Madi Weland Solomon • 4 minutes
In 1983, Thomas Wolf wrote an essay for Esquire about Intel Founder, Robert Noyce.
Noyce grew up in midland Iowa on a family farm and he revolutionised office design by incorporating the egalitarian landscape of his childhood. Gone were the cubicles, offices, and hierarchy. Instead, he introduced the first “open plan” space where employees were perceived as equals and had free range of the farm, er, I mean firm.
A few years later, the concept of Content Strategy helped shift editorial understanding of web content as more than just uploading a catalogue, to the art and craft of combining writing, linguistics, digital assets and technology in a hyperlinked medium. We have now entered a third phase where it is no longer sufficient to just provide content and products online, but to offer customers a full smorgasbord of consumables from products to rich media with social interaction that can keep them busy for hours. Businesses refer to this as maintaining their Feed.
What’s a feed?
A feed originally referred to a stream of syndicated news that was constantly updated. Today, media companies use Feeds to engage customers within their platform long enough to turn a snack into a meal. We haven’t travelled too far from Noyce’s farm with our metaphors as it’s not difficult to imagine Facebook as one giant trough. But I digress.
There are several flavours of feeds (I couldn’t help myself) that are currently in play in the 21st century and they are generally multidimensional flows of information and products that are niche, rich, and personalised.
Successful feeds rely on a deep understanding of their customers and use this data to create profiles upon which to build. This usually involves accumulating data from multiple sources: volunteered registrations, harvested through free services, purchased from large aggregators, pulled from open data, and good old fashioned qualitative and quantitative research. Add machine-learning technology to this suite and we experience the early clumsy attempts of recommendations engines. I call them clumsy only because the data patterns that the algorithms are processing are still incomplete. According to my Netflix recommendations, who has one of the most sophisticated data models for categorising viewer preferences, I am a gay black man because of my viewing choices (I love Rupaul). I take no offence to this, but they are missing out on many other opportunities that have not yet been codified.
Advancements in data will soon be able to accurately determine my mood, family and economic status, even my health, and adjust accordingly. Of course, this begs the question if I even want that kind of intimacy from a machine. Managing the delicate balance between creepy and helpful will be one of our future tasks.
Findability, discovery, and content management were the business priorities in digital transformations. Not so long ago metadata based on library standards were adopted and applied in a labour intensive exercise. It was a common thinking that the more robust the metadata model, the more granular the taxonomy, the better productivity. This, unfortunately, was not true. Technological developments have advanced enough to automate much of the basic functions without the use of a taxonomy and it turns out that less is more when it comes to metadata. Yet businesses often remain stuck in the cozy process of endlessly cataloging.
A few years ago, a division of an education publishing company wanted to understand how taxonomies might help manage modularised content that had been de-aggregated. Chosen members of the content management team attended a series of educational modules on the subject with guest lectures and workshops. Once sufficiently trained, they enthusiastically set off to design their first iteration. The results were revealing. They had built their taxonomies (there were many of them) based on the actual physical structure of a text book starting with their top category of “Chapter One.” The messages gleaned from this exercise were:
- the content managers were absolutely resistant to pulling apart a cohesive whole. Fair enough. In education, there are logical sequences and to de-aggregate was to de-value
- the structures were designed to provide editorial efficiencies for content creation and assemblage
Resistance and editorial efficiency were short sighted justifications and the efforts proved useless in an omni-channel world. The structure as designed did not prepare their content for re-purposing, analysis, re-aggregation, or help with measuring and benchmarking against student outcomes. Their customers did not care what a section was called and tracking behaviours around concepts such as “Chapter One” offered little insight into the value of the content.
Building online catalogues are fine, but efficiency is not a business proposition.
A recent Harvard Business Review article (Survey: People’s Trust Has Declined…) stated that trust in government, institutions, and businesses have reached their lowest levels in 17 years. People around the world no longer trust “the market” as the great economic equaliser and this should make us all pause. Attracting attention from market-weary customers will require different approaches than the one we use today. A better and fuller understanding of our customers is needed and this will require embracing some messy imperfections.
No matter how clever the algorithm or recommendations engine, humans will still follow impulses that are basically mammalian. Consider the recent reports on the vinyl record revival, or the sudden reappearance of independent bookstores. Or read David Sax’s excellent book, Revenge of Analog: Real Things and Why They Matter to learn more about how the backlash against the forced techno utopia are advancing new business markets that are succeeding and celebrating old school imperfections. Humans, in all our striving, still long for community, contact and touch. It’s time we incorporate these in our data models as well.
A shift in perspective helps create data models with more human elements. Big data capabilities are impressive but it cannot tell us what to look for. A clearer understanding of how customers interact with content can help inform direction in content and product creation thereby “closing the virtuous circle” of data and keeping the dialogue going.