The Death of the Web Page

In a post on PandoDaily we talked about Single Page Application Architectures (SPAs) and how APIs, coupled with Javascript would very likely be the major paradigm for Web Applications going forward. Such applications work by moving away from using separate HTML pages to loading a single HTML Page, a Javascript UI and a powerful backend API. Although it’s a relatively new phenomenon it’s already being adopted for Web Apps such as Gmail and now Zendesk.

Adopting SPA means a radical shift to a web built on powerful platform APIs + rich “desktop like” clients which are more interactive, faster and can be tailored to different delivery platform. The same API can then also be used to drive equivalent native mobile applications or even B2B integrations. This change makes a lot of sense in particular for Web Applications which are designed to be used to complete some task or interactively explore data.

The article raised the question though:

If this is true for Web Apps, what about the rest of the Web? The Content Web? Will these sites also move to Single Page formats and what are the implications of this?

In other words, if SPA architectures are good for “functional” Web Applications, why not apply the same techniques to content heavy sites?

Technology Challenges

The immediately obvious answer would seem to be “no, this wont happen any time soon” – in particular SPA architectures are only just emerging and there are a number of technical challenges which affect content sites in particular, not least:

  • Older browsers often don’t have the javascript performance to perform well for SPA sites – for general content sites where audiences are more diverse this has a stronger negative impact than than cutting edge tech sites.

  • Basic content navigation controls are also harder to get right – in particular, the management of browsing history – again a modern browser is needed that implements the full stack of HTML Navigation APIs.

  • SPA type sites, while potentially much richer in experience are harder to build and maintain – requiring a greater degree of programming skill and modern tools than simple page based sites.

  • Lastly, and perhaps most crucially, there are open challenges on how SEO for single page sites – while there are initial solutions, such as retaining landing pages and ensuring the single page contains “#!” URLS which google indexes (not all search engines do), this is a very real problem for content heavy sites.

Of these problems however ironically, the first two are being solved relatively rapidly precisely by the wider adoption of more and more rich Javascript front-ends for Web Applications. As such applications proliferate and as popular sites such as Pintrest and Twitter use more and more Javascript in their experiences, the browser upgrade cycle is accelerated and adoption of HTML5 capable browsers is growing quickly. Tooling for building Javascript heavy SPA style sites is also rapidly becoming available with frameworks such as Meteor.js making it increasingly easy to build them. As these trends continue building SPA versions of content sites certainly becomes more viable – although but this still leaves the SEO problem which we’ll come back to later.

… and Benefits

Before going in further however, the question really is – what would the benefit of SPA for content sites be? – just because something is possible does not mean it will be used or become widely adopted. At first glance there is not much reason to shift away from page based models if they already work. However, projecting forward there are a whole bunch of reason why this type of architecture will likely be be very compelling for the content web – and they closely mirror the reasons for the shift in functional Web Apps:

  • Single Page can produce much slicker, more customized and faster experiences for content consumption just as it can for web apps – in particular, it would allow tighter coupling of content types, peripheral content, advertising and consumption experience. Typical page driven reading/viewing still requires page reloads for every new piece of content and this creates a significant discontinuity no matter how well it is managed. The obvious example here is the example set by native app content consumption applications such as Flipboard. These create a phenomenal content consumption experience seamlessly – no pages involved. If Flipboard came to the web, it would be a Single Page App.

  • Just as with Functional applications, SPA would allow content heavy sites to converge experiences of the Web Based clients with those they have on native mobile platforms. While this seems minor, it is likely to become increasingly important as UI metaphors from Mobile establish themselves – the Web will need to “Catch Up” in usability – the browser back button may become alien territory for users who experiences the Web first on Mobile.

  • Lastly, content needs to reach wherever users are and in whatever form is appropriate for that platform be it television, tablet, mobile of desktop – all businesses on the web will become “multi channel” (see thoughts on this here and in the excellent comment thread on Fred Wilson’s recent post “Rethinking Mobile First”). This means that a powerful API which provides uniform and flexible access to content will become the major distribution hub even for content companies and overlaid Web Sites, Mobile Apps the APIS’s primary clients. This approach also opens up a “wholesale” channel for content to be re-used and re-sold through partner channels. We see such strategies from our clients constantly and it is perhaps the most compelling reason why content will also end up becoming API+SPA & App driven.

So – will it happen?

While this may seem a stretch, we would argue that the motivations in favor of this shift seem far too compelling to think that it will not happen – just as for Functional Applications, moving to a more “App like” experience for content is likely to be irresistible soon especially to large online properties (arguably it is already happening as sites like Pintrest for example rely little on URLs to identity content). As tooling becomes available this will likely propagate. The change may take longer than for Functional Applications and be much patchier, but it is likely to affect larger properties first.

So where would this leave questions such as the SEO question? Web Advertising? and the Web as a Whole? – this is difficult to predict but it seems very likely that:

  • SEO techniques will become sufficiently adapted to work for Single Page Cases – this is already becoming possible for Google and other search engines will likely need to follow suit. However, for SEO this may also be part of a very deep transformational change as content discovery changes from a pure Web search paradigm to one driven increasingly by “over the top” mechanisms such as social sharing, rankings and other mechanisms which could become embedded in rich interfaces. It may be the case that for some sites, SEO becomes less critical and this may even begin to present challenges for search engine business models.

  • For Advertising, it seems that as long mechanisms remain to address screen real-estate in SPA applications, advertising will still be a lucrative business model – potentially SPA will allow even more sophisticated integration of promotional content into applications than before.

  • The “Appi-fication” of the Web implies the channeling of content into channels which are served by custom interfaces. On the surface, this may mean a loss of the “connective tissue” which makes today’s Web so powerful – links and rapid navigation between sites. However, hopefully the firm grounding of this technology in HTML5, REST conventions and other very stable protocols such as openAuth will mean that all the elements for openness are retained. It is difficult to predict how this play out, but Web conventions will still play a powerful role in structuring interfaces – the question will be how it adapts to accomodate more and more changes in content structure.

In some ways the SEO / Advertising debate is an exact extrapolation of the problem currently faced on mobile – while much of the mobile debate is focused on “screen size” for advertising, the real issue is distribution for apps / advertising (and the shift in who controls it) as well as a complete rewriting of the rules of “SEO” on mobile since Appstores have become the dominant distribution channel.

In many cases the content delivered into apps once they are installed is simply invisible to any kind of indexing, no SEO can be applied and there is no cross linking between content in applications. This is arguably a completely broken version of that the Web is today. The assumption to date has been that such issues will not arise on the Open Web – however smart solutions will be needed to ensure this since the Appification even of content seems very likely to happen. The web has the right DNA to get this right, but there will still be a strong shift to a more App-like experience and hence technology is needed to ensure that discoverability, cross linking etc. need to be maintained to avoid some of the issues arising on mobile today.

What will replace pages? What about HTML?

The idea that much of the Web will move from a page model to one which is more application like is difficult to grasp, since our entire experience of the web to date has been based on a page navigation paradigm. No doubt any transition will take significant time and likely be only partly noticeable as user experiences shift.

However, the key insight is that while pages as a navigation paradigm may be less important, this does not change the importance of the structure of data and operations which underlies applications. In particular, applications will be driven by powerful APIs which serve up content and functionality in the first place – these APIs provide an underlying layer of accessible, potentially indexable and addressable content. These APIs arguably become a new lower version of the Web divorced from the final end user experience.

People have predicted the The Death of the Websites before – but it seems a little more subtle than this: a shift away from the “page” model most of all.

To SPA or not to SPA?

An obvious question given these thoughts is “what should we do now?” – build an SPA version of our site? wait? resist?

There is no hard and fast answer to this since every site is different – certainly for content sites, SEO is still crucial to survival and systems need to be built to deal with today’s reality. However, the following steps are highly valuable in future proofing any strategy:

  • Evaluate how an SPA+API type strategy could help improve the content experience for users and especially make them more uniform across mobile & web.

  • Ensure you have a well charted API strategy that takes into account the new multi-channel world that is arising and positions you well to deliver core assets (content / services) into whichever environments your customers are likely to demand it. (You can talk to 3scale about this!)

Rather than being a threat, in many cases this transformation of the Web is a huge opportunity for many players – the era of truly multiplatform rich content and applications is only just beginning and online behavior continues to integrate Web based services in more and more intricate ways. New technologies such as APIs and the Javascript frameworks, are making it increasingly easy to create phenomenal user experiences – so there is plenty of scope for progress.

Does this mean we’ll all be native apps soon?

If this sounds like doom and gloom – for the Web it isn’t. What will change is the way the Web is structured and individual web pages will become less important – but this does not mean the “death of the Web” in any way – arguably the Web will become even more powerful – as a Web of data and functionality addressable by API from many many new types of clients.

How long this will take to come about is hard to say, however the beginnings of the revolution are already apparent as more and more APIs become critical and leading edge companies begin to pick up new interface technologies.

Looking forward to people’s thoughts!