Dozens of Internet startups, many of them Seattle-based, want to help the big online publishers make money (and take a cut of that, of course). One of these is FizzyLab, a one-year-old company that recently launched the “Relevator” software, now being used by a handful of online newspapers and Snowball.com, a site for the young and the hip. The aptly named Relevator brings to a reader’s attention other related articles within a site or network of sites. The goal? To increase a site’s page views, click-throughs and eventually e-commerce transactions. Below, we asked founder and CEO Scott Hudson what distinguishes FizzyLab from the search-engine pack.–SJ
Seattle24x7.com: Scott, FizzyLab is often referred to as a search-engine company. How are you different from the other “smart” and powerful search engines like Northern Light?
Hudson: First of all, we do not search the entire web for related content. Our software works on individual sites or a network of sites, and it’s going to be the connective tissue, or the glue, if you will, of a site or network. Imagine if you could bring all the different types of content within a site together–the advertising, the products offered for sale, the editorial content. Our software establishes relationships among all those elements to generate more page views, online sales, etc.
Also, we are not involved in “key-word” searches [when you type in a word or phrase and system looks up related articles]. Most key-word search engines within content-heavy websites don’t work well because they simply look for how many times a specific word or phrase appears.
Seattle24x7.com: How is your Relevator software different?
Hudson: If you go to the Seattle Post-Intelligencer site, for example, and click on one of today’s headlines, you’ll see a “Related Articles” button, powered by FizzyLab.
To get you those “related articles,” we use a statistical approach. Our system not only looks at the number of key words in an article. It also weighs key words based on where they appear in relation to other words, and it then comes up with a model for that article that is compared to other article models. So we’re comparing apples-to-apples, so to speak.
Seattle24x7.com: Yes, I’ve already gone to the P-I and got sucked in because of FizzyLab’s “Related Articles.” Does your system rely at all on Artificial Intelligence (AI)?
Hudson: There is AI in the system. But it’s not a primary driver because AI slows down the search. Basically, we’re taking a statistical approach that is supplemented with AI, when time permits. For instance, we’re about to announce a new product–an email notification system–that allows for enough time to add AI to the search.
Seattle24x7.com: Are you getting paid?
Hudson: We haven’t done a single give-away deal, and our revenue per-account has been dead-on with our projections. How much we charge varies with amount of traffic, size of archive, etc. And the range is pretty significant–say, $500 a month for a small site to $10,000 for a larger site. Overall, we’re targeting sites (or networks) with 10 million or more page views a month.
Seattle24x7.com: Your success depends on Relevator’s ability to increase page views. Right?
Hudson: Yes, we’re very aware of that. One of the things we’re finding is that FizzyLab is proving to be more useful on business news vs. general news. In fact, in the next several weeks, we’ll be announcing two major business publishers as clients. And we’re constantly adding features, such as the ability to drill down into a specific company or industry mentioned. Also, within the next several months, the “Related Articles” headlines will appear automatically, without readers having to click for them.
Seattle24x7.com: There doesn’t seem to be much money in search services. What’s FizzyLab’s long-term plan?
Hudson: We want to do for online content what Cisco [Systems] did for networks. By this fall, we should have all the products in place to tie together all the different components of a website–editorial, advertising, e-commerce, etc. Six months from now, about 80% of our business is likely to be behind-the-scenes, backend services [that the site user doesn’t initiate].
The sweet spot, of course, is e-commerce. We’re going after those publishers who are aggressive about publishing content and selling products online, eventually including audio-content providers. For the large business publishers we’ll announce next week, for instance, there’s an opportunity to sell some high-margin financial products that have a high click-through rate.
Seattle24x7.com: So FizzyLab would point readers to related products or ads, as well as related articles?
Hudson: Yes, that’s one way it would work. Right now, companies like Engage are doing profile-based targeting. But we think a better indicator of what would be appealing to someone is what they’re reading. Relevator would make that connection.
Seattle24x7.com: What about Autonomy PLC?
Hudson: Autonomy is probably the most similar to us. But their core technology was not built to be served via a hosted model. It’s a “neural-net” system that resides on the customer’s servers and costs the customer about $200,000 upfront to install, plus it has to be maintained.
By contrast, FizzyLab’s system is hosted on our servers; it’s an ASP model. So there’s no upfront cost to our clients. We’ve competed with Autonomy in about 10 different bids so far, and we sell against Autonomy better than any other company. That’s probably because there is no other company currently that offers a hosted related-content service with all the benefits of ours.
Also, until recently Autonomy has focused on the corporate-enterprise market. We’re more focused on consumer sites, and that will continue to be the case.
Seattle24x7.com: But, Scott, haven’t you heard that consumer sites are worthless?
Hudson: Ha! I actually think advertising is going to work on the Internet, contrary to current thinking. We just have to figure out what type of advertising works. And what types of products sell well online. Interestingly, in the past few months, the direct-mail marketers have come online in a big way–especially the credit-card issuers, the financial-products companies, etc. They’re spending two to three times the money they did last year. And the results they’re getting are really good.
Seattle24x7.com: FizzyLab applied for a U.S. patent in January 2000. What’s that about?
Hudson: There are some 39 claims on our patent application, covering the entire process that we’re using, as opposed to just one algorithm. Our lawyers did a very extensive patent search and turned up a lot of technology that is close to ours, but nothing directly comparable. But you never know. We won’t find out about the patent status for another 18 months, and by then our system will have evolved significantly. Sure, the patent would give us more strength, but we don’t need it to compete effectively.
Seattle24x7.com: Where are you in terms of funding and employees?
Hudson: We’ll probably do another round of funding this fall that will include corporate investors. Right now, we have about 40 people– 10 in sales and 20 in software development. Our main need now is engineers. Also, within a year, we’ll probably double our sales force.
Soula Jones is Content Chief at Seattle24x7.com
Plop, Plop, Fizz, Fizz
FizzyLab Corp.
www.fizzylab.com
Market Place Tower
2025 First Ave. South, Suite 1150
Seattle, WA 98121
206-728-4282
Started: 1999
Executives: Scott Hudson, CEO (co-founder of Headbone Interactive and former Microsoft employee); Ken Willner, VP Marketing (formerly director of account management at McCann-Erickson Worldwide); Tom Haug, VP Engineering (formerly at DigitalXpress, a Boeing spinoff); Dion Long, VP Sales (formerly at eCharge and Spry/Compuserve).
Funding: $11.2 million so far. Investors include: WaldenVC (also invested in Snowball.com, a FizzyLab customer); Sandler Capital Management; Guide Ventures; and several individual investors, including Nicholas Negroponte, co-founder and director of MIT Media Lab.
Employees: 40
Customers: Seattle Post-Intelligencer; Snowball.com; 11 McClatchy-owned newspapers.
