Warning: file_get_contents(http://www.linkedin.com/countserv/count/share?url=https://freshpractices.com/recommendations-engine-uxp/&format=json): failed to open stream: HTTP request failed! HTTP/1.1 404 Not Found in /home/customer/www/freshpractices.com/public_html/wp-content/plugins/tk-social-share/tk-social-counter.php on line 145
Note to Readers: This post was written circa 2008-2009, but is reposted here as I have seen few real improvements since that time. If you are a user experience professional you need to think about how suggestions and recommendations impact your users in email, on-site, and even via retargeting ads on sites they visit.
A dangerous trend is making users of ecommerce sites and software unwitting victims in the quest for more revenue. Though money certainly does make the world go ‘round, and the strong flow of money aids the greater good, the latest features being demanded by marketing departments (not necessarily users) are “Recommenders” designed to keep visitors on a site, spending their hard-earned dollars.
User experience evangelists need to stand up for the users of their sites and software, and make sure this feature does not negatively impact the user experience. These tools may not only destroy a positive brand and user experience, they could ultimately result in the loss of customers if they aren’t implemented with integrity and an awareness of user perception when faced with a barrage of recommendations meant “just for them.”
Here are some of the things leading recommendation software providers say they can do for you:
• Rich Relevance will “make shopping experiences personal”
• Avail Intelligence promises to “inject the collective intelligence into every customer interaction”
• Aggregate Knowledge will “help people discover things they didn’t know they were looking for”
One of the founders of Aggregate Knowledge assured me they talk to users and use analyst research to help determine what users want regarding recommendations, and I’m sure all of these companies do. There are studies that demonstrate recommendations increase sales, but I’d like to see hard data on how they help the customer, in addition to data that reflects they know exactly what users do not want as well. Are all of the predictions and guesstimations based on The Truth about you, me, us? Because You are unique, unpredictable and only marginally stereotypical!
Amazon’s Recommendations Are Hit & Miss
In my own experience as an online shopper, I have rarely gained from something being sent to me, that I did not request, based on my previous purchases, or purchases of people like me, or in my state, area code, etc. as these unfortunate samples from Amazon will show. However, when I can go in and essentially search for things, I find lots of stuff I want – my Wish List there is chock full of more books than I can buy. This post mentions Amazon specifically, but really this applies to the majority of online retailers – they are all guilty of the same issues if they send you suggestions and recommendations… or follow you around the web via retargeting ads.
A few years ago, Amazon sent me two recommendations in email, comical in their inaccuracy, but still annoying. The first one was harmless enough… it is an email suggesting I buy the Zagat ToGo Pack Chicago 2008/09. The problem is, I don’t live in Chicago, never have, don’t recall the author Laren Stover, and could actually care less that Zagat is releasing something to do with eating in Chicago. And yes, I can always “opt-out” by clicking a link and doing something, but who opted me in to this farce? When thinking about the information architecture, and database collection, and marketing emails, and the many technical details that go into a recommendations campaign, did Amazon do some user-focused scenarios so that they could avoid boneheaded mistakes like this one? They are the ones who look uninformed and greedy, not me. I just collected yet another piece of spam to deal with, from a company that I have shared a lot of money with already.
The second email could be dangerous. If this came to a shared email account of a husband-wife, Lucy might have some serious ‘splainin’ to do about the recommendations for divorce books. Over 5 years ago, I dated an old flame who was purportedly getting a divorce, and purchased the book “How to Survive Your Boyfriend’s Divorce”. The relationship did not survive, and I’m not sure if the divorce ever happened or not, but apparently my need for divorce information has lived on in Amazon’s database. However, that was a long time ago, and I am married to someone else now. The correct recommendation, if you want to get technical, should involve “my boyfriend’s divorce” not “MY divorce.” Semantics are apparently not factored into the intelligence gathered.
These are examples of email recommendations, but going to the Amazon web site is also an adventure in sales suggestions. Today, miraculously, they mostly got it right, except for the big push for the Kindle reader that I have no interest in. The books on the personalized home page are centered on user experience, design, development, start-up businesses, etc. and of course, include some I already own, and some I don’t. But there was a day when I was working with some software developers and managers, in 2001, when we wondered if there were a particular book on a problem we were dealing with, and in a meeting I pulled up my Amazon site, to have books on gays and lesbians prominently recommended to me. I had purchased one book for a gay friend’s birthday, and apparently Amazon believed I was gay and wanted more like it. I quickly shut the site down, because I felt a little mortified at the sexually-oriented books being displayed to my business colleagues. And to this day, I still get tons of golf book suggestions from one purchase of a golf book for Father’s Day many years ago.
These are examples of how recommendations made that are unsolicited by the user, often get it completely wrong. It is a waste of the users time, creates bad will, and can result, believe it or not, in users looking for someone else to purchase items from that won’t submit them to the annoyance of the “gimme more” attitude that is so prevalent, and becoming more so all the time, online. And yes, I can change my recommendations and email preferences and everything at Amazon, after they have already set stuff up. It takes a fair amount of time, and not something I should have to do just to get out of the loop of lame sales attempts.
Great Customer Experiences Aren’t Hard: Keep the User the Focus
A better way, would have been to set up the site, databases and marketing based on things that I chose (not previously purchased), and allowed me the complete flexibility and freedom to personalize this at any time, with simple tags or keywords, not having to go through product by product and eliminate options cluttering my environment. That would provide value, and get me excited, as a customer. Because I do use some of Amazon’s recommendations, but when I choose, based on my interests at the time, not on their “predictions” based on a purchase that may or may not have been a random one time purchase (stepfather’s golf book) or fad I briefly entertained (sixties decor), not on my lasting interests.
I have a background in retail and visual merchandising, which is why I feel so passionate about creating great customer experiences, online and offline, and there are ways of doing business in the real world, that online sellers need to learn. “Upselling” is one method of offering a suggestion to a receptive customer, with the appropriate timing so that it is not construed as intrusive or out of place.
For example, a man is trying on a dress shirt and pair of slacks in a department store. The attentive (but not cloying!) salesperson will inquire, when he says he’d like to take them, about whether he needs a tie or not. If the man wants to see the ties, she will ask about color and pattern preferences, and because his shirt is off-white and his pants are black, won’t go to ties that go with brown and white, but will show only those that reflect the tastes of the customer and coordinate with the clothing. Then she might ask if he has dress shoes and a belt that will work with the slacks, to make sure HIS needs are met, not just to sell more stuff to a guy with a credit card. The difference can only be understood by people who want to do business in an ethical, moral, and mutually beneficial way.
This same experience can be created online, but it takes hard work, savvy programming, and dedication to the user experience by a team of people who care about their customers. iGoDigital has some examples of Shopping Assistants that I guess are better than nothing, but they lack a truly user-centric personalization that takes into account people’s unique personal tastes.
An example is the Auto GPS Product Advisor they did for Walmart. It offers a few questions, some of which I really don’t even know the answer to, like the size of unit I want or the brand. What I really want to know, being somewhat of a girly girl, is do they have one in pink? Because my best friend told me her husband bought her a pink one, and I don’t remember the brand and don’t know the cost, but I would love one in pink! Where is the keyword field, where customers can enter some of their own preferences? Where is the human element of this “product advisor?” Make recommenders smarter so that I can choose more attributes than a dropdown list or checkboxes can provide, and available at my command, rather than shoving products in front of me, and I’m sure they would be a lot more helpful, and drive even more sales.
Recommendations Are the New Personalization, With Consequences
Perhaps technology isn’t ready, or marketing professionals don’t know how to ask the right questions, but it is up to user experience managers, specialists and evangelists, to push for these types of ecommerce features to be implemented in ways that benefit, and do not hinder or destroy, the awesome user, customer and brand experiences we are working so hard to design, so that customers will desire what we are selling and not be repulsed by our efforts.
I work in both the marketing and the design realms of ecommerce and software creation, so I know there are better ways of doing business than what I am experiencing as a customer and watching companies do with my information online. This is a trend that must be altered, because it’s already ugly and headed in the wrong direction. With the economy suffering it’s only going to get worse, if we don’t make marketing professionals and companies we do business with, take a better, more user-focused direction.
What’s worse than recommendations based on what you purchased from a company you chose to do business with, are suggestions from people who know what you bought, that want you to buy more stuff. Given the ease with which online consumers give companies access to personal information (through mail, email, and phone numbers), the burdens of spam threaten to overtake us as the greediest of companies fight for our dollars. As I type this, someone somewhere is trying to figure out what else they can sell you if you own a hybrid SUV, an iPhone and pay for Showtime channels. Will they stereotype your personality correctly, or will you just receive a plethora of unwanted solicitations? Companies think their opportunities are limitless, if they can just crack the code to figure out what makes you spend your money. But are these opportunities to make more money, or to corrode customer goodwill?
I found a great post by Joshua Porter called Which Movie to Watch? An Overview of Recommendation Systems that you might also want to read.
Silly Waste of Time
I was delivered a great example of how pointless recommendations can be. Because of job hunting I was shown this helpful offer on Linked In. The problem is I am not a barista or coffee-related worker, don’t speak another language and so am probably limited to US-based jobs at this time. But thanks for the suggestion!
Do you agree? Are you concerned? I want to know what you think, as a user, about recommendations in your ecommerce experiences and whether or not suggestions benefit you, or get in your way as you navigate the online space.