Evil Conversion: When Optimization Goes Too Far
As online marketers and conversion experts, we want our sites to be effective. But when does conversion optimization go too far? When does it stop being useful and become, well, evil?
As online marketers and conversion experts, we want our sites to be effective. No, it has to be said: we want more than that—we want to turn our sites into kick-butt conversion machines, leaving competitors in the dust with our screamin’ throughput and sales. But when does conversion optimization go too far? When does it stop being useful and become, well, evil?
Most of you know what I mean, because you’ve experienced the evil firsthand. Maybe you’ve even designed it. Experiences like these litter the internet, raising blood pressure around the globe:
- We click on a free offering, only to discover later there was a recurring charge after 14 days.
- We click on a product link, only to find something completely different—and more expensive— pre-selected on the next page.
- We try to purchase something, only to be assaulted with page after confusing page of upsells.
I first started writing about evil usability back in 2007, with a short series of blog posts, and since then the techniques have only gotten… more. More sophisticated, more frequent, more aggressive and above all more optimized. It’s rather annoying, because I know it’s partly my own fault. All my work championing the causes of usability and conversion, and what do I get in return? An increasingly predatory web experience that leverages the very techniques I’ve advocated, but uses them to conceal, mislead and cause me to accidentally purchase crap I don’t want or need. That’s irritating.
A definition of evil
Let me step back for a moment and explain what I really mean by “evil.” I’m not talking about truly malevolent interactions like phishing and malware, that actively aim to harm users. And I don’t even mean pushing a ridiculously misleading product like this one:
Evil in the conversion sense is more a mismatch of goals, along the lines of this hackish definition:
“evil: adj. As used by hackers… evil does not imply incompetence or bad design, but rather a set of goals or design criteria fatally incompatible with the speaker’s (goals or criteria).”
So conversion evil-doing has more to do with business goals, the degree to which they differ from user goals, and how vigorously a company tries to forge an agreement between the two.
Take Facebook, for one very public example—the recent furor over privacy changes on the site illustrates a broad misalignment of goals. The goals of Facebook users are pretty simple—they want to share funny cat photos and play MafiaWars. They wouldn’t have cared much about Facebook’s ambition to be the uber-hub of the web, except that the path to that goal railroaded right through their online privacy.
Evil conversion is more active and granular—it misdirects individual user actions in order to meet a business goal, with or without the user’s understanding.
Let’s take a look at a few examples of evil conversion I’ve run into recently, and examine how they twist conversion principles to the dark side.
Example #1: RealPlayer—Evil button prioritization
Quick disclosure: I despise RealPlayer. I’ve despised it for years. But that’s a long, painful story I won’t go into here.
Last fall, I had to reluctantly re-install it for a client project that included RealPlayer content. I searched for “free RealPlayer”, and below is the page I found. Can you tell me where to download the truly 100% free version of the product?
If you guessed either of the large orange buttons labeled “free download,” you’re wrong! Your reward? A surprise monthly charge after two weeks. Here’s the correct link, up in the far top right (I’ve added the red arrow so you can see it):
This is an evil interpretation of a conversion best practice I’ve often touted:
“Prioritize and differentiate your buttons.”
Yes, it usually helps visitors if your important buttons are prominent; other buttons less so. In this case RealPlayer certainly has a clear grasp on their priorities, but the page design guides visitors to take an action they may not have intended. It provides only the illusion of user choice and control.
Addendum: since last fall, RealPlayer has updated this page to be much more user-friendly. See the improvements here.
Example #2: Kodak—Aggressive, unclear up-selling
Also a few months ago, I succumbed to Kodak’s claim of cheap, planet-saving ink, and purchased one of their newer printers. Their mini-site provided several examples of evil conversion.
To begin with, after adding a printer to my cart I had to wade through three up-sell pages before reaching the “real” checkout page. These included an especially creative offer of a second printer identical to the one I was purchasing:
Now these pages weren’t necessarily evil in and of themselves. Offering related items can be valuable and helpful for the visitor. But on each step the path to avoid purchasing more things was visually de-emphasized, hardly even looking like a button at all:
While less evil, in my view, than the RealPlayer example, the result is still likely to be user confusion and a number of unintentional purchases. Again, the path is being tipped to benefit the company, not necessarily to aid the visitor.
Example #3: Uhaul—Forcing opt-outs to avoid a purchase
My colleague Lance Loveday first mentioned this example to me over lunch, complaining of the “terrible” rental experience he’d had on Uhaul.com. And though it turned out he was more offended by it than I was, I had to agree with him on principle.
This site takes aggressive up-selling, adds evil button prioritization, and mixes in forced opt-outs. Below is one of several participating pages:
If you’re not careful here, you’ll mistakenly rent the pre-selected dolly and furniture pads, adding $17.00 to your total.
Not so evil, you say? The next page was even worse, pre-selecting nearly $50.00 worth of rope, gloves, and tie-downs:
This is a twist of the conversion principle “Make It Easy To Do”. Oh, yeah, it’s easy to purchase all these additional items all right – but the optimized action, again, is more beneficial to the business than to the visitor. On both pages I managed to avoid the large, conversion-optimized buttons and find the top-right “I DO NOT need these items” text link fairly quickly, but I’d been pre-warned by Lance. I had to wonder: how would an average Internet user fare?
The real question: How far should conversion optimization go?
Probably the real irony in this is that I often find myself learning as much from evil conversion as I do from “white hat” resources (to use an SEO term). As a conversion practitioner, I can admire an especially well-done page layout and notice why it’s so effective, even if its goal misaligns with my own. I can even wonder whether a similar approach might benefit a client, and how to tweak it to match a different business model. I can feel the pull of the dark side.
Two things counter and balance the attraction:
First, evil conversion techniques have a short-term viewpoint and short-term payback. The underlying mindset is “It’s OK to manipulate the user, because we don’t need her after this transaction.” Companies that I prefer to work with don’t share this opinion. They understand the value of trust, and long-term customers, and know it’s not in their best business interest to earn their visitors’ hatred.
Second, as a user, I can remember the frustration, irritation, and yes, disgust that such techniques cause me. It makes me a better conversion specialist, one who can (usually) see the trick, understand the human impact, and bring that insight into my reviews, teaching, and designs.
Where will evil conversion lead? Probably to a continued escalation of coercive techniques, until poor business results, lawsuits or regulation force a retreat. Until then, we’ll each have to balance business goals with those of users, and decide what the right optimization path is. Those of us who pride ourselves as being true user advocates have an interesting time ahead.
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.