, data requests, data transfer between IDs.
Any service can be an ID, linked to a provider by a particular type of trust link (so there may be a “provides” type of link in the ontology).
This should be open-source, and open-platform, extensible by anyone. Rigidity will be a millstone (as shown by the relationship between browsers and standards).
Should also be retro-fittable to existing offerings (like OpenID today).
It’s commercial advantage that will drive adoption. If someone is smart, they can set up early killer applications that encourage others to plug in.
Can People’s Data Be Held Anywhere?
It would be great if I say that any app can access my info, but also that I can update it from any interface. So, if I add a new delivery address on Amazon, it’s also available on ebay.
And if I put a photo on Flickr, I can also sell it on iStockPhoto, or use it as my profile picture on Facebook.
So, my ID doesn’t need to be a big database; just a record of uniqueID, my relationships, and their types.
In theory, any existing service provider could become an ID repository. PayPal, LinkedIn, Google, Yahoo!, MSN etc. would be ideal for day 1. I’m not sure how much of this OpenSocial and other similar projects are capable of now.
2 Links is Enough
I’ll probably have 100+ links to other people (some stronger and fresher than others). But I’ll also have hundreds with companies, organisations, products, and services. Even so, this will still be easier to manage than all my logins and spam today.
Trust referral/polling should always be as transparent as possible. Certainly in communications apps, like email/Skype.
Distributed Trust is Key
I won’t ask you how long your account’s existed, and how well connected you are. I’ll ask everyone else! It’s very hard to spoof.
Describing Links Neurally
Should links have a weight as well as a value? What about shelf life?
Today, I may highly rate a brand of coffee, a politician, a company, news story, car, ebay seller etc., but these perceptions can change quickly.
My trust in my kids, my wife, best friend, business partner, political party, religious organisation etc. can last longer and mean more to me.
Maybe relationships that are regularly used & maintained should be reinforced, just like a neural network! Those that aren’t decay and dissolve over time.
- A news story could be gone in a few weeks.
- But my trust in the provider of the news story could grow over time, as long as they keep generating good content, reinforced by all the 2nd-links from the stories.
- A business partner I strart trusting today shouldn’t be as important as someone I’ve worked with and trusted for 5 years.
A key question: How will changes in trust in 2nd-degree contacts affect the common link?
- i.e. If you recommend Jack to me, and I decide I don’t rate Jack, does that affect my trust in you?
A natural, neural network of trust relationships between all identities and objects that exist online, both now and in the future.
An ID owner may manage their own level of trust and exposure.
- A subject (me, our company, organisation, blog)
- An object (you, him, her, it, product)
- A value/strength (how strongly subject relates to/trusts object)
- Types/tags (that describe the nature of the relationship/trust, which may be extended to implicit permissions)
A lot of this information already exists, but the fundamental disconnect prohibits massive potential.
Great opportunities become possible when this foundation exists. Also, the early bird will get many worms.
More Strategy: The How!
My goal: A process for releasing something so GREAT that everyone HAS to get on board!
Killer Apps are the Key
Release one or more killer apps simultaneously with a full specification for extending your apps and systems to use this.
The killer apps could be:
- Spam filter
- Marketplace (a place where anyone can offer anything, or post a request for anything)
- Permission Marketing (a biggie!)
It would be way easier to jump on the back of an existing database (esp. Microsoft, Yahoo!, Google). There are also players like AOL who may be motivated to re-emerge as leaders. Netscape? Is there anyone still breathing that hasn’t been bought, and would benefit from riding the public service wave?
Profile of Ideal Developer
- Pain / Threat
- Brand benefits (open, making the world a better place)
- Capability & Cash
- Existing channels
- Existing database of members
- Advantage / position to defend
- Threat of a competitor’s having this first
- Need to enter markets:
- Social Networking
- Permission Marketing
Threats for Microsoft
- Low trust / brand value
- Need to build new revenue streams:
- Death of the desktop (Windows Vista is not good enough to save the cash cow); OSX & Linux on rise
- OpenOffice and online docs like Google Docs a threat to the MS Office cash cow.
- Google has a few strengths:
- AdWords (!!!!)
Opportunities for Microsoft
- I believe Permission Marketing could trounce today’s online advertising (and undermine Google’s Golden Goose).
- Trusted email and instant messaging
- Be the original Web3.0 brand
- Transparency & philanthropy
- New MSN, more than IM, a complete lifestyle online
- Own the vocabulary (mindshare) of the new web.
- Marketplace could compete strongly against ebay, craigslist, and Google Local.
- Out-Google Google basically.
- Strengthen media: photos, news, media, blogs…
- No antritrust, because it’s fully open. MS would just have the day-one advantage. It’s in everyone’s interest to have this totally open.
- Embracing & empowering competition.
Permission Marketing is Huge: How it’ll work
I say what I’m in the market for, now and in the forseeable future.
I also say what my tastes are (and the web can learn from watching what I do)
The net may also know about what my buying habits are. Do I book holidays early or late? Do I buy new or used cars? How long do I keep them? Do I always buy the same brand of aftershave, or try new ones? Do I like to buy flight, hotel & car hire in one purchase, or book separately?
It can also learn, by cross-profiling, what people like me do and buy, based on my tastes and history.
Marketing spend goes into better products and services, and discounts, instead of bombarding information to people who don’t want to know.
Instead of being interrupted by ads, I choose to view offers that have been selected for me.
I could even join a (MS) AdFree channel, which strips out all ads from my web browsing experience, and at the same time watches what I do for the purpose of profiling. This is all under my control.
Channels and web sites may register to be part of the AdFree channel, but they must comply to build trust.
MS runs the PermissionMarketing ID, which polices the system and sells introductions to vendors.
Data never goes to vendors; MSPM is the intermediary. All offers must go through them for vetting and must match the required level of trust.
Vendors pay to send product information, or offers, to N thousands of open and willing consumers.
Takes the “con” out of “consumerism”
So Audi could reinvest its ad budget into directly contracting willing buyers, R&D, and making even better cars!
The price point could be set by AdWords-style auction. “I have 10,000 people in socio-economic group A-B1 who express interest in buying a second home”
Win win win win win win win !??
MS News Finder: How it’ll work
I subscribe to news by channel, keywords etc. at first. Then I rate stories for interest and relevance (either actively, by clicking something, or passively, by reading them through).
MSNF learns what I like, what other people like me who like what I like like, and predicts what I’m likely to like out of tomorrow’s news, across all channels, video, etc.
MS benefits because it informs taste and preferences! It’s not about eyeballs any more: it’s about trust.
This service builds trust in MSNF (in the same way we all started trusting Google because it gave us better search).
It’s not advertising dollar-driven any more, because the ads have gone.
So where are the dollars?
There are plenty of opportunities.
- Maybe people could pay to push a story network-wide, and get the results (instantly) of how well the story is received. Good idea for politics & marketing; all policymakers.
- The data on “what people are interested in today” could be worth money.
- Ties in well to MS Permission Marketing. In return for getting all your content ad-free, you permit your browsing habits to be harvested, which generates PM offers. As long as you play the PM game (i.e. actually look at promotions, and spend), you’re a valuable customer.
MS Marketplace: How it’ll work
Anyone can get a temporary or permanent presence to promote their product, service, offering
There’s a fundamental distinction between supply (I offer) and demand (I need).
Anyone can also post a need: job, firewood, girlfriend, housesitter, dog walker, CEO, mentor, running buddy, kids’ bike, yacht…)
Marketplace takes into account:
- Tastes & preferences
- Socio-economic profile
- Geography, timezone etc.
…and simply matches supply with demand (something search engines don’t do).
Bringing it All Together
So we’ve got an open, neural network of IDs, trust, and services.
Links and data are all controlled and owned by the ID to which they belong (you). You decide who knows what about you.
The strategy is to release this to the world, together with a few commercial killer apps that will drive early adoption and rapid growth
Now – a relationship between 2 IDs could be a service, or indicate permission to read/write/store/gather my data.
It’s a 2-way consensual system of transparent trust and reputation that fundamentally changes marketing and consumerism.
To participate, all you have to do is link an account to a GUID (globally-unique ID), and agree to share data in common formats, and agree that ID owns all its stuff.
What does it all Mean?
All ratings/trust are now available, like rolling out all the ebay & Amazon ratings across all online services and other media not yet invented!
What you can do with this will depend more on your aggregate trust, and less on noise.
- So bloggers with creativity and integrity will get more exposure than commercially-driven partisan news channels.
- A company that provides a service will only succeed by delivering on its brand promises.
The consumer is fully in charge. “I own my everything!”
Back to Strategy – Grand Unifying Theory
We’ve got 3 killer apps: Content(news), Permission Marketing, and Marketplace. These 3 can work together brilliantly!
PM generates $$$ through outperforming AdWords. (Who wants to be advertised at any more, when you can choose from real offers matched to what you really need now?)
$$ are ploughed back in to the content providers who create the great content and gather usage data.
(Remember, my data are always owned by my ID. I can view/filter/make it anonymous as I wish. That’s the deal.)
Any content provider or service provider may generate usage data (with permission), which is then crunched and used to match more content, products and services that suit my tastes and needs.
Data can include:
- My calendar
- Where I am right now
- My tastes
- Do I preorder or buy on impulse?
- Work history
- Music I listen to
- My age
- My birthday
- What I buy, what I say I’m looking for (Marketplace)
- What I promote, offer & sell (Marketplace)
- Gadgets and software I’m interested in or use
- Movies and books I’ve seen or read
- TV listings
- The weather
- All the content I look at
- Even the content of my email
All this could generate useful intelligence, which is turned into helpful, personalised marketing messages.
Our Permission Marketing system pays content and service providers for their raw data that matches my ID.
How Everyone Wins
The Consumer Wins
In return for my participation, I get no ads, I get my pick of the best offers, and content that’s personally selected for me (without having to hunt for it any more).
I get to find out about stuff I didn’t even imagine existed, but just happens to suit me, just because someone like me finds they like it.
I can also discern the most reputable, reliable providers that most closely match what I need right now, without having to ask everyone personally or go through trial & error.
The more marketing I participate in (i.e. view, click through, or buy), the more valuable I become. Hey – I could even get cash back directly! So as a consumer, I also get to build my reputation and value by paying promptly, giving feedback, playing by the rules, and spending within my limits, and being disciplined about charging back or returning stuff.
Promoters get to save money by delivering their messages only to people who want them, so they can sell more stuff with less spend.
Plus, they get much more usable feedback from the system on how well their messages are received (“People with beards who own red Volkswagens clicked through most…”)
Content Producers Win
Everyone who writes a blog, takes photos, or makes a funny video on YouTube, can start earning real money just for doing what they love doing.
The playing field is also suddenly levelled, because the power is no longer concentrated in the hands of individuals who own the channels, but in the hands of the people who enjoy the content (like YouTube has shown).
And wait! Suddenly, have we also solved the DRM problem? People who write great music that gets matched to people’s preferences can get compensated fairly through micropayments when their output positively matches people’s tastes today.
The System Owner Wins
As already discussed.
Well, anyone who currently earns a living by throwing out loud advertising messages, and cluttering up our minds with noise.
Oh, and of course people who earn a living by doing business in an underhand way, delivering less than they promise.
And don’t forget the guys who get fat because they own media channels and get to dictate who makes it big.
And spammers. And phishers.
So let’s not worry about them right now.