In this third installment of the Bazaarblog partner interview series, Matt Eichner, VP Marketing & Strategic Development at Endeca, shares his views on where the field of information access is heading, specifically within the e-commerce domain.  Endeca and Bazaarvoice share many customers in common and have partnered to bring social navigation functionality to market through out-of-the-box integration, so we were eager to pick Matt's brain about what the future has in store for Endeca. 

1.  Google’s mission is to organize the world’s information.  As a B2B company that provides access to information, what is Endeca’s mission? 

Endeca’s mission is to inform – and influence – daily decision making.  This manifests itself in many ways across many different industries.  But the idea is that businesses are sitting on a wealth of information in a wide variety of formats that lives in different places.  If you can give people the ability to better explore, analyze and understand this information in a way that helps them make better decisions, the aggregate positive economic value of all these better decisions is mind-boggling. 

Take e-commerce as an example.  A consumer shows up on your website looking for a great digital camera for her 65 year old mother – they are at some stage of the buying decision process.  They aren’t just looking for a list of all the cameras you have.   They want information about those options that will help them figure out which one is the perfect match.  They may care about common things like brand, price or resolution, but they also may care about what features matter most to less tech savvy people like their mom.  This info may come from product catalogs, buyers guides and customer reviews.  The challenge is pulling it all together in a way that supports this unique buying decision…and making sure you do the same for all users: The amateur photographer…or the person looking for the best kids games for the Wii…or the shopper who literally has no idea what to get her 10 year old nephew for Christmas. Influencing a unique decision is worth 10’s or hundreds of dollars.  Influencing all of the unique decisions is worth 10s or even hundreds of millions of dollars when you think about the aggregate volume of commerce traffic.   

2.  I associate Endeca with search and “Guided Navigation”, but “discovery” has become a popular term in our industry over the last year.  Endeca released its Discovery Suite very recently, which includes a social navigation module.  How do you differentiate “discovery” from “guided search and navigation”? 

Search can be an effective tool for fact finding, if you know what you are looking for and how to ask for it.  Want to know how much a Wii costs…and if they actually have any in stock?  No problem, type “Wii” in the search box and voila.  But search is a poor tool for discovery, where you may not know exactly what is available and can’t ask for it precisely. Stores like have millions of SKUs.  How could anyone possibly know even a fraction of what might be available in a catalog that size? Shopping for your nephew and you want to know what are the hot toys this holiday season for 10 year old boys?  How would you describe that question?  Search isn’t going to help. 

Endeca can help with the fact finding questions – its search filter certainly includes the full set of popular search features – but what Endeca is really designed to do is support investigations…to guide people through the process of discovery.  Guided Navigation is a part of that.  Think about a hotel concierge.  You tell them you are looking for a good place to eat – the equivalent of typing “restaurant” in a search box.  Rather than just listing out every possible restaurant in town, they might ask whether it’s a special occasion, if you prefer certain types of cuisine, how much you want to spend – next step questions that are the equivalent of Guided Navigation. But more than just ask these questions, the concierge would suggest ideas, and expose tradeoffs.  Want that French restaurant?  The price just went up to between $30 to $80 a plate, unless you mean French Fries.  The concierge might also ask whether you want something close by or perhaps in a certain area of town, the equivalent of visualizing results on a map rather than in a list.

The Endeca Discovery Suite, taking the analogy a step further, is like arming the concierge with even more tools, like a Zagat Guide, so now, they can tell you things like these restaurants are really popular with couples, or the restaurants that had great food but where service was suspect.  The idea is simply that we want to give regular people more and more ways to express what’s most relevant without putting the onus on the person to ask the perfect question… or “the concierge” to be a mind reader.

3.  Many of our shared retail customers now offer the ability to filter or sort search results by customer rating and have seen significant conversion improvements as a result. Why do you think user-generated content is such a valuable attribute in search and navigation?

User reviews have long been a consumer favorite.  What better evidence is there of how something will work than someone else who is just like you enjoying that product (or not!)?.  The meteoric rise of companies like Bazaarvoice has really changed the game here – our ability to work with this type of information has quickly gone from nice to have to must have.

Information access techniques like Guided Navigation really unlocked a new way of looking at reviews; not by individual products but by collections of products; building community and driving business results by letting users advise each other through media like reviews.  Even more important, they open the possibility of new types of cross-product navigation, integrated with the other criteria buyers care about.  Interested in a gift for that young nephew?  Well, how about, “what are the favorite products in the $30 – $50 range among 15 year olds?”.  This takes the review content into a new context, combining it with product attributes to open up the kinds of advice that used to be only available when talking to a store’s best sales reps.  Best part – no need to hustle for their attention during the busy holiday buying season.

And the early success stories – the early techniques used to combine these features – are just the tip of the iceberg.  Over the next 12 months we’re going to see marquee sites introducing information access capabilities that will forever change the way reviews are used in online retail.

4.  What makes user-generated content such an interesting search/navigation challenge for Endeca to focus on?

Our ability to integrate and expose this type of information is critical to our success and that of our retail clients. Making that happen introduces some rather interesting user experience and technology challenges, and opens the door to many new ways to be creative in site design.

First, you have the problem of introducing new attributes into the environment.  As far as the user experience is concerned, we have to expose this combined information in a way that facilitates discovery, gives shoppers a myriad of ways to explore desirable products by characteristics most relevant to them.  And it needs to be intuitive. People need to be able to take advantage of this info using familiar techniques – you don’t have time to train your shoppers. It has to be done in a way than any shopper can take advantage of right out of the gate.  And, it can’t introduce any fragility into the systems IT need to maintain.  It needs to scale, and also embrace new scope as new elements of reviews get added over time.  For instance, we see sites expanding demographic information on reviewers to include other elements like clickstream analysis.

But including user reviews data goes beyond attribute navigation.  One key to making this information useful is taking advantage of information about the users reviewing the products in addition to the products themselves.  There is a subtle user experience issue here with dramatic technology implications.  Navigating by the characteristics of the review – age, gender, etc. – is not enough, alone.  It is important also to see the products these reviewers assessed, and also within that subset navigate by the characteristics of those products.  In this fashion, a user can ask those useful questions like “I’d like to understand the best inexpensive stroller options that were preferred by mothers like me.”

Yet another challenge is around data irregularity.  Product catalogs are very structured, organized in a more classic table format — a product has associated attributes like price, brand, size, color etc.  Reviews take many forms from numeric ratings to free form text to user-generated tags.  Unlike a product catalog, this kind of irregularity makes it virtually impossible to shove all of this info into a relational database.  You need to create a very flexible data model to bring these worlds together while retaining the context between related information.

5.  Several Internet search engines are integrating user feedback tools into results pages to ultimately make those results more relevant for the community as a whole.  Can you share any insight into Endeca’s roadmap in this area?

It’s an interesting concept and something we’ve watched for a while.  The challenge is two people who enter the same exact query may be looking for completely different things.  Likewise, a result that is perfect for one person may not be as relevant for someone else.  For instance, two people looking for digital cameras.  One may prefer portability for fun shots, another may be shopping for the ideal camera to capture some nature shots for an art class.  Or, imagine how even a single person may alternate between shopping for themselves and gifts for others.

A different community approach, and one that you’ll see in the not-too-distant future is something more similar to the tag-based approach popularized by sites like Flickr.  In this approach, the user will have an ability to add descriptive tags to either individual results or even aggregate lists of results.  These tags will instantly be added to the user experience and can then be used exclusively as personal tags for later access or can be shared with all users as another descriptor of that information. 

6.  You work closely with Endeca retail customers every day.  What are the biggest search/navigation problems these companies are trying to solve, now that solutions like Endeca’s are fairly ubiquitous?

This really is a fun market to be a part of right now.  Retailers have always been on the cutting edge of leveraging innovative technology to provide a competitive advantage.  There are quite a few advanced capabilities that retailers are looking at incorporating in to their site right now.  The challenge, and what is going to separate the leaders from the followers, is their ability to integrate these different capabilities into a cohesive and compelling user experience.

Beyond the social media challenges we’ve covered, other areas of interest and investment include multi-channel commerce, bringing in-store inventory and localized pricing into the online shopping experience.  The Home Depot recently took this challenge head-on and you can now use Endeca to browse both online and in-store inventory, unique to which ever store you select, side by side.  We expect to see a lot of retailers heading in this direction.

Another key area is mobile commerce, where there are unique challenges in replicating an experienced optimized for a full-sized computer screen into a handheld device.  Innovation like that seen in the iPhone will change the game forever, and we’re already working with companies to take advantage of this new class of interaction and user experience.  We’ve even gone as far as to incorporate the local, in-store inventory and social navigation – bringing user reviews into the Endeca experience – into our first prototypes.   

One Response to “Partner Interview: Matt Eichner, VP Marketing & Strategic Development, Endeca”

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>