The Numerati and how data-mining and personalized content may impact design." /> The Numerati and how data-mining and personalized content may impact design." />

Dmitri Siegel | Essays

Design by Numbers

Business reporter Stephen Baker’s new book The Numerati explores the way that marketers and retailers are leveraging personal data to create customized experiences and targeted messages. The book details the staggering amount of data we leave behind every day — making a call, browsing the Internet, using a credit card, swiping a Metrocard — and how these actions are tracked and stored in various databases. Baker’s book is an ode to the mathematicians and engineers (the Numerati) who manage to sort through these mountains of raw data, but what does the rise of targeted messaging and personalization mean for designers?

The kind of data-driven customization outlined in Baker’s book is already well established online. The same cookie that serves up the “you are logged in as…” message at the top of a web page tracks what you’ve browsed, bought, where you arrived from, where you exited to, even what resolution your monitor is set to. Online advertising is targeted based on your browsing habits and many sites already serve up personalized content based on your browsing history. Looking at how these innovations have affected designing for the web, can give us insight into the Numerati’s impact on the broader world of design.  

The old cliche goes, “Half of my advertising works, I just don’t know which half.” It has become a mantra of the Numerati to make this predicament a thing of the past. By harvesting personal data the Numerati promise to offer timely, actionable information that will vastly improve the effectiveness of content. Advertisers will be able to serve up only the most relevant advertising and response rates will increase exponentially because you’ll be showing the right content to the right person at the right time. But what the Numerati (and Baker himself) overlook is that this numerical approach ultimately feeds into a subjective process called design. Was an ad effective because of its placement and timing? Or because of the way the photo was cropped? Does an ad work better if the type is underlined? Would Trade Gothic have worked better than hand-drawn type? 

Vexing questions like these have created a distinct issue in designing for the web: the drastic increase in the use and sophistication of testing. In digital design, quantification goes way beyond focus groups and user experience testing. Multiple versions of creative executions are regularly deployed to audience segments — with results monitored in real-time. There is no decisive moment when a design is “delivered”: any image can be tested, tracked, tweaked, and then tried again, multiple times over based on results. Just as non-destructive editing software transformed the art of editing film, A/B testing has the potential to transform design. This may seem like a ghastly development, but testing and targeting can (and should) open the door for experimentation. The ability to “try anything” may degrade the decision-making ability of a designer, but it places a renewed emphasis on process — the act (and hopefully joy) of making. In the era of the Numerati, designers may have to trade the glory of fashioning great monuments of design for the pursuit of a fulfilling creative process. 

Another effect of customization and personalization is that it necessitates new forms for the delivery of content. Designers increasingly have to think in terms of systems and templates, designing not just content pieces, but channels into which that content can be piped. Digital design requires an extended initial phase that is essentially content agnostic. Instead of formatting text and image, the designer must create a flexible hierarchy of “content slots” that can adapt to customized content. This process also involves designing for two audiences: the public and the content administrators. They must design with as keen an eye for the content management process as for the user experience. Once a flexible content system is created the second stage of the digital design process involves creating the multiple, personalized versions of content. Designers must become adept at making subtle changes that will speak to different psycho-graphic groups identified by the Numerati. The same video may need a “bohemian” version and a “sports fan” version. A single Blog may need to be programmed for “Belongers” and “Achievers,” and so on. Designers are going to be called on more and more to translate the same ideas into multiple dialects. For design studios that serve multiple clients, modulating the visual language for different audiences is a familiar exercise but the distinctions between audiences are getting ever-more subtle and the sheer number of audience segments require flexible design strategies. 

The digital designer has to understand the implications of using data to drive creative execution. Often, the kind of messaging that yields the best short-term results, loses effectiveness over time — and degrades a brand in the long term. For instance, a garish flashing banner ad touting a deep discount will “perform” well and drive traffic, but the people you attract with this approach are what Baker calls “barnacles.” They are not bought into your brand; they are merely hanging around waiting for a deal. It almost always falls on the designer or art director’s shoulders to make these arguments. This is the traditional role of the creative side of the house, but even in the most “brand-focused” organization the Numerati will be able to make ever-more compelling arguments for short-term gain over long-term investment. Designers need to be able to advocate against data. In fact, I would argue that design in general would benefit from some strong arguments for its long-term effectiveness beyond vague notions of “brand-building.”

Beyond the impact on design, the widespread use of our personal data raises serious cultural issues as well. If your every experience is tailored to your previously established preferences, how will you ever expand into new experiences? Targeting ultimately limits experience, and as these methods move into areas like health care and voting, how will the public protect themselves from being unfairly discriminated against? Stephen Baker argues that the public has no choice but to become educated and conversant enough in the language of the Numerati to protect their rights (not to mention their privacy). This is true for the designer as well. There is a great and growing need for designers who can have a critical dialog with the Numerati. These designers will be able to not only digest and learn from the statistical analysis, but will offer a counter-point to the short-term incremental gains that it can offer.

Posted in: Arts + Culture, Business, Media, Technology

Comments [16]

How does one become conversant in the language of the Numerati?

And how do we find the language to communicate the value of ideas, concepts, visual seduction to those who only look at data?

How do I communicate to Amazon that just because I bought a map of Chicago 4 years ago before a meeting there does not mean I want every Chicago related item they offer? (London is another story.)

Perhaps I've found an idea to explore for my dissertation.


And how did the above post before I hit the button? And what happened to the "preview" option?
Michelle French

i find the last point very interesting.. the way we are actually shrinking our horizons for cultural experience...


As I delve more and more into the digital world, I find that message is no longer King, but subservient to many Masters. The overarching campaign giving way to something resembling advertising micro-loans, where content and ideas become candy-colored Chicklets: small and easily digestible, but in the end fleetingly unsatisfying.

Speaking for myself, my tastes are always evolving. If such advertising, becomes commonplace, I fear that we will be forever chasing millions of ever-moving nano-targets. Which in turn begs the question: when will though and concept be replaced by an automated preference algorithm based upon past behaviors?
James D. Nesbitt

Blonde in a Boat

There was a blonde driving down the road one day. She glanced to her right and noticed another blonde sitting in a nearby field, rowing a boat with no water in sight.

The blonde angrily pulled her car over and yelled at the rowing blonde, "What do you think you're doing? It's things like this that give us blondes a bad name. If I could swim, I'd come out there and kick your butt!"

Michelle French: How do I communicate to Amazon that just because I bought a map of Chicago 4 years ago before a meeting there does not mean I want every Chicago related item they offer?

Easy. Click the "Improve your recommendations" link under the search bar(assuming you're already on a recommendations page). Additionally, recommended items have a line at bottom telling you what specific other item triggered that being put in front of you, with a link to do something about it. Many places with recommendation systems offer some such function, for obvious reasons; there's no point in antagonizing you with stuff you don't care about.

Su. Thanks for pointing out the refinements to Amazon's recommendation process. This brings up an interesting point as well that certain venues (books, movies, music, etc.) tell us a great deal about who we are and what else we might like. Not only that but I can rate hundreds of movies on Netflix that I haven't rented from them because I've seen them previously. The same is not true for things like electronics or apparel or for sites offering original content. These sites generate recommendations on a very small set of personal information and so the decision-making becomes more suspect.
Dmitri Siegel

A thoughtful and timely piece of writing with many provocative ideas to consider. One item that caught my eye:

"Designers need to be able to advocate against data."

This point gave me pause. I understand the rationale for this line of thinking, but it ultimately seems to suggest a reactionary stance to the idea of metrics and these so-called "Numerati". Was this intended?

Part of your argument seeks to position the designer as a champion of the long-term value of "good design." Naturally. But in doing so, it evokes the traditional "us versus them" model. We are the defenders of the faith. Good design is good business. Etc. But let us not forget that at the other end of all that data are people. Some of them may be "barnacles", but I suspect that most are not.

You propose a critical dialog with the Numerati. Great. In addition, how about a dialog with our customers, users, and soon-to-be collaborators? Peer-production is here, and the speed at which it, and social media in general, is being used and (I would argue) already abused by marketers takes my breath away. There is an opportunity here, and it is being distorted and wasted in the name of business-as-usual.

The mining of data is simply a better way to sell stuff. More stuff. For the most part, this is what companies care about, and why they care about data. This is capitalism. But the use of metrics reinforces the divide between producers and consumers. It presents people (us) as a spreadsheet. Designers are in a position to counter this kind of thinking. Or we could be. This is an opportunity to talk about "good design" in a new way. A notion of design that is less about authorship or "brand building", and more about real people, collaboration and openness.


Hal Siegel

Whoever gets the personal data crunching, customization and delivery working better, could make Netflix work smoother, and make themselves millions of dollars.
Jason Tselentis

doesn't mean a damn thing.

designers design what they like if they are not the dogs for Industry. consumers buy what they want. if they are braindead to watch the chihuahua film, there's your masses for you.

numerati, a bunch of poindexters juggling numbers. you juggle it another way, you change its meaning. lying with statistics, basic statistics 101.

Corbin Johansen


thanks for a great post - numerati is on my (long) list to to-reads. have you seen the 1981 sci-fi thriller "Looker"? it is one of my favorites (albeit very low-budget, the ideas in it are amazing) and whatyou write about makes me feel that what is happening now is a kind of foreshadowing of the visual optimization technology pioneered by marketing research firms in the film's futuristic plot. a brief description from wikipedia, in case you haven't seen it (and you MUST! partridge family's susan dey stars..):

"The Digital Matrix research firm rates advertising models by using a scoring system that measures the combined visual impact of various physical attributes in television commercials. In an experiment to increase their scores, some models are sent to Dr. Larry Roberts, a Beverly Hills plastic surgeon, to get cosmetic surgery in order to maximize their visual impact. After the surgeries are performed, though the models are now physically perfect, they still aren't as effective as desired. So the research firm decides to use a different approach. Each model is offered a contract to have their body scanned digitally to create 3D computer generated models and then animating them for use in commercials."
Gong Szeto

First of all, "Looker" is awesome and totally recommend. It's a great hangover film.

Getting back to Hal's question about the "us versus them" model. I believe that dialectics are part of how our language and system of meaning work and can only be shaped (not avoided). The trick is to cultivate dialectics that are productive and transparent (art vs. commerce, form vs. function, etc.). I think designers should dive in and help complete the data vs.??? dialectic that will be so significant in their work in the coming decades.

I don't mean to imply that designers should always argue against data. But if someone does not form an interesting counter-point to the persuasions of the spread-sheet the conversation will end.

For example, in my own work I find it very useful to remember that all the data you gather only tells you about your existing audience, and may or may not be relevant to the audience you want to have.
Dmitri Siegel

Thanks. That was a very insightful post. I read a few pages from the book " Numerati " . However I dont agree to "Designers need to be able to advocate against data". I think designers also need data. I do not think we should be against it. I think todays world is highly interlinked and this kind of an approach it quiet isolating. I think the marriage of data and design would make a better design.
Nehal Parekh

Dmitri, I've continued to think about your "Data vs. ???" formulation, and finally got around to a response. It's here in full.
Hal Siegel


Fascinating topic. I'm happy to see this kind of discourse happening here, particularly the exchange between Hal and Dmitri. I've posted my own response on my blog as well. read it here.

Jobs | July 25