Reading Time: 10 minutes

Anti-patterns can be hard to spot. Anti-patterns are the inevitable outcome when a rule set is applied so rigidly that it yields the opposite of the original desired outcome.

User experience is no exception to anti-patterns. As UX principles and practices become more commonplace, enterprises are finding themselves faced with an increasing number of failures in the nooks and crannies of the experiences they are crafting.

latest report
Learn why we are the Leaders in API management and iPaaS

Interdisciplinary teams everywhere are sent into agile design meetings and asked to make decisions that balance all concerns. Agile — by its nature — asks teams to be swift and autonomous. This can cause teams to move counter to the intention of the un-represented specialty disciplines that are critical to operating at scale, like organic search engine optimization (SEO), performance, security, and operations.

It’s almost impossible to prevent teams from perceiving their counterparts as the peanut gallery and it ultimately falls to management to try and broker compromise while keeping everyone’s goals intact. Reaching a decision on these instances can be excruciating, because they exist where multiple philosophies collide and the solution will leave no one happy.

Anti-Pattern #1: The eyes have it

Audiences are the fertile ground from which so many dot coms sprout and thrive. These audiences demand intuitive experiences tailored to them and their tastes. But what happens when the experience design comes at the expense of drawing the audience?

It sounds odd, but there are times when SEO best practices collide with optimal experience design.

For example, let’s say you were running a site that sold cat posters. SEO best practice has the main titles (H1 headings) on a landing page align verbatim with the most popular search term: e.g. “Awesome cat poster.” The intended destination for your cat poster selling site has a search results page offering many awesome cat posters (although aren’t all cat posters awesome by definition?).

A page designer would naturally push the grammatically driven copywriters to label the h1 for the page as “awesome cat posters” in an attempt to have the heading maximize its alignment with the page contents. Given that SEO awards extra points for verbatim matches, details this small (making “poster” plural by adding an “s”) actually become relevant in unique sessions and page impressions for websites in heated competition for audience share. It can be the difference between being the top awesome cat poster search result and the second. With enough competition, differences like these can pull your site above the fold of the Google search results page.

This problem becomes even more thorny when you consider:

  • The difficulties in a/b testing SEO changes in a seasonal business
  • The difficulties in determining or predicting the precise impact of a singular SEO change
  • The problems that emerge with your design team if you successfully proved that very few people read headlines on search pages anyway

An interesting approach that might help bridge these two disciplines is the trend of increased irreverence from traditionally staid sources. Whether it’s Twitter’s fail whale, Google’s “aw snap!” or even the presidential campaigns 404 Easter eggs, everyone seems to be jumping onto the “web copy doesn’t need to be stuffy to be trust inspiring” bandwagon.

Anti-Pattern #2: Flaw in the honeycomb

Jakob Neilsen’s UseIt site used to be a real conversation starter in the UX space. An example of extreme minimalism on the web, many UX practitioners would deride the site, and Nielsen as well, as someone who understood usability but not UX because he paid no attention to aesthetic design or desirability (an often overlooked component of UX). Most of these discussion’s would end up pointing to Peter Morville’s UX Honeycomb diagram which spelled out a multi-faceted view of UX.

As the web has grown richer in content and capabilities, a new, previously overlooked hexagon needs to be added to Morville’s original work: Performant.

One of the original premises for UX practitioners was that experience design needs to come first and be treated separately from the technology that underpins it. This premise is designed to prevent design from being overly compromised either by existing implementations or by the perceived effort required to achieve the desired experience.

Once again, the original premise can steer well-intentioned professionals into strange territory, where efforts to design optimal experiences apart from technology concerns collide with the effort to create optimal experiences.

For example, a shop that regularly separates experience design from engineering is — by design — removing experience designers from creating the hyper-performant experiences that are inextricably linked to both the first and the lasting experience impressions.

The web performance community has begun to work on this problem in multiple ways. Amongst the most compelling is the work of Lara Hogan, engineering manager at Etsy. Hogan has spoken and published on her works to create a “Design for Performance” movement. While her talks speak to improving performance technically, she devotes much of her speaking content to teaching the craft to both engineers and designers, with an eye towards creating a sustainable culture that values performance as something that all disciplines contribute to.

Playing by the business rules

Who would have thought that the profit motive would provide the solution to design and culture problems? Talking about these issues in a business context can take the emotionally charged positions out of the conversation and get teams centered on designing for business performance. UX-centric thinking, like any other centric-oriented movement, will always lead to anti-patterns when there is no room for dialogue or contribution from other disciplines. The art is in knowing how to smell the arrival of anti-patterns and to invite other disciplines in for collaborative solutions.

This article first appeared on CMSWire.