Posts Tagged ‘search’
Monday, November 28th, 2011
Every economic cycle comes with its host of enterprise software trends. Big Data hs become a recognized phenomenon in 2011. In May 2011 McKinsey released the “Big data: The next frontier for innovation, competition, and productivity” report. It started with: ”The amount of data in our world has been exploding and analyzing large data sets—so-called big data—will become a key basis of competition, underpinning new waves of productivity growth, innovation, and consumer surplus”.
IBM, Oracle, SAP, Microsoft, SalesForce.com, and others are all aiming their development efforts at Big Data (see vldb.org). The amount of data produced, collected and stored by online activities to which companies, their customers, their partners, and their sales channels participate has grown enormously. Tools are being developed that allow affordable long-term storage. New columnar in-memory database formats have emerged that enable near-real-time analytics. Fast growing stratups and open source solutions have also converged with their own new NoSQL formats (InfiniDB, LucidDB, InfoBright, Hadoop, NoSQL, etc.) MatchMaker, Exoryte’s Universal Search platform, is the perfect answer to search within Big Data.
The challenges of search within big data are:
- Searching Big Data though SQL queries is simply too slow and inflexible – fuzzy or advanced search requires a search indexer layer or something different than traditional on-disk relational DB formats.
- Indexing large databases can be long, disruptive to normal database operation and require complex hardware infrastructures.
- Running complex queries and fuzzy logic requires so much calculation and lookups that new search strategies are required.
Exorbyte MatchMaker is made to address these challenges and our professional services team has proven repeatedly tht they can be addressed: Allianz (the world’s 12th-largest financial services group), German Finance Ministry, and more blue chip and government organizations tun o us each year for that very expertise.
What do you think of Big Data?
Tags: big data, database, enterprise, erp, hadoop, ibm, oracle, salesforce, salesforce.com, sap, search, search engine, search engines, software
Posted in Exorbyte, Government, Insurance, Search, Software, Technology, databases | 1 Comment »
Friday, September 30th, 2011
Oracle Larry Ellison has been attacking the rationale of HP buying Autonomy for $10B. One wonders if he wasn’t eyeing Autonomy himself for an acquisition before HP snagged it. This said, now that Fast, Convera, Autonomy, Netrics, have all been acquired, we are wondering if Endeca might not be the next one up for an Oracle acquisition this time?
See more here:
http://www.businessinsider.com/oracle-hp-autonomy-2011-9
http://www.marketwatch.com/story/oracle-fuels-fire-over-h-ps-autonomy-deal-2011-09-29
http://articles.businessinsider.com/2011-09-29/tech/30216980_1_oracle-qatalyst-slides
==== 10/18/2011 – Update ====
http://www.forbes.com/sites/ericsavitz/2011/10/18/oracle-buys-endeca-targeting-unstructured-data/ That’s it. Oracle announces they are buying Endeca. This is big news. Another confirmation for us that Search is a hot category for enterprise software but also that deeper integration of search is in the works. While this acquisition of may be a good thing for those Endeca users with full-text and unstructured data applications, other with massive structured data (ecommerce, CRM, etc.) will need to turn to other solution providers possibly. We are ready and waiting.
Tags: autonomy, endeca, hp, oracle, Oracle Acquires Endeca, Oracle Buys Endeca, search, search engine, search engines, software
Posted in Exorbyte, Software | No Comments »
Thursday, September 29th, 2011
Exorbyte doesn’t care but the feud between Oracle and the “cloud bunch” (Andreessen, Benioff of Salesfoce.com, and more) but we feel it has something else to teach us besides the low blows.
“Andreessen kicked off BoxWorks, the first-ever customer conference for cloud collaboration provider Box.net, this morning in San Francisco.
His firm invested in Box, he says, partly because he found that a lot of the other startups they were funding already used Box’s product.
As he put it, “Ten years ago, it was a joke: you’d raise $20 million in venture capital and write a $4 or $5 million check to Oracle, Sun, BEA, and EMC….When it started, Salesforce looked like a toy compared with Siebel. Look ahead five years later, it’s obviously better. Not a single one of our startups uses Oracle.”
Read more: Marc Andreessen: The “Clock Is Ticking” On Oracle - http://www.businessinsider.com/boxnet-2011-9#comment-4e84d76aecad04784000002d#ixzz1ZNOjydsW
Oracle is indeed under siege big time. But this cloud computing debacle is just one more front for them to fight on. SAP, their main rival in the DB business, has been taking effective swings at them regarding the lack of believe they have in in-memory databases (which are by-the-way the way forward for cloud computing also). Hasso Plattner from SAP has a field day here with this:
Our company sells SaaS site search for ecommerce and we also did away with Oracle from the start. Better yet, the vast majority of our customers are the same and this is a growing market!
Good luck Larry!
Tags: cloud, databases, in-memory, in-memory databases, levenshtein, real-time, real-time databases, search
Posted in Exorbyte, Technology, databases | 5 Comments »
Wednesday, July 20th, 2011
An excellent article on the BBC caught our attention this summer. Charles Duncombe, a Brtish ecommerce entrepreneur, says an analysis of his website metrics shows a single spelling mistake can cut online sales in half.
Mr Duncombe says when recruiting staff he has been “shocked at the poor quality of written English”. Sales figures suggest misspellings put off consumers who could have concerns about a website’s credibility, he says. Worse many site search engines for online stores are incapable of finding a product based on a misspelled query of if a company employee has misspelled a product’s name or description.
Needless to say, only a store search engine capable of cutting through misspelled queries, data, typos and all the unexpected inputs from visitors is capable of guaranteeing satisfied buyers and maximized revenue. Exorbyte Commerce does that better than any other store search engine.
See the article here: http://www.bbc.co.uk/news/education-14130854
Tags: commerce search, conversions, eCommerce, misspellings, online store, revenue conversion, revenues, search, search engines, spelling, typos
Posted in Merchandising, Technology, Usability, eCommerce | No Comments »
Wednesday, May 18th, 2011
Why do it?
This previous article about the long tail of search in ecommerce makes already a good case for better site search in ecommerce sites. This said, it is also important to remind people that past studies have shown that about half of all visitors gravitate straight to the search box. They don’t waste their time browsing for products to buy in the directory structures of your store catalog’s categories, they want something specific and Google has taught them to find this with search. This little roundup of the state of industry aims to show you where we think ecommerce search engine products are coming from and where they are going.
What is it?
The Search Legacy
Search software has been all the rage since the late 90s. When I worked for Netscape, and later Infoseek, I saw the search software industry explode and split into multiple major segments. The Internet search segment (Google, and co.) found pursued growth opportunities in media with added value portals and other advertising-supported business models. The cost of entry in this category became so high eventually, however, that the late arrival of Google on the Web was seen as a miracle-of-sorts. Search software editors on the other hand kept sprouting with new products and new approaches.
Some search products were outgrowth of web search ventures: Ultraseek, Inktomi, Google Appliance, etc. Others aimed straight at the enterprise with a mix of full-text search tools and database and ERP connectors (Endeca, Microsoft/Fast, Autonomy, SLI, etc.).
While that was happening no-one paid too much attention to the relatively small ecommerce opportunities. Enterprise search contracts often started at around $250K (sometimes yearly). So why focusing on ecommerce where the only enterprise-size businesses (eBay, Amazon, and a few others) mostly hired their own R&D and software development staff.
The Ecommerce Legacy
Strangely enough (as Tobi Lütke the founder of Shopify pertinently explains here) search remained anafterthought for most of the ecommerce software developers. Their initial focus was to automate the brick-and-mortar processes of the real world through advanced database powered applications. They never envisioned the volume of traffic that some ecommerce properties see these days and the very small portion of visitors, surfers, and product researchers who actually complete transactions.
Many ecommerce platforms were developed based on a mix of full-text search technology (crawler-indexer). These technologies are just not what the ecommerce platforms need because their data (products, services, customers) is already structured in databases. Furthermore, the search results logic that presides over content (documents, web pages) is very different from the way a shopper behaves with the catalog of an online store. Retailers need ways to serve search results that help the visitors find what they want to search for. Not just results to a well-prepared query. Online merchants also need to cross-sell, cross-promote, and present visitors with as many relevant opportunities to transact as possible.
What is Site Search for Ecommerce?
So at the end of the day, running simple queries against a database just isn’t enough anymore for powerful ecommerce search experiences. What is needed in terms of features is a mix of the following features:
- a powerful structured data search facility which can handle fast recall with error-tolerance like:
- misspellings
- synonyms
- taxonomies (related terms, catagories, facets, etc.)
- advanced interfaces like real-time AJAX search results controls and autocompletes (like the Google Suggest)
- Search Reporting and Analytics to allow continuous improvement of catalog inventory and descriptions to match highly sought after products (this data also needs to be exportable to third party business intelligence tools like Google Analytics or Omniture)
- advanced search-powered merchandising modules like “related products”, “contextual promos”, “you might also like” which can only get better if they are based on the same error-tolerant features cited above
- full-text search for non-catalog content on the store’s web site (web pages, support pages, corporate content, etc.)
What’s Ecommerce Search made of These Days?
Search-only offerings:
- Exorbyte Commerce
- Endeca
- Omniture / Mercado
- Celebros
- SLI
- Fast
- Nextopia
- Baynote
- Shoptivate
Ecommerce platforms with better-than-average search features:
- Prestashop
- Volusion
- Demandware (not on all stores)
What features and technologies?
Clearly, at Exorbyte, we have taken one orientation which we firmly believe in. Here are some of the features we believe every search app should offer to ecommerce users:
- a proprietary search engine that allows high error tolerance, very fast recall (for autocomplete) and phonetic, edit distance, and more matching algorithms
- merchandising features for easy control of search ranking, faceted search, and more.
- advanced reporting showing what people are search for, what they are finding, what they are not finding and allowing conversion tracking
- a system that allows for multilingual product data search (any language for product catalog data, any language for search matching logic above)
- hosted software-as-a-service mode for delivery (considerably more affordable for the online retailer)
- complete integration using AJAX technology ( no more jumping to a special domain or complex DNS integrations, search is delivered as if it lived on the store’s own servers)
- easy customization for special custom data attributes about products for rich catalogs and faceted search
So what do you think? Please comment.
Tags: conversion rates, eCommerce, factfinder, industry, merchants, nextopia, online retail, product catalog, revenues, search, search engines, shoptivate, sli, sli systems, transactions
Posted in Exorbyte, eCommerce | No Comments »
Thursday, April 7th, 2011
I just read a post from Mary Ellen at Librarian of Fortune. It talks about the different perspectives of Librarians vs. the general public regarding what makes good search results.
This is a typical haunting thought for those concerned with search engines and their quality. I want to share an insight which basically comes down to this surprising fact: a good search engine is more than a great algorithm. half the work goes into the user interface, facets, autocomplete, the data structure of the index, and all the other features which supports users in developing a good query.
Poor quality query + Great Search Algorithm = Poor Results
Great quality query + Poor Search Algorithm = Good Results
Needless to say, if both the quality of the query and the matching algorithms are good, results will probably be optimal.
Am I blaming poor search results quality on ignorant users who couldn’t develop an advanced search query like ["poor search * quality" -site:google.com -"marketing"] on their own? No, of course. It’s up to the software to make it easy to develop optimal queries, using suggestions, autocomplete, search facets, etc.
I have always been an avid searcher and loved to do research at the library during my college and grad school days in the late 80s and early 90s. Then I worked for Infoseek, Netscape, and a bunch of web search engine companies in the 90s. I often had the exact same thought as Ellen in her post.
The difference between the synthetic mind of the average search user and the analytic mind of the research pro is the following, I believe: The average search user would rather receive two relevant hits for his query and nothing else. At least that’s what the marketers “monetizing” the search engines with ads at Google and other companies would like to think ;-) He/she has little time and always wishes that the “magic of search” would guess the exact desired result even when his/her query doesn’t actually express it very precisely. The pro user is different. He/she wants to use his/her brain to analyze and discover the boundary of the semantic field being searched. Example: search for “Japanese Raw Fish Cooking” might lead him to other words, like “sushi” which will become part of the final, ideal advanced search query such as ["Japanese Raw Fish Cooking" AND "Sushi"].
This concept is important because now that I work for a company that manufactures custom search engines, we find customers that want all sorts of preciseness. We say that our software we can tune the “search corridor” to be rather large or precise. But beyond this optimizing between the conflicting influences of noise (too many irrelevant results) and silence (two few results, or even zero results), what became obvious for me over the years, was that just like in the library, a good search sessions involved careful development of one or several ideal queries. One often starts a session with a vague idea of what the search query should be. That’s why I am a big fan of Google’s instant search, its suggest tool, faceted search and other such tools which, in essence, guide the user towards the ideal query. We also have developed such an autocomplete tool which we even offer as a SaaS tool for ecommerce (Exorbyte Commerce – sorry shameless plug here).
Please comment.
Tags: information, interface, librarians, results, search, UI
Posted in Exorbyte, Technology, Usability | No Comments »
Thursday, March 31st, 2011
If you are familiar with exorbyte, you have read me and others boasting our unique error tolerance again and again and you may still wonder why this is so important to ecommerce. This is why: The “Long Tail of Search”
Search Logs Blog.exorbyte.com
The term “Long Tail” is recognized as a key concept to grasp the differences between the main categories of search queries online users submit into any search engine, on an online store or even a web search engine like google.com. The main idea behind the “Long Tail” is that while most people look at the top of a list of search logs (list of all search queries submitted by users on a web site during a given period – see chart above), the real important terms lie at the bottom of that list in what’s also called “The Long Tail”. The bottom of that list is what matters because that’s where most of the search queries lie. In many search engines on online and elsewhere the long tail is reported to be over 80% of the search queries for any given period. In other words, 80% of all queries made into the search engine are unique, while 20% are popular queries.
The Long Tail of Search
The first lesson here is that popularity doesn’t matter as much as some would like you to think. If you are selling lots of products, which most online stores are (because they can – no floor space limits, drop shippers, etc.), then you care about selling all these products not just the same few popular products over and over again. Inventorying is costly.
The second and most important lesson for Exorbyte was that many long tail search queries are unique because they contain typos, misspellings, phonetic spellings, and related but not matching queries. This is where error tolerance comes in. We can match more of these search queries than any other search engine. And this translates into more sales. Many of our customers report improvement of 5% to 20% of their conversion rates after implementing Exorbyte Commerce for instance. It’s that simple! So if you have an online store, wait no longer, try us out, it’s free, takes 10 minutes to install, 10 minutes to remove and it will change your business performance and the satisfaction level of your customers forever.
High conversions are a direct result of the wide search query coverage of Exorbyte's proprietary search technology
Tags: conversions, eCommerce, long tail, online store, popular, popularity, query frequency, revenue, search, search distribution
Posted in Merchandising, eCommerce | 6 Comments »
Monday, January 31st, 2011
The Google Suggest Autocomplete
Exorbyte continues to see interest from customers in Ecommerce and beyond for its autocomplete capability. The trend is not surprising as its been endorsed by some of the largest companies online. For instance Google surprised many analysts when it turned what many saw as a mere widget, its Google Suggest (see pic to the right), into a new core search interface: Google Instant Search.
I must say we at Exorbyte were not surprised at all. We have seen first hand that instant search interfaces (also called autocomplete, incremental search, suggest, auto-suggest, search as you type, typeahead, etc.) are capable of changing and enhancing the search process entirely. They bring a boost of ease of use and finadability to conversions for ecommerce and to just about any other application featuring a search feature. Our whitepaper on the topic of advanced autocompletes is worth a read if you want to know more.
But just to put this in context here are a few rules to help the business-focused person make sense of the differences. Ask the following questions:
- Are the search suggestions returned straight from real live (indexed) content or just popular search query suggestions?
There is a huge difference here. Real search suggestions from content can be hard to return every 10 ms after someone types the next character but they are infinitely more valuable to the user because they cut through the search process straight to relevant results.
- Are the search suggestions returned with a high degree of error-tolerance? Does it allow for phonetic similarities, aliases and synonyms, complex misspellings (a letter difference or 3 right in the middle or at the start of the word) , multi word queries, can the user erase and re-type, is the system available on search results pages too, etc.
Error-tolerance, or the ability to find close matching suggestions whatever the source of the error (unknown spelling of someone’s name, typo, bad spelling in the content itself, is a huge benefit here. The reason is that autocomplete suggestions happen at the very beginning of the search process as the user has yet no specific idea of what an ideal query for his desired result(s) (disambiguation process). Therefore error-tolerance is just not a nice-to-have here. It’s a must have because that’s exactly what the autocomplete is for: preventing errors that will return too much irrelevant results (noise) or zero results (silence).
- Does the autocomplete return results in an advanced interface? Is it configurable to include actual matches from content but also suggested categories, or other facets, images, etc?
Having a list of items is not enough. The autocomplete needs to also help the user disambiguate him/herself. Having a simple list doesn’t help as much as an organized list that also displays associated data (ie. prices of suggested items in an online store). The addition of images is a big plus but needs to be done only at no cost to the speed of the system (which requires a special image server like that of Exorbyte).
Hopefully, you will find these suggestions useful and you apply them the next time you are choosing a search system or offered a UI with an autocomplete that does or doesn’t meet these criteria. See Exorbyte’s own autocomplete at work on our Exorbyte Commerce Demo now if you need an illustration and don’t hesitate to leave us comments or questions below.
Tags: autocomplete, conversions, disambiguation, error tolerance, google suggest, incremental search, search, suggest, typeahead
Posted in Banking, Business, Exorbyte, Online Travel, Technology, eCommerce | 3 Comments »
Thursday, December 23rd, 2010
Why is Exorbyte Commerce better than its competition? To answer this, we first must contemplate the nature of search itself, in relation to e-commerce sites and sales. Of course, the purpose of an e-commerce site is to generate sales of the products featured on the site. The web presents unique challenges to the traditional sales model, where two human beings interact, be it in person or on the phone. One of the most powerful tools in sales is negated; the ability of the sales person to intuit the customer’s needs via verbal and non-verbal queues and guide them to the appropriate product. There are a variety of methods e-commerce sites use to attempt to compensate for this short coming, including: custom facets and categories on the site, animation or short movies that load and instruct the customer how best use the website, putting featured products or links on the first page of the website, et cetera. One of the most powerful ways to help compensate for the “mechanical” nature of a website, and one of the most overlooked in the SMB market, is robust search functionality.
Because of the ubiquity of Google with the web, more and more people are actually using the search bar as their sole tool to navigate a website and find the product they’re looking for. It is imperative, in order to capture a sale, and in order to engender brand loyalty, for the search box to be error tolerant, extremely quick, and extremely user friendly. These features are taken for granted in the enterprise market. Go to any very large website and you will see that the search not only corrects mistakes in spelling and semantics for the end user, but also suggests products as the person is typing while compensating for spelling errors. This is all in an effort to get the end user to the product as fast as possible, so they do not become frustrated and move to a competitor’s site. It also instills into the customer a sense that the site is extremely user friendly, increasing their chances of returning to that site when they are in the market for a similar product. The technology to provide true error tolerant, fast and intuitive search functionality to an e-commerce site has been, to date, very costly. It often runs into the 5 and 6 figures, putting it well out of reach for the average small to medium site. Exorbyte Commerce has solved this problem. Look out for part 2 when I spell out how we’ve done this in detail!
Tags: e-commerce, Exorbyte, nextopia, search, yahoo
Posted in Exorbyte | 2 Comments »
Thursday, December 2nd, 2010
Check out this important article by Adam Audette closing a year of intense changes in the search engine industry with great impact for store owners.
3 Important Ecommerce Trends To Watch.
Enjoy!
Tags: eCommerce, online retail, online store, online stores, search, seo
Posted in eCommerce | No Comments »
Every economic cycle comes with its host of enterprise software trends. Big Data hs become a recognized phenomenon in 2011. In May 2011 McKinsey released the “Big data: The next frontier for innovation, competition, and productivity” report. It started with: ”The amount of data in our world has been exploding and analyzing large data sets—so-called big data—will become a key basis of competition, underpinning new waves of productivity growth, innovation, and consumer surplus”.
