Building a Shared Web

Joyent is both a customer of GoInstant and GoInstant is a customer of Joyent’s. I’ll avoid all the back patting, but we picked Joyent because we needed rock solid infrastructure and AWS wasn’t cutting it for the sort of compute/serving we do. We’ve been extremely happy there.

So when the folks at Joyent asked if we were interested in contributing a guest post on their blog, we jumped at the chance.

We also took it as an opportunity to start a conversation about the Shared Web.

You can read the entire post over on the Joyent Blog: GoInstant and The Dawn of the Completely Shared Web.

The web is always evolving. As the underlying technology of the web has advanced, we have been able to use the Internet in more creative and interesting ways. We have gone from documents filled with static content to a web of applications, each providing different types of experiences, data and information. We have gone from static shopping cart structures to useful collaborative filtering shopping suggestions to social shopping. We have gone from downloading movies and videos to sharing movies and videos with friends.

In the past year or two, we have entered a new and incredibly exciting phase of the web – The Shared Web.A few years ago, the idea of a 100% shared Internet seemed impossible. The connections were too sluggish and unreliable – in particular on the mobile side. The Web application frameworks did not do very well with situations where extremely low latency was required to keep users happy. But new technology and changes to the core of the Internet are making a lightning fast Shared Web possible.


Real-time and the three pillars of the web

We have arrived a few definitions for the web in the last few years and I believe we are finally getting to a point where there is enough differentiation that we can think practically about the impact of each component of the web as we know it.

Social Web

I have spent a lot of my time thinking about the Social Web and what that means for business and the web in general. It would be hard to argue that Social has not arrived — it is clear that it has. While there is still a lot of maturing to do on the enterprise side, consumers are leading the charge in establishing social as a primary component of the web.

Mobile Web

The Mobile Web has grown at a stunning pace in the last 5 years and continues to break new ground. Much of the excitement around the Mobile Web has been focused on the iPhone, but there are other trends that are having a real impact. Mobile is driving much of the consumerization of IT and mobile apps for the enterprise continue to be a foothold for more creative uses of the web by enterprise IT.

Real-time Web

Until now the concept of the Real-time Web has been relegated to being a euphemism for the rapid delivery of data streams to the end user. “Real-time Search” for example is focused on delivering the same search results more rapidly. Real-time has not as of yet matured to the point where it is widely regarded as a category. In fact, Wikipedia defines the Real-time web as “a set of technologies and practices that enable users to receive information as soon as it is published by its authors, rather than requiring that they or their software check a source periodically for updates.” That day is coming quickly however.

I would argue that the opportunity and function of the Real-time Web is largely misunderstood today, much as the Social Web was when it first began to emerge.

The opportunity for the Real-time Web is to create intimacy and awareness. Where the Social Web gave us a wealth of relationships across a great breadth, and the Mobile Web gave us an immediacy of relationship, the Real-time Web will allow us to connect more closely with those we know with a higher fidelity interaction.

Of course, these are pillars of a single web, and there is a significant amount of overlap between each one, but as we continue to better understand each pillar more clearly in its own right, we are able to see new opportunities that may not have been possible with a singular view of the web.

The advent of Real-time brings with it some very new and compelling opportunities. For a long time now we have mixed “collaboration” in as a use case for the Social Web, but so much collaboration is actually in real-time. That is why we have meeting rooms and why the phone continues to be so useful. So far the software that enables Real-time collaboration has been limited, bulky and poorly executed. It has all been built as a native application (think Screen Sharing, MS Groove, etc). We are however finally at the stage where we can deliver Real-time on the web at scale.

As a sort of Friendster of the Real-time web Google Wave was the sacrificial lamb of the Real-time Web. A beautiful and inspiring first experiment in creating a Real-time experience at scale, it was an incredibly intimate and deep experience as a user. The quality of the interaction was evident but the main problem was that we assumed that all of the use-cases of the Social Web applied in Google Wave. There was no framework by which users could determine what problems to apply Wave to and it became too uncertain of a place. “Should I be here?”, “What am I doing here?” — we didn’t know why we were using the Real-time web.

Perhaps, to be crass, Google Wave’s technology should have been applied to dating. It might have made more sense to an initial audience.

Unlike Wave however, Quora has begun to integrate more and more of a Real-time experience in to their application. As other people are responding to a question, voting up and answer or completing other tasks, Quora tells you in Real-time what is going on and who you are interacting with. It is a very engaging experience and creates a new sense of life within the application. It is a subtle but important use of Real-time.

Moving forward

The Real-time Web has a long way to go but I suspect that the rate of its growth will be rapid. As Mobile and Social experiences have become incredibly pervasive in the last 5 years, Real-time will leverage that critical mass to catch up very quickly.

In the enterprise
Enterprise process is dominated by finite tasks. As social software has been making inroads in to the enterprise we have seen an attempt to utilize socialization of process and software as a way of adding value to those process. The problem however is that they are fundamentally tasks which need to be completed directly.

The Social Web provides an overarching context (“who should be completing the task”) but it does not provide a better way of actually completing that task and moving on to the next one. A combination of Social and Real-time technologies give us an opportunity to leverage what we have developed with social software inside the enterprise.

Real-time collaboration on the web inside the enterprise is currently limited and it is not pervasive. It is limited to single applications (such as Webex, Gotomeeting, etc) which are catch-all solutions based on arcane technology. Future Real-time applications inside the enterprise will be more pervasive and integrated with existing Social and Mobile efforts.

Gaming is the one place where the Real-time Web has had a foothold for a long time. Social and Mobile are only really coming to gaming now, but MMPOGs have been around for over a decade and multi-player gaming continues to be very popular.

Consumer Web
Facebook already has Facebook Chat, but I think you will see them take a page from Quora’s book and more tightly integrate the real-time experience. Rather than having new Wall Posts streamed to you when they are complete, you will see “John M. is currently typing . . .” when you visit a wall and it will give you a better sense of the existence of other users as part of an ecosystem, rather than as simple content-producers.

There are natural limitations to the application of Real-time Shared Web Experiences, but there are many.


The applications are endless, but they are also vertical. Where social can be effectively applied to broad problem sets, Real-time is necessarily narrow but deep. There will be no “Real-time for busineses [linked-in]” or for Dating, but instead there will be applications of Real-time within those and other domains in order to accomplish outcomes that were not previously possible.

So I have to ask Euan, Rob, JP, Susan, Andrew, Ross, Jeremiah, Ray, Dennis and others who have all been instrumental in definining the Social Web for me: What does the Real-time Web mean to you?

E20 vs Social Business?

I thought I would make a helpful slide for the next time an “Enterprise 2.0” vs “Social Business” debate comes up.

Feel free to click to get the full sized detailed view. I made it as pretty as I could.

E20 and Social Business each both need the other. The more that IT can focus on becoming more comptent in rapidly testing, deploying and monitoring E20 applications in the enterprise, the faster and more effectively Social Business will be a reality.

Enterprise Startups survey

I have created a short questionnaire that I am sending out to people who have built or are building enterprise focused software startups. This is not a formal survey, and I am not doing it for commercial purposes. What I want to do is collect the insights, advice and experience of people who have been-there-done-that, or who are in the thick of it right now.

Please take a look and fill it out if it applies to you

Hinchcliffe & Company Is Acquired By The Dachis Group

The news is finally out that Dion Hinchcliffe will be joining the Dachis Group team.

Dion is someone I have wanted to work with for a long time. The depth and volume of insight and quality thinking that he has delivered over the years has been astonishing. Dion has also played a key role in the advancement of Enterprise 2.0 thinking. His blogs are a trove of IT oriented strategy and guidance.

I am excited to help welcome Hinchcliffe & Co to the team and we are all looking forward to bringing their capabilities to bear with our existing and new clients.

The Personal Enterprise

Pete and I have been talking a lot about our Organizational Design work as we have reflected back on 2009 and what that means in terms of Social Business Design as we go forward. For each piece of strategy that we deliver, we always need to be able to back it up with concrete changes at the organizational level for customers who are ready to act. There will be a post from me on the Collaboratory (You can subscribe to the RSS feed here) later this week with some more thoughts on this.

The reason I bring that up is that in looking at better ways that an organization can operate, we often look at the current service delivery model. How are services shared, delivered and managed? Which functions and resources are centralized and which are distributed? Does centralization mean less flexibility? Does distribution mean less reliability? We look at services inside the organization through the lenses of People, Process and Technology.

This is a long way of saying: I’ve been thinking about “shared services” and how that concept is going to change very soon.

Then over the weekend JP wrote about the The Facebookisation of the enterprise.

JP’s dream is more than just the Facebookisation of the enterprise, he is talking about the personalization of the enterprise.

imagine an “enterprise” world where:

  • You chose your own phone
  • You chose your own portable computing device (which may be your phone)
  • You chose your own desktop computing device (which may be your television)
  • You chose the operating systems you put on these devices
  • In other words, the IT department had “lost control of the device”.

I’m not sure whether we need IT departments to “lose control” of the devices used for work, so much as we need them to “give up control” of the data that is available to whatever devices the user chooses. I think that the Personal Enterprise is one where the definition of “shared services” includes thinking about what the user provisions for themselves, not just what is given to them.

The personalization of the enterprise is already happening. It couldn’t be more obvious these days. People are literally carrying two laptops and two cellphones with them. Sit down in any meeting (although I notice this trend far more in the US than in Canada right now) and you can be sure that a handful of the people there will reach in one pocket for their Blackberry, and then they will reach in to another pocket for their iPhone.

How do Police officers coordinate and communicate while working? Ask some of them: More and more it is happening on their personal cellphones. The radios and dispatch systems provided by most police departments are insecure, lack privacy and have centralized dispatching.

So, where do you draw the line? Where is the walled garden and where are the open spaces? I think this is the trick, and where a lot of enterprises will get caught.

The Personal Enterprise (or the Facebookisation of it) is not about picking and choosing which services get opened up and which have controlled delivery, instead it is about opening up as much data as possible and creating an ecosystem that allows personalization to be developed.

We are now coming out of an era where IT delivery was more efficient when it was done as an end-to-end controlled process. I do not believe that has been anything malicious or wrong about how things have been done to date. Controlling the platforms, devices and information that the end user can have access to was simply more efficient: Laptops were expensive, so you wanted to do a volume deal. Blackberries and service plans were very pricy, and then you needed a “Blackberry Enterprise Server” on top of that, and that only really worked with a full on Microsoft Exchange server setup.

Don’t forget how much Databases (still) cost. Doing millions of transactions? You used to have just a few options. For some of your data those are still your options, but you can now deploy MySQL or Postgres at a web, departmental, workgroup and even personal computer level. So you have 4 of the 5 major database types covered right there with free and highly reliable software.

Lets also not forget how “standards” never really used to be standards at all. They were lock-in tools, and the vendors who supported them gave you just enough rope to hang yourself. Having DCOM flashbacks here. Now you can feel free to tell your vendor to bake that RESTful API in from the start, otherwise you are happy to have it built yourself.

Open the data and create incentives for developers and designers to create practical and beautiful things. Users will scream loudly, designers will listen and developers will build. Utopia perhaps, but not far away to be honest.

The thing that is standing in the way of the Personal Enterprise is that we believe that being complicated is a prerequisite for IT. The truth is however that the services that mean the most to end users are some of the simplest and need the least “engineering”. They are rich profiles, SMS (or BBM) messaging, Email, phone and document sharing.

The personalization of the enterprise is happening in communication first, and the sooner that IT departments can get out of the communication business and focus on data, then we will be in a much better place.

Etherpad is now Open Source

This is worth taking note of in the enterprisey world: Etherpad has gone open source and it immediately becomes one of the best Enterprise 2.0 tools available. It is built on a Java/Scala/Mysql stack.

Etherpad is a real-time collaborative editing tool. In short: you can work directly on editing text with other people and be fully aware of what they are doing. In Social Business Design terms it is the ultimate in Hiveminded content creation tools.

Here is a document I created that you can also edit. It is Public.

If you go to the doc you will see just how simple the environment is. No fancy word processing features, just straightforward Rich Text editing.

A lot of energy inside organizations is spent writing documents, sharing notes and creating small pieces of text. Until now, as a public webservice, etherpad hasn’t really been an option for mid-large organizations. Writeboards were an attempt to solve this problem, but they have been a huge pain to use well in my experience.

Why did it go open source? The team behind Etherpad has been hired by Google to start working on Google Wave.

Which brings me to my next point. You can think of Etherpad as a sort of useful and focused version of Google Wave. Where you probably said “Why would I use this?” or “What does this do?” when you tried Google Wave, you will know immediately how Etherpad could be used in your organization.

I am hopeful that Etherpad development will continue now that it is Open Source, but it is hard to tell if that will be the case. Integration in to existing enterprise systems (specifically ECM and Document Management systems) will not be simple. There is a lot of work ahead.

In the meantime, you can try Etherpad at or just play around in the document I created. itself will not be live for much longer.

Interview with Shel Israel on Enterprise Social Media

Shel just published an interview he did with me about using Social Software inside the enterprise.

For all the shit he has taken lately, I have to say that Shel is a great interviewer who did his homework first and certainly made me feel comfortable and in control through the process. None of the usual stuff where the reporter is just fishing for a particular angle.

Go read it if you dare, it was fun to do some reflecting on the past and thinking about the future again.

What should be free and what should you pay for?

Figuring out what’s what in Enterprise software is tough, and it is definitely no easier when it comes to choosing Enterprise 2.0 tools.

So, a question comes up: What should you pay for, what should be cheap, and what should be free?

What should you pay for?

This is a lot more critical than you realize. Having reliable and reasonably priced support is critical to making sure that you don’t disappoint your users.

Vendors used to traditionally “milk” their customers when it comes to support fees. This is changing as more and more minor support issues are handled by user communities. This is an example of Enterprise 2.0 eating its own dogfood. Vendors and consultants are then able to offer much more focused technical support without having to worry as much about the “It’s all in spanish?!?!” calls.

Data Backup and Accessibility
Most people would say that this should be free. I agree: It should be. But it is also something I am more than willing to pay for. Ideally, you can set up an external repository, such as an Amazon S3 account or a private file vault inside your own datacenter. If you are using a SaaS vendor, then they should automatically back up to this external repository in a non-proprietary XML format.

The reason that you should pay for this is so that you can ensure some sort of minimum service level. If something goes wrong, you don’t want to find out that all your backups are a week old.

User Experience
We all like to think that design comes cheap. And sure, you can go on Elance and get a cheap graphic designer who can make something reasonably pretty. The truth is however that because every organization is different, you are going to use software in different ways. You need to be willing to invest in the User Experience if you plan to keep your users happy.

Many pieces of new software look really nice, but when you start using them, the experience degrades quickly. Non-standard interfaces and poor design decisions are the primary culprits.

Future Innovation
This is a hard one to measure, and you can only really base you decision off past performance. Will this vendor or developer keep innovating and updating tools in the future? The reason you need to be willing to pay for this is that they need to put food on the table as well.

What should be cheap?

I believe that hosting and provisioning costs should be broken out and separated from software cost. Current SaaS pricing models obscure the hosting cost and bundle application delivery in with the cost of the actual software. This is a mistake for a few reasons, but the primary one is that highly reliable vendors cannot easily differentiate their service level from a cheaper alternative. Everyone makes the claim “99.99999% uptime”, but it actually means very little.

If single vendors allowed multiple choices for delivery, then the customer could choose what their comfort zone is in terms of cost.

Integration is not a simple matter, but this is where your own IT folks need to play nice. You should expect them to know what the integration points are in your current infrastructure and they should be familiar with how to do that integration. Whether it is a link on a Sharepoint site, or access to an LDAP server, those options should not be difficult to execute on.

What should be free?

Design Patterns
There really is no such thing as Intellectual Property in the Enterprise 2.0 world. The difference between one vendor and another is not that they know more than the other, it is execution and design capability. Some vendors are incompetent when it comes to integrating features, designing workflows and in paying attention to the overall flow and “tightness” of the product. Do not let yourself be sold on “proprietary” anything.

Community Support
Low level support provided by other users and the vendor itself should always be free.

Deployment Speed
There should be no premium to pay to get tools deployed quickly. Vendors should be ready to deploy a base platform immediately which can start adding value to your organization.

What do you think

There is no way I covered everything here. What do you think? What should be free? What should be cheap? What should customers pay big money for?

Annals of Enterprise 2.0: Picking up Drag Queens

My use of the term “Drag Queens” to describe some enterprise software companies was outed today by Vinnie.

What did I mean when I said “old enterprise companies dressing up like a pretty E2.0 babe”

Here is the story of Tim, an Enterprise 2.0 customer who is a little timid, but has decided go for it and party the night away.

Dress up baby, we’re going out tonight

Tim, a friend of Charlie, is out at a club, the lights are low, the music is loud and he has had a few drinks.

The place is packed, and it seems like it is all new faces. Interesting people, great dancers and they are all buying you drinks. There are even a few celebrities hanging around. He dances until 3am, and before the lights go up, Tim ducks out the door.

As he is walking home through the rain, Tim is feeling pretty good. He had a fun night of dancing, good friends and he now has a girl on his arm. She’s pretty and says all the right things.

This story ends with an IT Project Failure, to say the least.

Who is he/she anyway?

The story is much the same for customers in the Enterprise 2.0 world. Those customers who are actually making purchases right now are a little timid and not sure exactly what to expect, and usually it was a friend who took them out to the party.

When the lights go down and the drinks start flowing however, things aren’t as clear as they were before and it isn’t always obvious who you are getting in bed with.

The Drag Queens of Enterprise 2.0 are those old Enterprise software vendors who haven’t done anything to change their products, but instead they went out and have bought a nice dress and have put some eye shadow on their football player physiques.