Archive for August, 2010

Better Ecommerce Search for All Online Stores

Thursday, August 26th, 2010

Ecommerce has been one of the earliest identified benefits of the Internet along with email, web pages, and few other early features that engendered large investments in the late 90s. Tobi Lütke, CEO of the now famous ecommerce platform Shopify, puts it this way:

…my theory of what happened in the ecommerce industry is that Netscape is to blame. (…)

When Netscape filed for the first big Internet IPO (mid-1990s) they needed to convince Wall Street and the rest of the world that the Internet would be successful. (…) People understood commerce, everyone sees it around them every day, so Netscape convinced the world that the Internet would be big because of ecommerce. They did such a good job of convincing everyone that a lot of companies were created to supply the software for this imminent gold rush.

This meant that most software in the ecommerce space was written in the 90s, long before we figured out how to make a good web application. (…)

The net result is that you have two things: a lot of merchants with post traumatic ecommerce stress syndrome, and a lot of software that is stagnant because of lack of growth. This nuclear winter existed until well into the last decade.

(source: http://37signals.com/svn/posts/2378-qa-with-tobias-ltke-of-shopify)

Toby might be right or wrong but a quick look around the online retail industry in search of great search software does lead one to wonder how neglected this key feature of online stores and shopping platforms seems to be. Amazon and eBay had decent search infrastructures and UIs but as one who worked with some of these large ecommerce organizations to better their search, I can tell you two things:

  • I can count on one hand the companies that can afford the massive investments they have chosen to make in this area
  • The results are not necessarily impressive. In many cases the search experience is not optimized, leading to dead ends, zero results searches, useless suggestions, and rarely capable of locating rare products without some unusual persistence.

That’s one of the several reasons why we are so excited about Exorbyte Commerce Search and Autocomplete (if you are in the UK: http://commerce.exorbyte.co.uk):

  • There are not many choices for advanced product catalog search features like what we offer with our SaaS ecommerce search add-ons for existing online stores.
  • It works with many ecommerce platforms (hosted or installed).
  • It’s affordable because we mutualized the resources across many customers.
  • It’s installed in minutes.
  • It comes with amazing performance: searching millions of products with sub-10 milliseconds response times.
  • Error-tolerance: ability to automatically correct spelling mistakes and entry errors, in any language, based on on actual products available in your catalog.
  • Autocomplete feature: allows to install on top of your existing search engine and helps people find what they are searching for.
  • Advanced reporting allows you to see what people are looking for, what they find, whith what keywords, what’s missing from your stores, etc.
incremental search 300x237 Better Ecommerce Search for All Online Stores

incremental search and autocomplete for ecommerce

It’s simple to understand the business goal here: visitors who find products are more likely to buy them, thus enhancing conversion rates and growing revenues. Additionally, it’s easy to understand how, with tens of thousands of small and medium online stores in the US, there is a huge market for these improved search features. Users have been educated by Google to be used to great search software and they now expect it everywhere. So, boost your store’s revenue with Exorbyte Commerce Search and Autocomplete (if you are in the UK: http://commerce.exorbyte.co.uk).

Low Hanging Fruit for Most Online Stores: Better Search Features

Wednesday, August 25th, 2010

We just launched the public beta version of Exorbyte Commerce Search and Autocomplete (if you are in the UK: http://commerce.exorbyte.co.uk). It’s a hosted SaaS monthly subscription version of our popular search products (MatchMaker and SearchNavigator) especially designed for ecommerce.

We also serve a large selection of other industries with the same core products for search within structured data. However we chose ecommerce, and we did so on purpose. There is plenty of data showing how many online stores are behind in terms of satisfying users with better search features to help them find what they want to buy. Here are  some of the reasons why Exorbyte Commerce is worth many times the very short time  (only minutes) setting up its free trial requires:

  • More than 47% of online shop visitors don’t become buyers because of poor search or navigation features.
    “According to iPerceptions’ E-commerce Industry Report Q2 2009, 38.6% of the 360,000 or so visitors to the 160 websites tracked were in the research stage of the buying process, while just 17% were at the buying stage.

    online shopper flight1 300x177 Low Hanging Fruit for Most Online Stores: Better Search Features
    source

    Source: iPerceptions' E-commerce Industry Report Q2 2009

    iPerceptions was able to establish several reasons why visitors aren’t purchasing. The main reason was that visitors weren’t able to find what they were looking for (34%), while price and navigation/usability issues tied in second place (13%) followed by shipping policies (9%).

    why customers dont buy1 300x200 Low Hanging Fruit for Most Online Stores: Better Search Features
    source

    Source: iPerceptions' E-commerce Industry Report Q2 2009

  • In his book The Humane Interface, renowned user interface guru Jef Raskin compares incremental and delimited search: “With a delimited search, the computer waits for the user to type a pattern and delimit it, after which it is the user who waits while the computer does the search. When using a delimited search the user must guess, beforehand, how much of a pattern the computer needs to distinguish the desired target from other, similar targets. With an incremental search, he can tell when he has typed enough to disambiguate the desired instance, because the target appeared on the display.”
    Jef Raskin goes on to say: “In spite of near agreement about the desirability of incremental searches on the part of both designers and users, almost all interface-building tools make it easy to implement delimited searches and difficult or impossible to implement incremental searches.”
    Jef Raskin even ventured to say in a footnote that search is either “incremental or excremental”. We kind of agree with him. icon wink Low Hanging Fruit for Most Online Stores: Better Search Features

While ecommerce catalogs are structured database tables, many online stores still use lame full-text search engines or slow database queries to provide catalog search results to their users. This is a real problem for the following reasons:

  • Full-text search engines usually have to crawl pages to index the data, making the update mechanism of search results slow and often out-of-date (not reflecting frequent changes in inventory levels, prices, and descriptions retail database undergo daily).
  • Database queries and full-text queries are simply not capable of handling advanced fuzzy search like structured data or database indexing engines can. Too slow or simply impossible to implement.
    Just try to build an advanced multi-field fuzzy search facility for a database of millions of SKUs (ex: online travel store, industrial parts, online electronics store, etc.) and keep it fast (under 10ms round trip) using a database query or full text search engine! Have fun and let us know if you want tips on how to make it work with a different approach.

Something that would be much better, much simpler, much faster and much more natural. An ecommerce search system that yeilds more conversions. Simply a system that guesses what you want to find even if you don’t know what it’s called: That is what we strive for every day at Exorbyte! What do you think?

Exorbyte Technology and Features Detailed, but Never Enough for Some

Monday, August 23rd, 2010

We are often asked by online requests from prospects and from potential users to give them extensive details on our data matching technology.  That’s understandable.  A few of the claims Exorbyte makes regarding the speed and error-tolerance of our software are pretty unique and attract interest by users and developers alike.  This said, it’s often with regret that we give partial information but cannot divulge the more proprietary parts of our system’s capabilities.

I would like to set the record straight.  Like all software companies, we have to retain a degree of privacy around the most valuable of our inventions.  However, we often see developers come to us for information and then attempt to build an application rivaling MatchMaker in speed and error tolerance.  The results are almost always disappointing.  The reason is simple:  there is no silver bullet.  

There is no single invention or trick to how MatchMaker can return Levenshtein or Soundex query results on many millions of data records in under 10 ms.  The performance our software features comes from a large series of architectural, indexing and algorithmic innovations.  Furthermore, every build of our core MatchMaker platform is tested for speed and overall performance at night (after our developers go home) to maintain optimal performance no matter what fancy features we add to the system.  This has been going on for 10 years now, so the quality, speed, depth of functionality, flexibility, and reliability of MatchMaker are no miracle.  We work at it every day.

We just released a new portion of our web site which describes there characteristics in more detail called Exorbyte Structured Data Search and Data Matching Platform.  Check it out and tell us what’s missing if anything.