Matthew Koll on Search

Excerpts from a SearchEngineWatch interview with the founder and CEO of Wondir:

I’m still waiting for the great breakthrough in visualizing search results. Clusters, folders etc are nice. And I’m continually attracted to new attempts at graphical or spatial displays. And I like the new A9 interface, Nevertheless, nothing I’ve seen really has the intuitive clarity and simplicity needed to be a major improvement in efficiency for the user.

[Google is in the business of] advertising. That’s where their revenue comes from, so that’s their business. The Google site just happens to be one of their best distribution outlets (pretty convenient for them huh?). I think of Google as two companies. And it wouldn’t surprise me if one day there were to split in two.

I don’t think its possible to overemphasize the revolution that GoTo (now Overture now Yahoo) and Google have created in the world of advertising, For content and media companies, search companies, any company that is in the business of having users tell them something about what they are interested in – GoTo and Google have created a superb enabling environment.

I’m not expecting true natural language understanding any time soon. But we can continue doing a whole bunch of little things – such as specific results for known questions, recognizing known entities. We can make better use of user feedback, and of clues to the user’s intention. I think that far more than personalization, tuning results to the users task at the time will produce greater improvements in relevance.

I think searching by voice will be huge. When I have a question, I want to just say it out loud and have results appear.

[In 5 years, we will see] Voice access. Task integration. Unified natural language access to all different kinds of information: web documents, desktop documents, network documents, contact info, definitions, media segments, background info, and live Q&A of course.

Why Open Source is Unsustainable

Financial Times has an essay (which I don’t agree with) by Richard Epstein:

The difficulties with the open source movement , moreover, go deeper than the problems with a single provision of the GPL. The open source movement shares many features with a workers commune, and is likely to fail for the same reason: it cannot scale up to meet its own successes. To see the long-term difficulty, imagine a commune entirely owned by its original workers who share pro rata in its increases in value. The system might work well in the early days when the workforce remains fixed. But what happens when a given worker wants to quit? Does that worker receive in cash or kind his share of the gain in value during the period of his employment? If not, then the run-up in value during his period of employment will be gobbled up by his successor – a recipe for immense resentment. Yet that danger can be ducked only by creating a capital structure that gives present employees separable interests in either debt or equity in exchange for their contributions to the company. But once that is done, then the worker commune is converted into a traditional company whose shareholders and creditors contain a large fraction of its present and former employers.

The bottom line is that idealistic communes cannot last for the long haul. The open source movement may avoid these difficulties for outside contributors who work for credit and glory. But how do the insiders, such as Linus Torvalds, cash out of the business that they built? And in the interim, how do they attract capital and personnel needed to expand the business? Traditional companies have evolved their capital structures for good reason.

But suppose this analysis is wrong. One clear policy implication remains: this novel form of business association should succeed or fail on its own merits. The do-or-die question is whether open source offers a low cost solution to particular problems. Ordinary companies will make just those calculations, but government agencies may be swayed to take a different tack, as has been suggested by a number of EU studies. That temptation should be avoided. Governments are bad at forcing technology by playing favourites. If open source is less effective than proprietary software, that gap should not be ignored by positing some positive network externalities that come from giving it a larger base. Proprietary systems also show positive network effects from increased users, as software designers are always attracted by a larger installed base. Its a tough world out there, in which no one should be exempted from the general competitive pressures of the marketplace. The fiduciary duties of government to all citizens demand no less.

Metadata for the Masses

Peter Merholz writes:

what if we could somehow peek inside our users thought processes to figure out how they view the world? One way to do that is through ethnoclassification how people classify and categorize the world around them.

Were beginning to see ethnoclassification in action on the social bookmarks site, and the photo sharing site Flickr. Both services encourage users to apply their own freely listed tags to content tags that others can then employ when looking for content. See a web page that looks interesting, but dont have time to read it? Post it to with a tag that will help you find it again.

Lets consider another classification challenge. When Im looking for documents on Adaptive Paths intranet (which I helped design), Im often frustrated because Im unable to uncover items that I know should be there. There are a number of reasons why picking topics from a pull-down menu is arduous, the topics we currently employ are not sufficient, and updating the tool with new topics is too time consuming. Productivity declines as we hunt for documents with cryptic filenames.

The primary benefit of free tagging is that we know the classification makes sense to users. It can also reveal terms that experts might have overlooked. Cameraphone and moblog are newborn words that are already among Flickrs most popular; such adoption speed is unheard of in typical classifications. For a content creator who is uploading information into such a system, being able to freely list subjects, instead of choosing from a pre-approved pick list, makes tagging content much easier. This, in turn, makes it more likely that users will take time to classify their contributions.

Microsoft’s Istanbul

AlwaysOn Network has an AP story:

Microsoft said it is launching a desktop application that aims to seamlessly integrate e-mail, instant messaging, video conferencing, traditional phone service and Internet-based calling.

Microsoft plans to debut the product, code-named “Istanbul,” sometime in the first half of 2005 to compete with efforts by rivals like IBM Corp. to link different channels of communications onto a single platform accessible from a computer.

Microsoft is testing Istanbul with corporate clients, and expects to run similar tests with consumers in the next few months before bringing the product to market, said Anoop Gupta, vice president of Microsoft’s real-time collaboration efforts.

The effort dovetails with Microsoft’s announcement in July of plans to offer business users interoperability between its instant messaging program and users of the other two leading services, AOL Instant Messenger and Yahoo! Messenger.

Istanbul will enable users to communicate with others regardless of which brand of instant messaging either party is using, said Gupta, who announced the plans at the VON Fall 2004 conference on voice-over-Internet phone technology.

Istanbul will feature so-called “rich presence” technology, using “buddy” lists to indicate which colleagues or friends are available for a range of communications rather than merely through the IM service itself.

Then, users could choose to immediately respond to an e-mail via instant messenger or another method such as voice-over-Internet within the same application, rather than switching back and forth between applications.

The product also includes capabilities to initiate impromptu, real-time, remote “meetings” by bringing together groups of people without having to arrange Web or phone conferences with dial-in numbers and passwords.

Ed Brill has a counterpoint.

Guide to Technical Trade Show

Don Marti has some tips:

1. Booth space is for suckers. Join a large vendor’s “partner program” and sweet-talk your way into their booth. This is especially easy if you make hardware — software companies don’t want to schlep and set up boxes. And hardware companies want to have some cool stuff on screen, component vendors want nifty boxes built with their parts, you get the idea.

3. Marketing’s job is done before the show. Leave them home and send at least two people: one big cheese (that’s you) and one sales engineer who knows the products inside out, can answer questions about them, and can fix them under pressure. Marketing people might look cute in a company shirt, but why fly and house someone just to have him or her say, “Hold on and I’ll find someone who can answer that for you.” Don’t try to do a trade show alone. Bring a consultant who works with your product if you have to.

6. Make your collateral work with the web, not against it. Bring lots of business cards and maybe useful doo-dads with your URL, but attendees don’t want to carry big glossy brochures any more.

8. Follow up! Don’t complain about all the money you’re wasting at trade shows when you’re letting all the business cards and lead printouts you bring back gather dust after the show. Take down key facts about prospects at the show, and ask for permission to follow up.

The Sales Learning Curve

Ross Mayfield blogs a talk by Mark Leslie:

This method of establishing a sales force is called the sales learning curve (SLC). It’s a concept adapted from the manufacturing learning curve (MLC), which is widely accepted in the manufacturing sector. The MLC states that the cost to produce the early units of a new product normally is high, but over time, as the production team learns how to optimize manufacturing and wring-out costs, volume increases and per-unit product costs decline sharply.

When we apply the MLC to sales, we come to the following conclusion: The time it takes to achieve cash flow breakeven is reasonably independent of sales force staffing. It is, instead, entirely dependent on how well and how quickly the entire organization learns what it takes to sell the product or service while incorporating customer feedback into the product itself. Because the entire organization has to come up to speed, hiring a large initial sales staff does not speed up the time to breakeven, it simply consumes cash more quickly…

Rather than starting with a large sales force, a company using the SLC is better served by hiring a small team of sales execs with the analytical skills and patience to lead the company through an iterative learning process that includes the continuous discovery and solution of small but crucial problems…

By adhering to the sales learning curve model of sales force staffing, you can safely toss out that old adage we started with. It doesn’t have to take longer and cost more than you planned.

TECH TALK: My Life as an Entrepreneur: Part 3

How did you learn how to set up a business? Are you self-educated? Did you have any formal schooling or a really good mentor?

I learnt business from two sources: my father and my failures. My father has been an entrepreneur since the early 1970s. Unlike me, he has done a wide variety of things from consulting to construction to owning a marble factory and an edible oils factory. While he never managed to build any into a really large business, it was very instructive to see him making efforts. He never gave up. He always believed in the power of knowledge. In 1983, he invested a significant portion of his earnings into buying a computer even though he has never learnt to use it! That computer changed my life from wanting to be a civil engineer like him, I decided then that I wanted to become a computer engineer. He always had subscriptions to most of the international magazines even though they were quite expensive.

When I left for the US, my father told me one thing that I never forgot and which governed my stay in the US: I came back in the mid-1960s from the US after my education and working for a couple of years. You should do the same. So, the decision to return to India was made even before I went to the US.

Even as I floundered in my early ventures, my father let me make the mistakes. He did give me advice, but never insisted that I follow it. His belief was that the only way I will learn is on my own. And that is why failures in my life have been the second teacher.

During the period 1992-94 immediately after my return to India, I did a lot of things wrong. It is actually hard to realise that one is doing things wrong. Only later did I realise my mistakes. But for me, that was my real education in the real world. At times, I have felt the lack of a formal management education, but then I feel that it is the real world which is the best teacher.

Even now, I have a more philosophical approach to failure. I accept failure and learning as part of the process. I also believe that there is a good reason why things happen even though the reason may not be obvious then. One has to always look at the brighter side of things as an entrepreneur.

Do you have a normal job, or do you survive solely off of the income derived from your business? If you have a job, would you ever consider leaving it to only work on your business?

I dont have a normal job. My entrepreneurial venture is my job! My only work experience was in the US at NYNEX.

In your business do you have any partners? If so, what do you think is the most effective way for partners to resolve conflicts and make good decisions?

At present, there are no partners in Netcore. I do have a strong management team, so most decisions are made in consultation with them.

My view on partners is that there has to be one person who has to have the final say in case of conflict. Equal partnerships can paralyse a business and decision-making. There has to be one person in charge.

Tomorrow: Part 4

Continue reading