Content migration, information decay and technology delay

Content migration, information decay and technology delay

I think Nimbus has the potential to be the most impactful product I’ll ever work on.

When I tell you that Nimbus is a content migration tool, you’ll think that’s hyperbole. But I believe that ineffective content migration tooling is one of the biggest impediments to adoption of technical innovation in the digital landscape, and the standard ways of dealing with it are contributing to a digital dark ages. We’ve never created more content, nor preserved so little.

Unless, that is, we solve the content migration problem – and I think we can.

The content technology problem

Poor content migration capabilities slow down adoption of the best and latest content technologies until the value of the new technology exceeds the value of the content that will be lost in the process.

I wrote once, not too many years ago, that we were heading into the 7^th generation of digital content platforms. Don’t bother looking for the article, it died with the 3^rd Gen platform that housed it, already outdated at the time the article was published. The platform was outdated, that is; the article would still stand up, and I’ll end up having to reconstruct it soon to talk about the 8^th generation that’s now emerging (that’s a teaser right there, subtlety fans!).

I mention this for two reasons: the first is that digital technology has been moving at breakneck speed for 25 years – each of those 7 generations has been a major shift in how content is stored, assembled and presented, and inconsistent with all of the previous generations.

The second is that as an industry, we’ve made major progress in technology and development mobility – the shift towards loosely coupled, composable platforms with proper separation of duties embodied by the shift towards headless and hybrid headless systems allows for a more incremental approach to technological change, supported in working practices by a shift away from waterfall delivery methods.

But we haven’t developed equivalent tooling for adapting and preserving content that retains value when we switch platforms at a rate of about once every three years.

So, as an industry, we have built a terrible decision cycle: first we restrict ourselves from progressing in new technology to avoid losing the value in our content libraries, until we can no longer bear how far we have fallen behind. Then we burn those libraries to the ground, taking only simple content we can transport with minimal effort or that we’ve flagged in limited collections for what amounts to manual re-creation because we can easily see that the content still has enough value to be worth special effort.

The content value problem

Poor content adaptation capabilities leads to premature deletion of content with limited current value due to undervaluation of its future, latent or potential worth.

Stewart Brand’s old aphorism was that information wants to be free (because it is easy to copy); and that information wants to be expensive (because it is so valuable). Today we have a new version: content wants to be preserved (because it is hard to create); but content wants to be forgotten (because it is hard to update).

There are a few factors in play here. What is the current value of an old piece of content – does it still have meaning and does it still have an audience? If it has neither, does it have any archival value, as a record of the time it was created or as context for other content? On the flip side, what is the cost to update the content? Does the presentation of the content need to adapt to new design components or consumption channels, does the content model need updating for a new CMS, do the actual words need rewriting to adapt to new information or context?

Let’s pick on a famous example of an organisation undervaluing the future worth of content: the BBC Archives. Thousands of hours of TV programming and hundreds of thousands of hours of art, craft, and human creativity (also some episodes of Dr Who) were destroyed because at a particular moment in time, the medium was considered more valuable than the message stored upon it. Destroy the content, save the tape.

In digital, we’re re-enacting this process constantly, typically driven by limited metrics of current value that are easily measured: primarily popularity and search engine rankings. And while we’re not so worried about the price of the storage medium any more, we are worried about the cost of reviewing updating content. Reviewing an article, updating the contents, updating the presentation for a new design system, copying it into a new content model – the time adds up, and it’s time that could be spent on new content instead.

So who can stop to assess and sift potential value when there are impatient years of technical standstill to cut away and renew? Set the fire, rescue the obvious jewels, build anew from the ashes. And don’t worry about the value of what you’ve lost until you’ve lost the audience – if you ever even notice.

The content experience problem

The ghosts of your forsaken content haunt your new content experiences. This can be clumsy and blatant, with vestigial navigation, cul-de-sac calls to action, truncated content tags, and redirect strategies that placate search engine authority but leave visitors from external sources marooned on high-level landing pages. Or it can be subtle, a patchwork of missing colour that once gave richness to the weave of the content you have chosen to transfer. Like Derrida’s différance, the underlying shape of content left behind can subtly alter the content carried forward; old case studies that gave credibility to claims of performance, knowledge bases for old products that may only rarely have been consulted, but which told prospective purchasers for your new products that you will support them in the long-term.

This is another form of content value that is difficult via traditional metrics to assess, and another form of damage that is difficult to evaluate. But instead of the damage being to the content you didn’t value, it subtracts from the value of the content you did. It is a subtler, more insidious damage.

The content mobility problem

Content value is multi-modal, spanning language, imagery, sound, design and presentational as well as semantic relationships, and even modern content systems do not model all of these modalities in programmatic interfaces.

The content repositories in 6^th and 7^th generation CMS platforms are much more rational than in older platforms. Headless platforms have ensured that content models can be designed to the break down content into structures that are human readable and understandable, rather than only being accessible to the platform creating them; that separate form from content so that content can be distributed to many channels. And in being designed for multi-channel consumption – well, surely your next content platform is equivalent to just another content channel, right? So problem solved?

But of course it’s not that simple. The situation is better but the problem isn’t solved. For a start, effective content models aren’t consistent across platforms; more importantly, they aren’t consistent across redesigned experiences. Your content model will change over time, over technology, over design.

Illustrating this, for all the genius of content grid approaches like Netlify Connect, borne of Gatsby’s Valhalla technology, and the platform migration assistance provide by allowing you to present content from multiple CMSs into a single experience, they don’t actually migrate the content, they give the illusion of migrated content to give you more time to tackle the migration itself. Eventually, unless you want to maintain multiple content platforms or models in perpetuity, you’re going to want to retire older properties, at which point the illusion shatters and you need to complete your actual content migration.

And you might not be this fortunate. Your old content model might be badly designed. You might be migrating from an older platform without a content delivery API. You might be migrating from a newer platform that privileges web delivery with a visual layout management layer that isn’t represented in the content model.

And this takes us back to that separation of form and content, which contains a major misconception at its heart: form and content are not independent, and form both contains and influences meaning in its own right.

So if you take a data-driven content migration approach, whether that’s API driven or database-focused ETL, you will miss out a key modality of meaning that’s typically represented in your front end logic, not your content model. Hundreds of thousands of sad, ugly, hard-to-understand pages of content clumsily moved without context or nuance or consideration for content objects and visualisations that no longer render bear testament to this loss. It’s like trying to read the whole web through the Wayback Machine.

The content project problem

Content itself is undervalued in technology-focused digital projects. The effort and skill required to preserve content value is underestimated and under-resourced, and therefore under-delivers.

Technology – whether in respect to content delivery or content consumption advancements – has been the driving force behind change in the digital landscape, and for digital delivery projects, for all the reasons already described.

So when the time comes to replatform, redesign, transform, and you have the backing to create new, compelling experiences for those users backed by powerful new technology, you know, as someone responsible for the value of content, that you have reason to be fearful. You know from experience that the content migration will be allocated too little time, too few resources, too little senior attention, until it’s too late and you’re left to deal with the problems: delayed launches, paralysed content operations and lengthy content freezes, overworked content teams trying to learn new systems on the fly while they ignore their day job, teams of temporary workers without investment in the value of the content making decisions, clumsy rule-based automated migration systems that do not and cannot understand the content value of the – to them - abstract data they’re transferring.

Nimbus: our solution

Digital content will not start to be properly preserved and updated if we don’t bring down the costs of migration, whilst simultaneously expanding the coverage of migration tools to cover more sources of value.

Nimbus is our solution to these problems. It makes content migration more effective and more efficient. Using Nimbus, we not only bring down the cost and time needed to migrate content between platforms, but we preserve a wider range of content value than any traditional approach, whether manual or automated. And we restore lost value, adapting content to modern delivery standards and identifying areas where content needs to be updated.

There is an AI component to Nimbus. It’s a critical component – it generates the semantic understanding of the content that enables us to understand the value that needs to be migrated. We use our AI to understand your content, how it fits together, and how it needs to work in new content models, new propositions, new experiences, new cultures, and new audiences, to balance current, latent and future value and distinguish between publication, adaptation and archival needs. We analyse the actual user experience of content in context, just as a human editor would, but with higher attention to detail and absolute consistency.

But Nimbus is not just AI, and it’s not about AI – modern AI was just the final piece we needed to build out a solution to a problem we’ve been wrestling with for decades.

Beyond AI, Nimbus is also built on a massively parallelisable processing framework that can easily handle thousands of pages a day in multiple passes, so you can migrate your content both quickly and iteratively, without cramming it in at the last minute or imposing lengthy content freezes.

It’s also a flexible integration connector, that can deliver content into any modern delivery platform or multi-platform stack – not just CMSs, but DAMs, PIMs, WMSs, KMSs, LMSs, DXPs – to any content model.

And it’s a human-connected service, powered by experts who care, empowering content teams who care.

We have big plans for the future of Nimbus – we think the potential to solve really foundational problems in modern content is huge. And this is why I think Nimbus has the potential to be the most impactful product I’ll ever work on.

But this isn’t a hypothetical solution. Nimbus is here, it’s live, and we’re already delivering projects with it. We’d love to tell you more about how it can help you today – please get in touch.