Posts from October 2003

Spam Filters (Continued)

Last week i slammed Postini for having a system that allowed spammers to route spam around them. Then later in the day i posted a short post that showed how they were able to stop this from happening. It’s not an issue with their architecture, its an issue with the way we set up their system (and the way a lot of other companies did too).

I have gained a lot of respect for Postini in this whole process. I think they may be the best spam filtering software out there. They are certainly the most popular with businesses and i have heard that after AOL, Yahoo!, and MSN, they are the fourth largest recipient of email in the world. That’s because the filter so much spam for ISPs and enterprises.

They have an incredibly hard job. I think its going to get harder, not easier, in the near future as spammers try ever more clever ways to get their mail through.

I had an email exchange with their CEO yesterday and i got a sense of how hard they are working to try to keep their customers happy. I told him that customer communication/service will most likely become the most important part of his business as the way customers use the filter may have to start changing over time in response to changing behavior of spammers. That’s what has happened in the virus world, and i see the same thing playing out in the spam world.

It’s really too bad that the great thing about the Internet – it’s openness – is also the thing that allows these nut jobs who send spam and make viruses to thrive.

Search

The impending IPO of Google, which is making noises all over Silicon Valley and Wall Street has gotten everyone thinking about search, including me.

There are a lot of people who think that digital media is a pull model, not a push model, and that search is the ultimate user interface for a pull media model. The data certainly bears that out. I just came from a meeting with a person who runs a large digital media business and i heard that a survey they did showed that search was the number one way that people access digital media.

My experience as a consumer of digital media bears that out. I use Google to search like most everyone else. I also use Yahoo! to search since its still my start page (because the My Yahoo service is still better than any other start page i’ve ever found). But even more importantly, when i use Amazon, i always start with a search. When i use Moviephone or Fandango, i always start with a search. Even when i go to the New York Times, i generally do a search after scanning the front page (and the New York Times’ search is the worst i’ve found on the internet for a major site). So I do a lot of searching when i am on the Internet.

I also use search as the primary way i access TV shows on my DVR. I use search as the primary way i access music on iTunes. And i use search as the primary way i access files on my laptop.

But i don’t totally buy the push model argument. You would always need to know exactly what you want to find in a pure pull model. So the push model where media companies have editors determine what’s likely to be of interest still has relevance to me. And i bet it does to most Internet users.

I also think we’ve only seen the beginning of what search can do. Google, which is arguably the best search on the Internet, is still pretty bad. If you are a librarian or know how to write really good search queries, you can get pretty decent results. But most people don’t. And the ecommerce aspects of Google still aren’t that great. If you type in a product that is available widely on the Internet, the first links you see are rarely to the most popular sites (Amazon, Walmart, Best Buy, etc). They are usually to some no name ecommerce site you’ve never used.

So a lot needs to be done in the search world. Maybe Google will continue to innovate and deliver better and more relevant search. Maybe Yahoo! will come screaming back and start to roll out stuff that’s better than Google. I am sure that’s their plan. But i wonder if someone else may come into this search business with something new and different and change the game. Amazon and eBay clearly have the opportunity to do more with ecommerce. And the traditional media companies, particularly the newspapers, could do something interesting if they wanted to take some big risks. Certainly, they have a lot to loose if they don’t. And maybe Microsoft has something up their sleeve that is game changing. They have done that a lot in the past when they were late to a market.

I also think social software (blogging, social networking, etc) has something to add to this mix. Self publishing is an important filter that could add something to the way users find relevance in the digital media world.

Audience management also has the possibility to add something. The more a search service knows about each user and what their behavior has been in the past, the easier it will be to deliver relevance to them.

So i think that search is a big deal. And that we are going to see a lot more innovation around search going forward. But i do not think search is the only thing that matters in a digital media world. I think editorial still matters. I think audience management matters. I think one to one communication via email, IM, or some other communications medium matters. And I think social software and self publishing matters. Which is to say that search isn’t the only thing i am thinking about. And i think it shouldn’t be the only thing on the Internet that wall street, silicon valley, the digital media companies, and VCs should be thinking about either.

Social Networking Software

This is another technology area I am thinking a lot about these days. It’s software that lets you connect to people and opportunities via your network of friends, colleagues, business relationships, etc. The first company that tried to build a business around social networking was called Six Degrees (for the Six Degrees of Seperation concept). It was started by Andrew Weinrich here in NYC in the mid 90s. It was always interesting to me but I never wanted to invest in it because I couldn’t see a commercial application.

Well there are a bunch of companies emerging that are taking Andrew’s basic concept of a social network and turning it into specific applications that do make a lot of sense.

There is Friendster which does this for online dating. You meet people who know people who know you. And that’s a good filter for a lot of people. As I understand it, Kleiner Perkins and Benchmark have or are about to invest in Friendster.

There is LinkedIn which does this for professional networking. You get to people you want to meet or build a business relationship with via other people you know. I have heard that Sequioa has or will shortly invest in LinkedIn.

There is Tribe.net, which i am an investor in and recently wrote about. Tribe applies social networking to the classified ad business (job listings, find a roomate, sell a car, etc). I think that’s an interesting opportunity and its run by my friend Mark Pincus, who is a great entrepreneur. I am sure Mark will find a way to make money in social networking with Tribe.

There are a bunch of other social networking software comapnies that are working on other interesting applications. In fact, there is even a whole weblog devoted to this market at Weblogs Inc.

The area I am most interested in is applying social networking software to enterprises. I like it for two reasons. First, there are a bunch of important enterprise applications that can be vastly improved with social networking software. Examples are sales force automation, customer relationship management, and human resources software. There are probably a bunch more. That sounds like a good market to me. And the second reason is that i think a tightly controlled social network may have more utility to its users. You can apply privacy, trust, rules, and controls on an enterprise network that you can’t apply to a public network. And that means that the participants in the network will be willing to share more of their knowledge and relationships without worrying about what bad stuff could happen. I think that leads to more utility and more usage.

I am aware of two companies going after this market, Spoke Software (in Silicon Valley) and Visible Path (here in NYC). I am going to learn more about their businesses. And I’d love any insight any of my readers can offer about this market and these companies or any other companies pursuing this opportunity.

Spam Filters (Continued)

I just found out that there is a fix to the Postini problem i wrote about this morning. I can program my firewall so that it doesn’t let any email through that doesn’t come from Postini. This is good news for Postini and any other “pass-through” spam filters. It’s a bit of a problem if Postini ever goes down, but its better than getting all the spam that’s figured how to route around Postini.

I still think it shows the lengths that spammers will go to in order to get their bogus mail through.

Spam Filters (Continued)

We use a spam filter for our email. It’s from a company called Postini. Its a “pass through” solution which means the spam filtering software runs on Postini’s servers, not ours.

Last month, in my post on Spam Filters, I complained about Positini.

Well, yesterday I found out why Postini was doing so poorly. Some spammers (a growing number) have found a way to route their mail around the Postini servers and directly into my mail box.

This is terrible news for Postini. And bad news for me too. Because it means I need to get a new spam filter.

Anyone who uses a spam filter that is ASP-based and doesn’t run directly on your mail server should check to see what’s going on with your email messages. If your message headers show the mail coming directly from the sender to you, instead of through your spam filter, then you’ve got the same problem.

This is another example of the spammers being one step ahead of the good guys (us and the tools we use to stop them). And it makes me wonder a little about my most recent post on eradicating spam.

Eradicating Spam

No I am not dreaming. Although I do dream about eradicating spam every day and night.

I get about 400 emails a day from senders i never want to get email from. That’s how i define spam. About 300 of them are caught by Postini, my spam filter. It’s not perfect. About 100 more spam messages get through. By spelling Viagra V-I-A-G-R-A. And by lots of other tricks that annoy the hell out of me.

I am an investor in two great email companies Bigfoot Interactive and Return Path. Both companies help legitimate commercial emailers who have permission to send emails do just that. But they are not growing as fast as they should because their customers are reluctant to spend more money on a channel – email – where four out of every five messages is garbage. And I don’t blame their customers. It’s a huge problem.

Seth Godin convinced me in 1996 that direct mail without stamps was a great business. And it is. But its a great business for a lot of jerks. And that’s what they are. Because they are violating Seth’s number one rule for direct mail without stamps – you must obtain permission in a world where sending costs are almost free.

But if the jerks could be stopped, we’d have the greatest marketing channel ever invented. I get Clear Channel sending me a weekly email on all the new concerts coming to NYC and when i can buy tickets. This is great for me and great for Clear Channel. I get ~10 emails a day from Jim Cramer telling me what stocks he is buying and selling. I get a daily email from the Italian Wine Merchants telling me about new wines that i might want to try. I read these emails. I act on these emails. And i am better off because of them. And the companies that send them to me are too. It’s Nirvana for direct marketing – the smart/pull marketing that works (spam is unfortunately the proof of that). Not the stupid/push marketing that doesn’t.

So what do we do about Spam? We fix it. I am convinced we’ll fix it in less than 12 months. There is too much money at stake. Bill Gates said the following at ETRE, according to David Kirkpatrick, “”Clearly e-mail should be verified. There are proposed standards for that.” He said Microsoft would be able to make spam “a truly minor problem.”

I know for a fact that all of the major ISPs and email providers are working on standards to verify the sender of commercial email. They will roll this stuff out in the next 6-12 months. And it will take time to have an effect. But it will clean up our inboxes. And that’s a great thing. Someday soon my dream will be a reality.

Reuters Knowledge

I like to invest in startup companies that provide software and data services to the wall street world. I’ve had a lot of success in deals like Alacra, TheStreet.com, and Multex, which was sold earlier this year to Reuters.

One of the reasons that Reuters bought Multex was they needed Multex’s data to feed a new product called Retuers Knowledge. Reuters Knowledge is a desktop application that will allow investment bankers and research analysts to get tons of information on public companies quickly and easily.

Last night, Reuters had a launch party for Reuters Knowledge, which they kindly invited me to. I have to say the Reuters Knowledge product is very slick. The user interface is excellent. It’s intuitive, easy to learn and use, and there is a ton of data available on public companies. I can’t compare it to Thomson ONE Banker which is the other major investment banking deskto product on the market today because i haven’t seen the Thomson product. It also competes to some degree with CapIQ, which does not have the comprehensive data that Reuters Knowledge and Thomson ONE has. But CapIQ has some innovative features like a networking tool that’s really useful and also has private company data.

But all that said, I’ve got some issues with Reuters Knowlege. Here’s what i think they got wrong;

1 – its too expensive. They want $700/month per desktop. That’s a ton of money when this information is available in other places and many of the banks have integrated a lot of data onto their intranets. I have to wonder if the dollars/desktop model is gone for good with the web and corparate intranets.

2 – its a desktop application. who wants to install and maintain desktop applications? its a total pain. its costly. and there isn’t much benefit to having a desktop app. Reuters claims that they need a thin client to provide some of the really nifty charting stuff. but BigCharts does most, if not all, of that via the web. i would think Java would work if they really need a client.

3 – its missing private company data. the fact is that private companies matter – a lot. if you can’t tell an investment banker or research analyst what’s happening in the entire market, including the private companies, they are still going to have to go elsewhere to finish their work. and the whole point of this product is to put all the information in one place. so that’s a big deal. and the real irony is that Reuters used to own a good private company database called Venture One. For some reason, that i can’t figure out, they sold that to Wicks. they’ll probably have to buy it back now.

So in summary, i think the battle for the desktops of investment bankers and research analysts has been joined. But i have to wonder if they are fighting the wrong war.

Dean vs. Clark (Continued)

Dick Morris (I know he’s got credibility issues) wrote an opinion piece in today’s NY Post called “Wes’ Losing Plan”.

If you’re interested in the summary, Dick suggests that Clark’s decision to blow off Iowa and downplay New Hampshire is a big mistake. He suggests that Clark is giving Dean the opportunity to win both of those primaries and emerge with big time momentum.

And, even more interesting to me, is Dick’s assertion that “cyber roots” has supplanted “grass roots” as the essential vitality of a “come out of nowhere” political campaign.

This takes me back to the point i made in one of my earlier Dean vs. Clark posts. Clark’s campaign has traditional political operatives, big time contributors coming out of the woodwork, and supposedly better poll numbers against Bush. But as Dick Morris points out, Clark lacks real backing. And Dean has used the Internet to build lots of real backing.

And now Clark is effectively giving away the first two primaries to Dean. Not a great way to build momentum. It makes me wonder the exact same question that Dick Morris ends with. Are Clark’s managers “like the generals of France who enter each war perfectly prepared to win the last one”?

Options

There is a very interesting discussion of stock options going on at VentureBlog, the excellent blog written by the guys at August Capital.

The post was initiated by a survey by Deloitte and Touche that found that 73% of public companies surveyed have either decreased or have plans to decrease the number of options they issue. And David Hornik’s concern was that reduced employee ownership levels would ultimately result in reduced innovation, which is clearly a bad thing.

I personally think this is a natural response to the overheated talent market of the late 90s when you had to pay exorbitant amounts of equity to retain and keep valuable employees. Now that the job market is back to normal, equity ownership levels will come back too. I think that is all that is happening.

I loved the comment made by Tubby Bartles that this isn’t something to get all that worked up about because markets are efficient. Tubby argues that the cost of options has always been something sophisticated investors have known about and have taken into account and that reducing the number of options issued now that they have to be expensed isn’t going to make a difference one way or another.

I agree with Tubby. First, cash flow is what matters. Expensing options won’t impact cash flow. Sure options impact the true ownership. But venture capital firms and sophisticated public investors have always looked at “fully diluted” ownership levels so they have been taking into account the dilution from options for a long time.

Markets are self correcting. The people who buy public stocks or buy companies (in other words the people who ultimately pay us for the ownership we buy or earn in startup companies) will figure out what the companies are really worth and won’t have too much trouble dealing with whatever the new accounting rules throw at them. And the people who run companies (CEOs and their Boards) will do whatever it takes to attract, retain, and reward the talented people who are needed to make their companies succeed.

That is capitalism at work. And that’s why its the greatest economic system ever invented.

Afibs

These are irregular heartbeats. My dad gets them. Genetics being what it is, i’ll probably get them. Jeff Jarvis wrote a nice description of them in his blog today in case you are interested in the subject.