Welcome to Generation 3: VEPs are the next CMSes

A Grand Evolutionary Theory of All the Software Behind Digital Life, From <img> Tags to CryptoKitties

The evolution of online publishing software since 1991 gives us a road map for the next two decades of digital life. Out of the raw muck of early-nineties HTML came high-end Generation 1 CMSes for publishers and corporations. Then in the early 2000s, stripped down CMSes fostered blogs, a mass of single cell organisms that quickly speciated into the kingdom of social media platforms, Generation 2 CMSes that today occupies hundreds of content niches and overshadow their print-based ancestors.

Virtual Experience Platforms (VEPs) are the next stage of this progression. As virtual and augmented reality experiences become more common, organizations will begin to adopt VEPs to persuade, cultivate, entertain and serve customers and employees. Some organizations will use VEPs to supplement their content and social media strategies. Some VEPs will create entire new marketplaces, just as Wikipedia, Youtube and Facebook exponentially exceeded the ecosystems they entered. Whether you’re a programmer, marketer, rabbi, coach, lawyer, educator, therapist or short-story writer, you should start steering your career with this future in mind. 

Web oldtimers can read what’s below as a continuation of my 2002 essay forecasting the ascent of social media and resulting eviseration of traditional publishers. Fans of Racery can just focus on the Generation 3 section. 


>>Generation 1

Each year between 1995 and 2005, web pros began emitting a collective groan sometime in February.

“Uh oh. Damn. We’ve still got last year in the footer of every page on the site. We’ll look like idiots if we don’t update those.”

Businesses, charities, universities, political campaigns, newspapers, governement agencies and pro teams all suffered the same pain. As they graduated from websites with ten pieces of content to 100 to 1,000, their original method for pushing content to “The Internet” — getting The Webmaster to hand-build each Web Page in raw HTML and then manually code URLs — was not scaling.

In the short term, the solution was to update the copyright year on the front page and hope nobody, at least nobody in management, noticed last year’s date on other pages.

Eventually, many websites had copyright dates that were 2 or 3 or even 10 years out of date. The solution became obvious: put everything into a database. Store all content components — articles, bylines, mastheads, tabs, images, web design, headlines and yes, footers — separately so that each category could, if necessary, be managed in bulk. Not only were these systems easier to manage (once everyone in the organization accepted their inevitability and stopped insisting their department had special needs that couldn’t be met by the ‘plain vanilla’ CMS management wanted to buy) but they created a stylistic consistency across a site that made content far easier to consume.

Thus, the Content Management System became The Thing. Every organization that was serious about the Internet wanted a “CMS.” Smart managers RFPed; egotists hired teams to build their own.

At the high end, a slew of tools emerged from traditional newspaper publishing systems to become the standard for periodical publishers who wanted to move online. Elon Musk made his first millions building Zip2, a CMS for US newspaper city guides, and selling it to Compaq. In the UK, Triniity Mirror spent 100 million pounds to build a CMS for just three newspapers. Racery’s parent company, Pressflex, was (and is) also a player in this space, renting websites to newspaper and magazine publishers in France and UK. At the opposite end of the publishing food chain, tools like Userland, Blogger, Moveable Type, and WordPress bubbled up to serve the needs of individual writers and, eventually, larger organizations and groups that weren’t traditional publishers. Open source projects like Source Fabric and Drupal aimed to fill in the middle ground, catering to non-traditional publishers who still wanted to empower multiple authors/editors.


>>Generation 2

The period from 2005 to 2015 marked the ascent of second-generation CMSes. Blogs had empowered individuals to publish article-like ‘posts’ — now individuals wanted to publish other types of content. Platforms expanded quickly to serve new niches: Wikipedia (crowdsourced, attribution-rich articles), Craigslist (classifieds), LinkedIn (resumes), 4Chan (memes, gifs and random musings), Facebook (personal information), Reddit (links and comments), Goodreads (book lists and reviews), Github (software code itself), Twitter (140-character musings), Youtube (video), UranDictionary (crowd-source definitions), Pinterest (curated images), Instagram (doctored images), Snapchat (shortlived updates), Giphy (animated GIFs), Behance (designs), Twitch (videos of game-play), Strava (cycling routes/data), Udemy (online courses), Slideshare (presentations), and Thingiverse (designs for 3d printables.)

An additional layer of functionality enabled viewers/readers to subscribe to particular authors, and let authors limit who could subscribe. This layer transformed most of these CMSes into what they’re most commonly known as today: social networks.

(Looking at an 80%-baked version of this post, a web oldtimer pointed out that most people would not see a CMS in the core DNA of Twitter or Facebook. He wrote me: ‘Twitter grew out of the messaging space, and Facebook was created for social connection.’ In fact, hindsight notwithstanding, Twitter’s roots were almost entirely in CMSes (Ev Williams and team had previously built Blogger and Odio, with zero SMS experience) and Facebook was based on the template of a book printed annually by Harvard College — a user-populated, access-controlled CMS for student profiles. Its prime use, said Zukerberg at the time, was that “people can search for other students in their classes….”)

>>Generation 2.5

Since roughly 2010, we’ve witnessed the advent of CMS generation 2.5 — the social media management platform. This CMS is essentially a layer that exists to deal with the complexity and volume of content flowing to/through Gen 2 services. Tools like Hootesuite, Tweetdeck, SproutSocial and Buffer give individuals and organizations the ability to manage publishing and reading at scale across multiple social platforms. (More of these services are listed here.)

Today, any organization of ten or more people invests time (if not money) in a CMS. No self-respecting cabal or taco stand tries to compete online without a strategy for content and social media, even if no strategy is explicitly articulated. (My favorite restaurant in Chapel Hill, Italian Pizzeria 3, has no web page to speak of, but does a great job on Facebook, in part because its exuberant owners never copy-edit their posts. Most posts end with the salutation “IP3 — place to be!”)

So what’s ahead? CMSes for virtual worlds.

>>Generation 3

Virtual worlds as we know them today are essentially platforms created and monatized by game-focused professional organizations. They’ve been around a while, ranging from Second Life to World of Warcraft to Runescape to Minecraft to Strava. In the publishing model I’ve sketched above, they’re like the newspapers and magazines of yore, professional organizations each with a core mission of creating a particular type of content/experience for a discrete category of consumers/users/actors.

If my model holds, in coming years we’ll see the emergence of services allowing other organizations (or eventually even individuals) to create their own virtual worlds, in essence to act as “experience publishers.” (Dare I say: experiential bloggers?) Though vital in spirit and reach, the “published” virtual experience will not be part of an organization’s core competence — just as @ford is not one of Ford’s core products, only 1% of the 2005 “art of the heist” alternate reality game was about Audi A3s, and the Westworld storefront at this year’s Comicons didn’t “play” on HBO. These new world-sponsors will use their creations to engage with customers, stakeholders and eventually even friends. The content “published” will not be static, but ever-changing. Users will invest energy and ideas in personas within the new world. The content will be competitions, catalogs, puzzles, games and simulations. But mostly, they’ll be categories of life-altering experience we can’t yet imagine — just as the concept (and power) of Instagramming your food would have been unimaginable for a SLR owner in 1995.

The first iterations of Gen 3 CMSes are custom built for their sponsors, just like the early CMSes built by newspapers like the New York Times. They’re super expensive and take years to conceive, like that little game that Google spin-off Niantic built for Pokemon. (Next up, Niantic is building a Harry Potter virtual world.) At a smaller scale, celebrities are dabbling in a form of augmented reality by launching games based on their personalities, interests and even real-time travel plans. Kim Kardashian’s hugely popular Hollywood virtual game app mirrors ongoing events in her life, including vacation destinations. Also from LA, Ayesha Ayesha Curry has a new virtual chef game that mirrors her cooking show. 

If history is our guide, two seemly opposite trends will converge as virtual experience technology matures. First, previously nonexistent niches will become articulated, just as Pinterest and Strava emerged, unprecedented, from a digital cloud. On the other hand, we’ll see standardization within those niches as users develop consistent expectations and idioms emerge.

Eventually, as with the evolution of article-focused CMSes, the price of serving customers/users prices will drop and cloud based software services will emerge to allow businesses to ‘publish’ virtual experiences tailored to their users’ needs, desires, and interests. For now, let’s call these nascent fourth generation CMSes “Virtual Experience Platforms,” VEPs.

We can already see hints of future VEPs in various building blocks and boundary line stakes. These come in two forms.

  1. Virtual objects and spaces. Second Life is funnelling profits into Sansar, a “metaverse engine” where individuals can build virtual objects that others can play with or incorporate in their own scenes. AltspaceVR hosts a variety of spaces for particular activites like standup in your friend’s living room powered by VR headsets. (See AVR’s virtual reality ‘help desk’ video for a great example of how a proto VEP might be deployed by a business.) Philip Rosedale, creator of Second Life, is focusing on democratizing access to VR components via his company High Fidelity.  Companies like EntryPointVR aim to allow drag-and-drop creation of interactive spaces. Closer to hand, every iPhone X owner now has the ability to animate a poop emojii with her voice and facial movements.
  2. Codes, games, idioms and narratives. IFTTT is effectively a CMS to manage and share the algorithyms people create to interact with the web’s many APIs and algorithms. Services like BotPress effectively function as a CMS for a business’s chatbot. (The logical extension will be a similar service helping individuals manage their own chatbot ‘publishing‘ to family and friends?) SuperBetter is building and testing challenges to boost users’ resilience.

Someday soon, businesses will have dedicated ‘VR’ teams to inform and amuse customers, in the same way they today have content and social media teams. They’ll adopt one or another VEP to serve their customers, just as they’ve adopted WordPress (Gen 1), built a Facebook strategy (Gen 2) and used Hootsuite (Gen 2.5) to manage multiple social media communication and interactions for their products.

And this is where Racery fits in as we strive to perfect one type of idiom. Racery’s DIY route-builder allows charities and businesses to customize virtual routes that their stakeholders can use to compete on and in for fitness and/or fundraising. (The Hogwarts Running Club just used our platform to host a two week virtual race that raised $20k for six charities; the North Carolina Department of Health and Human Services is using the same platform to take 2,000 3,000 people a virtual marathon and ultra in Paris.)

It’s obviously still too early to predict what core features and characteristics will distinguish VEPs. But based on patterns in the evolution of electronic publishing since 1991, these arcs seem likely:

  • VEPs will allow the participant or sponsor to limit participation to a select group. The sense of privacy increases people’s willingess to share; in turn, intimacy and specialness boost participants’ sense of privilege. (Access control was arguably what turned CMSes into social platforms.)
  • Early VEPs will do less rather than more — note that Twitter and Instagram and Pinterest all succeeded by severely restricting the format of that users could publish. This simplified both the creation and consumption of content.
  • Corollary: Like all truly disruptive technology (in Clayton Christensen’s sense), VEPs will win with underserved customers who are willing to “satisfice,” long before the VEPs mature and displace incumbents in existing, high value niches. (Again, think about Facebook serving college kids a decade before its products and appeal matured to overshadow that of the entire US print publishing market.)
  • VEPs will combine fun and self-improvement — just as people use Twitter to see and be seen on the cutting edge of news, and people use Instagram to prove and improve their tastes, customers will participate in VEPs for a mix of pleasure and edification.
  • VEPs will leverage pre-mass-adoption tech like AI, wearables, bots, face recognition, Oculus Rift, blockchain and Alexa. To get a sense of why this piece is important, think about how platforms like Facebook and Twitter blew up because they optimized for the format of the fastest-growing new consumer technology — mobile phones. (Think DIY platforms for a whole kingdom of CryptoCreatures.)

The VEPs may be different from earlier CMSes in one important way. The first three generations of CMSes are tools for communicating (and archiving) different types of information for and by different categories of people. While these systems escaped the Gutenburgian categories of paragraphs, articles, editions and books, they’re still fundamentally locked in the paradigm of publishing; they’re platforms that are focused on push.

Yes, readers can ‘interact’ with authors, responding to content with likes, claps and comments, but all three generations of CMS platforms are still built around binary roles: actors and audiences. Which is to say Generations 1-3 are about publishing in the same sense it was used in 50 and 500 years ago. Perhaps VEPs will let people both interact as autonous equals in virtual worlds AND collaborate on creating virtual worlds.


1) An equally valid, distopian plotting of the same datapoints:

2) To better understand Gen 2 CMSes that a grew to be social networks, it’s worth mentioning a few innovative services that flatlined, perhaps because they were too niche, at least for their time. Graphicly tried to become a CMS/socnet for graphic novels and comic books; Odeo (by Ev Williams and Biz Stone just before Twitter) did podcasts, Findings did quotes from books and articles; This. did one insight per day per user; Svpply did shopping lists; our service Pullquote did (and does) quotes.