Emeryville, California-based Ask Jeeves <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: ASKJ)") else Response.Write("(Nasdaq: ASKJ)") end if %> operates a consumer Web search site as well as a corporate service that allows users to seek out information on the Internet using standard English. We spoke with President and CEO Rob Wrubel about the synergies between the two operations and why you don't want to be the number-four pets website.

TMF: I guess the best thing to do for starters would be for you to give us your description of AskJeeves.

Wrubel: Well, AskJeeves is a natural language question-answering service company, and we provide a powerful set of services to consumers directly through our consumer site at Ask.com and through our corporate service in which we provide a customized, outsourced service for corporations who want to have an easier, more intuitive, and more intelligent way of interacting with customers on their website.

Our mission for the company is to build a very large network or exchange of question answering that enables the consumer to get answers to their questions whether they come into Ask.com and ask a general question or whether they ask very specific questions of a specific company such as a Microsoft <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: MSFT)") else Response.Write("(Nasdaq: MSFT)") end if %> or a Dell <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: DELL)") else Response.Write("(Nasdaq: DELL)") end if %> or any number of our corporate customers in which they want specific information about products and services.

"I think e-commerce sites have finally hit their capacity."

By building a large integrated network we really hope to take advantage of what we think is the next wave of really powerful business opportunities on the Web, in which you're really leveraging your intelligence and volume and scale through a business-to-consumer (B2C) business, which is what we do on our website, into what is a very rich opportunity in the business-to-business (B2B) infrastructure market, and you're just continually leveraging intelligence and knowledge about people's language, questions, problems, and the needs they have for products and services from our high-volume platform into what are effectively lower-volume environments and corporate websites, but where the value of the transaction and the relationship and the customer value actually is higher.

TMF: At the risk of getting into something perhaps extremely technical, could you talk a little about how your technology works and how it was developed?

Wrubel: The way the technology works is the company was started in early 1996 by our CTO and founder Dave Warthen. The problem we were trying to solve was how to make it very easy for people to interact with Web content, and typically what we had looked at and seen was that on the Web there's a tremendous amount of content. The challenge has been allowing people to have an easy, intuitive way to find answers to questions, but also really to expand and enlighten them to possibilities they haven't thought of. So what we developed was the natural-language-parsing technology that would essentially take a plain English question, look at it both semantically and syntactically -- so both look at the word, the phrases, and the concepts you're using as well as the context and the intent of the sentence -- and then take the parsed understanding of that sentence and connect you to a couple of different systems.

[We have] a large database that we're developing that is composed of questions or abstractions of questions that have been frequently asked by people before. It's essentially a giant database of frequently asked questions. Each of those question abstractions has a whole set of possible answers that can point you to any Web resource whether it's a dynamically generated page, database or URL, so it's very open in its characteristics in that you can point to any place for an answer. And we also provide automated support backup to that with metasearch technology that enables us to supply answers when we don't have it in our database.

"Human beings will only take the time to get acquainted... with a couple of brands."

We capture all the customer intelligence about the questions they ask, the language they use to ask those questions, and the things they select in our system to better understand what customers need and want and what is successful on our system and, most importantly, what we don't have. So we're constantly improving in response to the questions people ask.

And then the fourth component is that we use an editorial system for populating those knowledge bases and expanding and enhancing the semantic system we use to process questions.

So those are really the four components: the natural language question processing system, the database, the intelligence-capturing system, and the editorial system.

The last couple pieces, which we just announced recently, are new extensions of that technology, and that is that we have a decision support engine which enables you to take complex problems such as buying consumer electronics like digital cameras or annuities and healthcare programs, and we engage you in essentially a preference-based dialogue that enables you to go beyond just question answering and finding a document.

The other piece of technology we just announced was the acquisition of a live help service so that if we don't have the answer we can automatically route you to an agent expert adviser anywhere on the Web to engage in an interaction via chat, text messaging, or cobrowsing so you can always have the best real-time personal service.

TMF:: I wonder if the best way to discuss the way your business operates is to handle the consumer side and the business side separately?

Wrubel: Yep. Let me tell you how we make the money on both and then let me tell you about what's really neat when you begin to integrate the power of the two together.

The consumer service -- think of it much this way -- we've assembled a very powerful, statistically relevant database of questions people ask us all the time and the answers. We supported it with very powerful metasearch technology, so when Jeeves doesn't have a match you can always look and find resources in other places -- and it's very automated and so highly scalable -- and what we've developed is a service that answers 3 million questions a day, learns every day. But underlying it is a very powerful targeting service for any advertiser or e-commerce vendor or someone who has great content because when someone asks a general question, we capture a lot of information about what their intent is because it goes beyond just the two-word keyword string. And we can serve highly relevant ad banners so Jeeves' consumer site makes money through banner advertising, e-commerce linkage, and lead generation services as well as sponsorship opportunities, and effectively what we have is what you can kind of consider a hosted active server page (ASP) service in which you subscribe to our question-answering targeting service to bring you new customers.

The corporate site is unique in that there are really three revenue streams, but two primary ones. We get paid for the customized development of a question-answering knowledge base for your corporate website that points to your content and deals with the questions your customers have every day. So companies pay us for that service, and it's almost as if it were a professional development service. Then what they also pay us for is the ongoing maintenance and support of the service, because what we do is we maintain a team that actually makes it get smarter over time and respond to the customer questions you have. And then there's a huge component in that, instead of licensing our technology, what we want to do is be committed to making sure you answer lots and lots of your customer questions, and we actually commit to a volume of questions that we will have answered and you pay us to achieve that volume. The deal sizes can range anywhere from a couple hundred thousand dollars on up to over a million dollars for that service.

TMF: Can you talk a little bit about the way in which you see the two divisions becoming profitable over time?

Wrubel: The way the two revenue streams work is that our corporate revenue stream is based on the number of deals we do with corporate clients and size of the deal, but we're also very conservative about the revenue recognition. We recognize that revenue over a 12- to 15-month period from the time we sign the deal. So that means by nature that revenue growth will be moderate in a sense because what we're effectively doing is building a very significant backlog of deals that we've run and worked out over time.

The consumer site makes money through all the advertising and we recognize that revenue immediately. Today if you look at the split, corporate and consumer, corporate was 6% of our revenues in Q1 and in Q3 it was about 34%, and we would expect over the balance of next year is that corporate will increase as a percentage of revenues, but because of those accounting characteristics it won't be this dramatic change in the way that we see the revenue mix over time.

TMF: But the cost structure is different between the two divisions as well, isn't it?

Wrubel: Yes, your original question was on profitability. The way the costs work is if you look at the two cost components for each, consumer, as with any good kind of targeting business or any kind of good classic exchange, you have a fixed cost in terms of your product spending, and then your variable cost is marketing customer acquisition dollars. And so consumer's essential profitability driver is when you finally get your volume, your reach, and your dollars per customers to a high enough level, and your customer acquisition costs go down as a percentage, then you get to profitability. So it has very many of the qualities of a media platform.

Corporate has as its key driver the variable expense of how many deals you want to do, and how much production costs go into the deal, and so the issue with corporate in terms of profitability -- as we grow and as we expect to grow that business very aggressively -- [is] our financial model because we recognize the revenue over time and take all the cost of the deal right up front. It means and implies by its nature that up front, in the early stage of growth in that business, you will take losses and then in the back half of the business you will have a highly profitable operation because what you end up doing is running through very high-margin revenue streams off the contracts and in your backlog that don't have significant costs associated with them anymore.

TMF: Though you seem to be saying the corporate division of your business is really going to be, probably, the more quickly growing of the two, I wonder, where do you see the consumer search sector going over the next five years?

Wrubel: I would say that we see very dramatic growth characteristics for both. A lot of people have focused on a consumer [division] and a corporate [division] and how do you see the two playing out over time. When I think of what we're building, when we build an integrated service, imagine that with our corporate service we do what we want to do, which is serve the top thousands of companies on the Web providing products and services to customers. If we build intimate question-answer systems for each one of those companies and really help them improve their customer reaction, the customer and the company win.

Once we've built that, the power of our system and the underlying technology is that each of those knowledge bases can be networked and integrated back into a larger whole. So if I build out question-answer services for the key technology players, all of that product and service information can get integrated back into Ask.com. So, we see dramatic growth characteristics for both, but it's driven by the synergy of the two working together.

TMF: Then what about for the sector at large on the consumer side? Do you think that there's going to be sort of a forced consolidation as companies that don't have something like another niche like your corporate side or simply don't have the scale, won't survive as stand-alone business?

Wrubel: I think the market's already consolidating. If you think about it, the first level of consolidation is where VCs will put their money and then the second level of consolidation is where consumers will spend their time. I think you've already seen the forced consolidation around -- you don't see a lot of CNET competitors being funded by VCs, and so media-intensive information sites kind of were the first ones to [experience attrition]. Now I think e-commerce sites have finally hit their capacity, so I think they're already going to see consolidation after this Christmas. How can you be the third or fourth pet site on the Web? And then I think it'll happen in the infrastructure companies as well. There's going to be consolidation and you either have to be number one in a category or number two at worst, or you have to figure out something.

TMF: Is it the nature of the Web that's made that sort of a reality of business?

Wrubel: I think it's the nature of capital markets in terms of who's going to fund who -- well capital markets fund wherever the momentum is. When you come to branded entities, human beings have how much shelf space in their brains for how many different brands providing different services? And even when they're on the Web, which doesn't have location and time as barriers, I still think that human beings will only take the time to get acquainted and have their trust with a couple of brands.

And think about what it takes to get to know even a Web company. We used to think that it was such a frictionless medium, but in fact an e-commerce company has got to prove to you that they'll ship on time, has to prove to you that they've got things in stock, has to prove to you they'll take your returns, and have to prove to you that they'll answer your question and that they'll design a great website -- and human beings don't like to invest time like that. So I think natural human laws of physics demand that you have fewer leading providers within any given category.

TMF: Well, Mr. Wrubel, I really appreciate your taking some time.

Wrubel: Oh, thanks, Dave.

Related Links: