An Interview with Craig Bassin
EasyAsk was one of the first search vendors who demonstrated access to structured and unstructured data from a single interface. Our paths crossed in the mid 1990s. The firm landed some blue-chip accounts such as Lands End. In 2005, Progress Software acquired EasyAsk, and I checked on the company every year. I learned that Craig Bassin, the choreographer of deals at B2Systems acquired EasyAsk from Progress Software in May 2009. A former partner in an investment firm, Mr. Bassin had the technical and financial expertise to leverage the EasyAsk information retrieval technology. I spoke with him on November 15, 2010, at Sorrelina’s Restaurant in Boston, Massachusetts
The full text of the interview appears below:
Tell me about EasyAsk and what it uniquely brings to the market?
EasyAsk has been in business since the late 1990s. Larry Harris, one of the early pioneers in natural language technology, founded the company and is still active as the technical visionary. EasyAsk developed a natural language technology platform that solves key business problems in the e-commerce search and merchandising and business intelligence markets. It was an independent company until 1995 when Progress Software bought it. Then in 2009, I bought the company and made it independent again. Today, EasyAsk has over 200 customers, is profitable and growing.
EasyAsk’s unique natural language technology helps people find information faster and easier by enabling them to ask questions in plain English, converting the query into SQL and retrieving an accurate answer. The technology is used in two products: EasyAsk eCommerce Edition, an e-commerce search and merchandising solution that has proven to drive the best buyer conversion rates in the industry, and EasyAsk Business Edition which offers the easiest, most intuitive manner to search and explore corporate data.
What are the differences between EasyAsk and systems like Endeca or Mercado?
One of the interesting problems with e-commerce has to do with the way people were taught to search. Google, the predominant Internet search engine, has trained people to search by keyword. You enter one or two keywords, see a set of results, enter more keywords, see refined results, and continue this repetitive process until you find what you are looking for. So Google has taught people to search for generally what they want, not EXACTLY what they want. This process requires a large number of clicks before you find what you want.
When searching vast volumes of information on the Internet, consumers may tolerate this iterative process. However, consumers looking for a specific product on an e-commerce website will not. Studies have shown 80% of visitors will abandon a site if the search results do not quickly find what the user is looking for.
Endeca, Mercado and other e-commerce search products, at their core, still offer keyword-based search. In an attempt to speed up the process, these products use what’s called navigational search, whereby the user continually clicks on different attributes until they find a specific product. This approach still requires many “clicks” before the user finds what they’re looking for.
What about faceted search or what I call “point and click” navigation?
While EasyAsk also supports the navigational style of search we go much further in helping customers find what they want quickly. EasyAsk’s natural language approach allows buyers to enter an entire descriptive phrase of exactly what they want. The natural language and semantic processing engine understands the context of the search and returns accurate results on the first page, greatly increasing conversion rates. With EasyAsk, customers can chose how they want to find products, and they will find them faster.
Natural language technology also drives stronger spell-correction, relaxation and stemming capabilities, eliminating a much greater number of “no results” pages than other types of e-commerce search technology. When a customer enters a search and the site returns a page with no matching results, the visitor is more likely to leave the website and find somewhere else to get what they want. EasyAsk enables e-commerce websites to always return search results, reducing the number of lost visitors.
What other use of NLP is in EasyAsk system?
In addition to enabling natural language queries in the search box, EasyAsk extends this technology to its suite of merchandising tools. This enables merchandisers to define their own offer and promotion rules in plain English, driving an unprecedented level of ease of use. And, since the merchandisers are able to manage offers, promotions and other merchandising techniques on their own, EasyAsk helps create more agile marketing and merchandising groups, who can quickly response to customer and market needs.
Can you give me a flavor of the type of solution EasyAsk delivers?
In the e-commerce market, Travers Tool Company is a classic customer solution for EasyAsk. Travers is a tool and die products distributor. They pride themselves on being easy to do business with. They want to maintain that, while at the same time moving 30% to 50% of their business online.
Travers has a complex product catalog that contains over 100,000 SKUs with intricate product attributes and sophisticated relationships between the products. They needed high performance e-commerce search software that could meet the scale of their search indexes, but also made it easy for their customers to quickly find products.
With EasyAsk, Travers Tool customers can perform complex searches in their own way. This has led to an 8% decrease in the time spent on site, coupled with a 12% increase in average order size. Clearly, customers are finding what they need faster and, the merchandising tools are enabling Travers to do a better job cross-selling and up-selling.
What are the three key technical attributes of EasyAsk eCommerce Edition?
That’s a good question. Let me hit a few key points.
First, natural language, which I already mentioned. EasyAsk’s natural language technology enables better search and merchandising. Customers can use the search box to describe exactly what they’re looking for – “red sleeveless dresses under $50” – and e-commerce websites can deliver more accurate search results, faster. Plus, natural language and semantic processing eliminate “no results” pages, keeping visitors on your website.
Next is scalability. Our advanced indexing technology supports extremely large, scalable product indexes. The advanced EasyAsk search indexing techniques support indexes of twice the capacity of typical in-memory, high-speed, searchable indexes, providing a significant performance boost in searches across large datasets.
Then ease of use. We know from our customers that EasyAsk’s merchandising tools are extremely easy to use, allowing merchandisers to create targeted offer and promotion rules on their own, in plain English. This also reduces information technology costs required to maintain the search engine.
Are their any metrics you can share with me?
Our advanced indexing technology supports extremely large, scalable product indexes. Our indexes can scale in multiple dimensions, including number of products or SKUs and the number of attributes used to drive user searches, giving EasyAsk the ability to index extremely large datasets. We support indexes of twice the capacity of typical in-memory, high-speed, searchable indexes, providing a significant performance boost in searches across large datasets.
EasyAsk also scales easily with Web load balancers and industry standard clustering technology. Load balancers can balance HTTP and SOAP requests across multiple EasyAsk servers to scale front-end search load. Clusters can be used to scale EasyAsk instances to drive larger volumes of search data and results. And, we support 64-bit operating systems, allowing us to load high-speed searchable indexes completely into large memory address spaces for extremely fast access.
EasyAsk’s natural language technology provides an extremely high degrees of search quality without the need for expensive search tuning, providing excellent “out of the box” performance.
When you look out nine to 12 months, what are the major trends in your sector of the search market?
We’re seeing a lot of small and medium sized e-retailers taking advantage of more sophisticated search and merchandising capabilities. They’ve realized they can compete with the big e-retailers as long as they have the right e-commerce capabilities. We’re a perfect fit for these smaller companies because EasyAsk provides superior search right out-of-the-box and incredibly easy-to-use tools for merchandisers to manage their products. We can provide them with and Enterprise-class product at a price that works for them.
Another trend we see is more packaged e-commerce implementations using the Software-as-a-Service model. EasyAsk’s out-of-the-box tuning and easy dictionary building provides a rapid, low-cost implementation model that allows e-commerce search to be up and running in as little as 2 weeks, as opposed to other systems that may take months.
Let’s switch gears to EasyAsk Business Edition. How is EasyAsk Business Edition different from enterprise search?
As you well know, the typical enterprise search product is geared towards allowing users to search unstructured or semi-structured data using keywords to find documents they need. This is good when a user is looking for a specific document, like a contract or performance review.
EasyAsk Business Edition addresses a completely different problem – giving casual users faster, easier access to corporate data. At our core, EasyAsk is all about Natural Language linguistic processing, that is, understanding the ‘intent’ of any given question, or query. We’ve extended our intuitive search capability into corporate data allowing users to search, analyze, report and share results. Users ask questions in plain English—“which customers purchased products but not service contracts last quarter,” or “What was the top sales promotion for each product last year,”—and EasyAsk returns fast, accurate answers.
Who is your target user for Business Edition and why?
Great question. We designed EasyAsk for casual business users who need immediate access to data so they can make informed decisions improving their ability to increase sales, service customers and execute operational processes. These casual users don’t typically have the skills or training to write ad-hoc reports. And, they can’t wait a few weeks for IT or a data analyst to get them a custom report.
What’s an example?
Imagine a sales manager who is always under pressure to find new sales opportunities and bring in more revenue. Using EasyAsk, that sales manager could ask questions like, “which customers bought products last year but not this year” and “what were the top 5 installed base promotions and offers by revenue.” Using the responses to those two questions, the sales manager can quickly create an ad-hoc sales program that targets existing clients for upgrades, using the most effective promotion offers. The sales manager can then share the target list and offer with the sales team and they are off and running, creating new revenue. All done in a matter minutes, without the need for any help from the IT team or business analysts.
Why is this the right time for natural language solutions for business intelligence?
Well, the first part is easy. EasyAsk Business Edition is a new and exciting solution providing the ease of Natural Language search without months of complex BI design and deployment.
Traditionally the IT teams that made the decisions about which BI tools to use were worried more about capabilities that would streamline their multi-million dollar data warehouse and analytic projects than those that would enable business users to perform the deep analysis and reporting that management needed to make long term decisions.
One major factor shifting the landscape is the ever-growing popularity of SaaS based applications for finance, sales, service, order management, human resources and more. With a large number of front-line workers using SaaS based applications, they’re not willing to tolerate long lead times from IT staff or administrators to get new reports written. In addition, many of the SaaS application buyers have fewer IT resources – one of the benefits to using a SaaS-based application.
A natural language search and analysis toolset integrated with a SaaS application, is the perfect toolset for the front-line SaaS application users. They can get whatever data they need regarding customers, sales or service performance, financial transactions, order and shipment status, and more, right from EasyAsk. They don’t have to wait for an administrator to create a new report.
The added benefit in a SaaS application environment is that EasyAsk already understands the underlying data model and business terms. This makes deployment easy – connect EasyAsk to the SaaS application and start asking questions immediately. No need to build a data mart or extract the data into an external analytic data store.
I noticed you recently announced a version of EasyAsk Business Edition for NetSuite. How will a NetSuite user use the product and what type of questions can they get answered?
EasyAsk Business Edition for NetSuite was announced in September, 2010 and is a product we are very excited about. We have been teaming with NetSuite since April to integrate and deliver both the eCommerce Edition and Business Edition products on the NetSuite platform.
EasyAsk Business Edition for NetSuite understands the NetSuite data model and all the related NetSuite business terminology out-of-the-box. So any NetSuite user can ask questions about their specific business or operational function to speed their execution.
The sales manager example I gave before is a prime example of the type of question a NetSuite customer could ask. A shipping manager could ask ad hoc questions about Sales Orders and Item Fulfillments to see which orders have shipped, which haven’t and why they haven’t. A finance manager could ask questions about Invoices to see which invoices have been issued, which have been paid, and which are late.
EasyAsk Business Edition is designed to provide answers to questions for senior management as well as C level executives. If a board member asked the CEO about service issues at some of the top customers, the CEO could simply ask EasyAsk “show me all open service cases for my top 10 customers by revenue.”
If I were to describe EasyAsk Business Edition in one sentence, I’d describe it as follows: You ask EasyAsk questions the exact same way you’d ask it of your VP Sales or CFO.
SAP Business Objects recently introduced a data exploration tool, how does EasyAsk compare to that, as well as SaaS BI companies such as Good Data, PivotLink or Cloud9?
The SaaS BI companies are just taking the entire BI and analytic development process and turning it into a SaaS model. Users still need to define extract processes, build a data mart, perform analysis and create reports. SaaS BI offerings are really geared toward lowering the cost of traditional business analysis rather than addressing the need for everyday users to search and analyze their data.
Similarly to EasyAsk Business Edition, SAP Business Objects Explorer tries to put intelligence into the hands of the casual business user. But that is where the similarities end. With Business Objects Explorer, you still need to build an expensive data mart or data warehouse that contains the information for the users. In fact, the initial version of Business Objects Explorer only worked with SAP Business Warehouse.
The second primary difference between EasyAsk and Business Objects Explorer is the interface. With EasyAsk, you can ask a question to any level of complexity and you get an accurate answer. If the user knows what they want, they can get their answer on the first try. If the user needs to explore, they can ask multiple questions and refine their results until they get what they want.
Business Objects Explorer only supports a navigational style interface, requiring the user to click and explore different paths to find the information they are looking for. This type of search will always take longer for the user to find the information they’re looking for.
The difference between EasyAsk Business Edition and Business Objects Explorer is similar to the difference between EasyAsk eCommerce Edition and other e-commerce search products such as Endeca and Mercado. In both cases, the other products provide users with only a navigational approach to finding what they want. EasyAsk’s powerful natural language engine drives both a direct style of asking for exactly what you want as well as a navigational, approach, providing greater versatility and ease of use.
Where can a reader get more information about EasyAsk products?
There’s a really good resources section on our website. I recommend watching the demos. They’re short, concise and provide an extremely good overview of EasyAsk’s capabilities.
EasyAsk combines several functions that are getting considerable attention in organizations struggling with key word search. EasyAsk allows the user to ask questions or point and click, combining NLP with faceted navigation. EasyAsk can acquire, index, and make findable content in structured and unstructured sources. Under the new leadership, EasyAsk is poised to capture customers who want a scalable, cost-effective solution to enterprise and eCommerce “findability” and business intelligence challenges. Worth a look and a demonstration.
Stephen E. Arnold, November 16, 2010