much of coding and design is collapsing into one another, and the discipline of digital building is essentially becoming “logic.” many new startups enabling this, and today’s #wwdc announcements around swift/xcode are inline with this transition. — Scott Belsky
Code is a powerful abstraction due to its dual nature: both machine executable and human readable. It bridges the divide between two worlds: relentless, automated execution of billions of operations per second and squishy, intuitive, provocatively clever human minds. Code is a means of communication between teams, and it is precisely this human readability that allows software projects to scale.
The rise of dual-purpose representations will enable entirely new classes of “soft automation” — allowing humans and machines to collaborate seamlessly on problems that previously had been impossible.
In visual design and visual creative tools, systematic, contextual understanding of visual media via deep learning is unlocking higher-level abstractions that bridge the photographer’s understanding of mood, composition and subject with intricate, structured representations that are suitable for machine consumption.
By bridging this conceptual divide, soft automata enable both:
- New ways to scale production outputs via automation; and
- More precise communication between human teams
That is, soft automata are both operational and understandable.
In visual design and visual creative tools, the abstraction between editor and language is weaker, and editors tend to neglect the communication aspect completely. This is beginning to change with tools like Figma, Abstract, or Invision, which are fundamentally centered around the communication of complex visual ideas.
Even in this new generation of tools, however, designers are still mostly limited to communicating outcomes, not process or frameworks. This is starting to change with Design Systems, e.g. in “Declarative Design Tools”, Jon Gold describes a new level abstraction in building design systems
Now that we have the language to describe why our current tools are slow (they’re imperative! we have to move our meat-hands and tell the computer how to do our work for us!) and what we would like them to be (powerful aids to explore the full permutational space of our designs!), we can formulate alternatives.
Towards operational creative
Proliferation of media channels and demand for richer brand experiences is shifting advertising costs away from media back towards production — this is an ongoing trend that has been accelerating.
For example, today Nike spends > 80% of their ad spend on production, only 20% on media buys — the market average for such “non-working spend” at > 40% on average for digital and accelerating. This trend is fueling the growth of “new design” companies like Canva, Wix, Ceros, etc., as well as increases in headcount at production-focused agencies.
Consider the “last mile problem” in digital content production: how do I connect the branding design work done by Gin Lane or Red Antler to a 21st century marketing or e-commerce stack? Hire a production agency, freelancers or ask in-house creative resources to coordinate photo & video shoots, retouching, video editing, color balancing — how do I make more of it?
Today companies have enormously powerful engines for running marketing experiments and optimizing targeting and audience segmentation, but there is little comparable technology for optimizing and scaling creative production. You can only A/B test as many asset variants as your creatives can produce, leading to a production bottleneck: the copywriter who needs to come up with ad copy, the production agency who needs to produce a 5s instagram spot, etc.
More importantly, how do we empower creatives to be able to take on more of this work themselves?
Facet tackles this problem with “Patterns”: clearly specified design systems that control the look and feel of your raster content. Patterns are a base transforms every image or video must inherit from to fit your brand concept. And Patterns are dual-system: suitable both for operational production implementation, as well as communication among stakeholders.
Operational design systems such as Facet Patterns are a must-have for content production. Once a creative at your company had ported over your existing branding system to Facet, everyone has to start using it in order to keep in sync stylistically. Once you’ve demonstrated the desired outcomes once, Facet can work out how to extrapolate across photoshoots and video clips.
Automated creative processes allow creative teams to explore a greater number of possibilities and achieve more while connected content variations in real time across platforms, audience segments, all while adhering to many different format sizes and requirements.
For visual content, Facet enables more precise and coherent communication of style guidelines, giving creative teams the tools necessary to scale their image and video production efforts and interface better with increasingly automated marketing and e-commerce solutions.
Soft automata bring dead pixels to life
At its heart, Facet Patterns area system for structuring “dead pixel” content and imbuing it with a natural semantics. Pixels cohere into objects and relationships: a woman with brown hair is sitting in an office chair; a house-plant casts long shadows along the floor; background highlights are flat magenta and cyan. Armed with a limitless amount of knowledge like this, Facet can intuit the intent behind the artists actions: marquee selection of a color region becomes selecting skin-tones; flattening tone curves locally becomes balancing visual rhythm, a flurry of dodging and burning becomes brightening a smile.
Better machine learning algorithms are the primary catalyst for soft automation: bridging human context via natural language or visual understanding. As machines come to understand the context of human experience better and move up the abstraction stack — the nature of human-machine collaboration changes.
When Facet structures an image, it unlocks all of the hidden relationships and meaning. This can power queries like “find me all the images with this model wearing this item of clothing”, but also semantic edits: change the background to match my brand guidelines but keep the foreground imagery intact.
Facet Patterns extend the analogy of design systems to imagery and vfx/video editing, empowering artists to unlock and structure systems out of pre-existing parts and relationships, rather than having them build everything from a blank canvas.
“How do you get your creatives to talk to your engineers?”
The change that we are seeing today is the start of a tighter integration of design into the tech stack. In the future, design and engineering will be seamlessly integrated, with no need for clunky intermediate “human parsable” formats like presentations, design briefs, style guidelines or mood boards.
Facet’s content and design analysis platform sits at the highest level interface between human and algorithm: a human designs a creative experience and Facet handles the internals of interfacing that with your digital asset manager and the rest of your marketing stack
Instead, all design choices, layout, scaffolding, typography, color and “style” will be crafted by a creative in human understandable terms, and then “compiled” into production assets with the correct look and feel. In the same sense that a programming language bridges the divide between human software engineers and the underlying platform, Facet patterns allow designers and creatives to express design constraints in meaningful terms, rather than pixel by pixel.
Previous in the series:
Sign up for our preview release at https://facet.ai