I’m fascinated by companies who do the simple things well.  And right at the top of that list is Zappos.com.  It’s everything from how they talk about delivering first rate customer experiences to critical deviations to a massively successful e-Commerce website.  So when I came across an eMarketer.com interview titled “Zappos talks about personalization“ (with Brian Kalma,  Director of User Experience at Zappos.com), I had to read it.

Kalma talks about personalization and its relationship to recommendations and although it’s nothing groundbreaking (Kalma doesn’t reveal ancient eCommerce secrets that’ll boost your conversion rates), he makes a great point about historical recommendations and how they may not jive with new sessions.

Full disclosure: I’m a bigger believer in session based personalization than recommendations.

To Kalma’s point - “It gets tricky when shopping for other people”.  Let’s take it a step or two further: last week you browse women’s dresses and skirts for your wife (at her request).  You look at a few different styles, brands, etc… but leave having not converted.  You come back to the site a week or two later but this time, you are looking for men’s dress shirts for you.  Based on previous browsing history, you could very well be presented with “targeted recommendations” of dresses and skirts.  Some may might find that offensive and most would find it irrelevant.

Another example: if you are anything like me, it’s an enormously difficult task keeping surprises/gifts from loved ones.  It’s a curse, I swear (I’m open to suggestions, too).  Imagine you’ve gone to such great lengths to hide ideas and deny all  guesses, only to have your super secret gift spoiled when your loved one uses the shared computer.  They reach one the sites you recently browsed in search of this gift and the recommendations are all in similar/hinting themes.  Perhaps your loved one isn’t that savy, but that’ll have you sweating bullets.  Or maybe the messaging is that targeted … “Did you still want to buy that engagement ring you were looking at?” … Yikes.  That could be bad.

In more extreme examples like those above, historical recommendations can be hit or miss.  Session based personalization is more reliable, as its all based on real-time actions of visitors.  And it never spoils gifts, either.

All this said, one thing really stands out.  Brian Kalma rolled recommendations and personalization up as one-in-the-same.  Interesting, but I see recommendations and personalization as two completely different entities.  Personalization can mean branding, messaging, calls to action and more via behavioral targeting, referring domains/URLs, unique/specific segmentation, etc…  Simply put, its not limited to relevant up-sell/cross-sell opportunities.  It’s about a session wide dialogue to heighten audience engagement and experience whereas recommendations are driving additional sales or conversions.

It’s well established visitors/customers don’t respond well to constant ‘buy this-buy that’ messaging.  That wears out quick.  The online marketing community has done a good job understanding that customers want an informative relationship that doesn’t push for sales on every tweet, message or email blast.  Under that logic, are product recommendations that different?  Don’t we run a similar risk of burning out visitors by only offering product recommendations?  Clearly it’s a valuable tool or tactic that has its place, however, it’s not the only weapon in your eCommerce arsenal.

You know what side I lean on.  Question is, where do you stand? Recommendations or real-time personalization - which do you see more long term value in?