What’s Next for Digital Product Design?

See the State by Bret Victor

Working in the fast-changing world of digital product design, there’s always some new and exciting tools and techniques to learn. The flip side of this coin is that you must keep learning to keep up with the industry or your skills become irrelevant.

So naturally it would be helpful to know how our practice will change in the future so we can figure out what to learn next. This article is my attempt to look into this future.

Paul Saffo said that if you want to forecast what’s coming, look back twice as far. Luckily, I’m pretty old. So looking back to, say, the beginning of the web is largely a matter of clearing out the cobwebs in my brain.

From the early 1990’s until now I see four overarching trends in digital product design:

  1. More direct design: Direct design is when we directly craft the artifact, as a blacksmith does when hammering a piece of iron. Indirect design is when we document the design first before making it. As websites grew larger and larger and we initially grappled with how to document the design of increasingly complex systems. Now we design products more directly using frameworks and prototyping tools rather than documenting all the ideas first.
  2. More validation: In the early 90's validation of our design amounted to QA testing. Over time we’ve adopted more human-centered research and testing methods.
  3. More sophisticated tools: We’ve progressed from cutting up Photoshop assets and hand-coding pages on static websites to using rich prototyping tools to design for a variety of interfaces and technologies.
  4. More sophisticated skill sets: The jack-of-all-trades Webmaster role has been split into many specialized roles each with a deep skill set.

To help make sense of these trends, I’ll break them down into three waves of progress roughly divided into decades:


In the 90’s digital agencies and advanced internal groups generally copied the process that worked for advertising agencies. This is opposed to IT-based groups who generally trailed in design maturity and were just starting to inject design expertise into their development-centric processes.

  1. Direct Design: Dreamweaver helped non-developers build websites, but mostly we coded web pages directly by hand, and made simple wireframes in tools intended more for desktop publishing. Mobile devices were so simple that little design was necessary.
  2. Validation: The more progressive designers adopt usability testing.
  3. Tools: As browser creators competed with non-standard features and stymied front-end progress, most of the advances are on the backend in the form of middleware and databases to design for.
  4. Skill Sets: The Webmaster role is broken out into designer and developer, and then copywriter, information architect, front-end developer, and back-end developer. IT groups still look to business analysts but start to hire specialized roles.


The 2000’s saw explosive growth of the web and mobile accompanied by new methods and tools.

  1. Direct Design: We actually did less direct design in the 2000’s, documenting our designs more as we grappled with how to design for new technical capabilities and how to communicate designs to others.
  2. Validation: Agile, which was invented in the 90's, spreads beyond early adopters and releases product to customers earlier and more often. Lean Startup leads to Lean UX and further increases use of design research.
  3. Tools: jQuery made development of rich interaction easier and HTML 4 & CSS 2 encouraged more robust web design, but mostly these helped developers implement design.
  4. Skill Sets: The progressive use of web applications and the emergence of touch screens require more interaction design skills. More specialized roles such as design researcher, interaction designer, and product manager grow in popularity.


  1. Direct Design: High-quality CMS templates (e.g. Wordpress, Squarespace), front-end frameworks (e.g. Bootstrap), and graphics libraries (e.g. The Noun Project) offer us robust, responsive elements that chip away at the practice of indirect, bespoke design.
  2. Validation: Lean UX methods become common; Agile passes 50% adoption rate and continuous delivery accelerates live testing. A/B and multivariate testing tools increase the use of quantitative testing.
  3. Tools: In 2007 I made a call to the design community to make tools, and in my warped perception everyone listened! With target platforms reduced to iOS, Android, and web standards, design tools are released faster than we can track them. Purpose built digital design tools such as Sketch and Axure replace static tools such as Photoshop and Visio. Tools like Axure allow us to document a design and make it interactive faster than if we drew each individual interaction state. Hybrid tools like Framer can act as an integrated development environment for designers. And there’s now many options for online testing.
  4. Skill Sets: Responsive web design goes from possible to required. Roles become further balkanized and specializations such as design researcher, product manager, and content strategist become firmly established. More mature groups internalize agile and lean methods so well they integrate it into processes of their own design.

What’s next?

Predicting the future in writing is sure to bite me in the ass later on. To keep it simple, I’ll ignore black swan scenarios and simply extend these trends to guess what might be coming in the next few years.

The industry zeitgeist now is one of well-understood business models driven by an expanding number of technological platforms: The Internet of Things, voice-activated assistants, machine learning, virtual reality, and augmented reality. What if these same four trends continued but accounted for these new platforms? Let’s jump ahead a few years and find out…

Direct Design

Screen-based prototyping tools like Sketch and Framer now output perfect production-ready code, so on many projects there’s no need for front-end developers unless we’re creating something esoteric.

Craft plugin for Sketch

And in fact there’s less need for traditional screen-based designers in general, since our algorithmically-generated design tools allow non-designers to create quite nice simple artifacts by themselves.

The Grid

Instead many designers focus on the hard tasks of solving interaction challenges of machine learning, virtual reality, and augmented reality. Once they progress beyond the sketch stage of design, prototyping tools for these technologies are targeted at developers who co-design with the designers by their side.



Prior to about 2007 usability testing was expensive because it required renting a lab and recruiting participants. It was more expensive than designing. Once usertesting.com and other services launched testing became cheaper than designing and it made more sense to be lean and test everything first before building it.

Now with developers using continuous delivery there’s more pressure to feed the engineering beast. To keep up, designers practice continuous usability testing; they constantly have tests running to validate new hypotheses or compare competing design directions. To facilitate continuous usability testing the new generation of prototyping tools publish designs directly to testing services. The testing services in turn connect directly to agile workflow tools so the winning design automatically gets prioritized in the backlog.

Ultimately we validate exponentially more ideas than we did back in 2016 (and spend less time in meetings too).


Skills we used to perform manually have migrated to our tools. For example, no one “designs” responsive websites anymore. The algoritms in our prototyping tools offer us options to test and testing reveals the best approach.

Platforms as a service (PaaS) have become so easy to use that designers set up their own backends. PaaS evolved from easily provisioning a server (Amazon Web Services) to eliminating the need for server configuration (Heroku) to click-and-code (HyperDev) and now Google offers drag-and-drop API creation. Designers who only need existing functions can set up their own backends without the help of a developer.


No one uses Arduino or Raspberry Pi for IoT prototyping anymore because the programming took too long and in the end they couldn’t do much. We now have IoT prototyping parts that do everything with plug-and-play ease of use and they’re all the size of a fingernail. Think littleBits for professionals. True industrial design, however, continues to belong to industrial designers; 3D printing never threatened industrial designers, and as devices are only getting more complex real industrial design skills are still needed.

Particle Maker Kit

Google open-sourced their voice-activated assistant in an effort to dominate the market. Like search, it’s futile to try and document new behavior for an intelligent personal assistant without prototyping it. Designers use new software tools to mash up language hooks with IoT prototyping tools and IFTTT-like functions to create entirely new products.

Skill Sets

The trend toward specialization continues as technologies such as IoT and machine learning require deep skill sets that don’t overlap much with each other. New organizations have popped up to support designers in each sub-industry. Universities continue to offer more general design education but gave up trying to keep pace with state-of-the-art techniques.

So What Should We Learn?

If the above trends resonate with you, it could be helpful to ask yourself how your own craft could advance in the future. These specific questions might help:

  1. Are the tools I’m using the state-of-the-art, or is it time to learn something new?
  2. Can I create a new tool to improve something manual I do now because the technology has become fast and cheap enough?
  3. Are there tasks I’m performing manually that could be automated?
  4. Are there decisions we’re making based on gut feelings that could be tested easily and cheaply enough with testing services?
  5. Do my skills align with the current or next generation of technology platforms?