Full text of email interview with Mark Goodstein of X1 (see my column in WSJE and FEER this week)
— Who are you aiming at with this product?
Not to be too simplistic, we’re aiming at two groups: consumers and professionals, specifically those who have a lot of email and files and who spend more time than they want searching for information on the Internet or intranet. The free version offers a substantial set of features that we hope will entice legions of users to use the product at
home and work, for all their information finding needs. The pro version has features that power users will demand, like indexing network drives and viewing files in their native formats, regardless of whether they have the native application installed. Both versions will continue to get richer over the coming weeks and months, as we add more consumer features, like media-specific tabs (pictures, music, etc.) and more powerful web searching and eCommerce-related features. The pro version will get support for indexing attachments, contacts, events, PDFs, and archives. We think these two prongs will encourage great numbers of people to use the product and will eventually allow us to crack the enterprise market, which is straining for simple interfaces to complex data: X1’s specialty.
— I’ve always thought this kind of product was really basic, and when Enfish came out in 1999, I assumed it would be massive. But it wasn’t, and nothing since has really caught on. Why is this? Does it have to do with new paradigms, or just the product wasn’t right, or people aren’t ready for it, or what?
Our approach isn’t that much different than others, but we’re staying focused on simplicity and speed. X1’s interface is visceral and innovative: allowing the user to winnow the searches down from all to just a few, instantly, as opposed to the normal none to many (sometimes with a coffee break) of today’s search engines and desktop search utilities. This interface gives the user the feeling of control over chaos, which is hard to underestimate. Many people have built up complicated directory structures for storing their files and email, all in an effort to just keep track. X1 allows the user to stop caring about the organization and more about the work!
This is a difficult question to answer because it seems like Enfish and others have done many of the things we’ve done, but several years in advance. I’m not sure why they failed to catch on like you assumed, but I don’t think the fundamentals have changed. The amount of data we’re responsible for is large and always growing; it’s in disparate formats and locations; the tools that help users wade into this sea of information are, maybe justifiably, difficult to understand and use; and there’s no incentive for market leaders, like Microsoft, to innovate. It doesn’t help that the dotcom bubble excited expectations and the companies responsible never followed through.
That said, we really do think we’ve created a beautiful interface to complicated data sets. We think of it as something between a spreadsheet and a database. So, like you said, Enfish should have caught on big, and didn’t. Just like databases were supposed to catch on big at the end-user level, and didn’t. Spreadsheets have tried to fill the gap,
becoming more database-y over time. But that’s a little ridiculous, as many people have come to realize.
— What’s under the hood? Presumably these programs have different technologies underpinning them? Could you explain a little of the challenges to minimize the downside of such programs — index size, performance loss, ease of use, success ratio of finding what you’re looking for, etc?
I assume most indexing technologies are actually pretty close cousins, separated by clever coding and intelligent choices. We all deal with the same limitations of compression, physical memory, disk space, etc., and all have to make trade-offs to deliver a product to market. X1 has an inverted index with all sorts of clever tricks to manage memory and
processor use to keep the indexing as invisible and painless as possible. Our goal, from the beginning, was to make a product that was as simple to use as possible, as fast as a machine would allow, and as invisible as possible. We’ve had success on all fronts and we’ll continue to improve and innovate as time goes by. We think the bottom line here is speed and simplicity. Speed allows us to skip all those complicated, frankly under-used, search features, while allowing the user to iteratively search (quickly) through their data. They may search twice before success, but certainly it’ll be faster and more satisfying. This is compounded by our innovative multi-field search interface. That’s it.
— Where do you see this going? Is searching a hard drive going to get more sophisticated a la data mining? Or is this a rough and ready product that will always fit the brute force approach?
Not to harp on this too much, but we honestly believe that our mission will be fulfilled and we’ll achieve big success if we stick to our dual goals of speed and simplicity. We can let Oracle do the OLAP while we do away with the DBA…