Wesabe Steps Out
Over a year ago when we first started thinking about personal financial management tools on the web, I had this OMG moment when I realized that a web based personal financial management service was going to be way more disruptive than just putting Quicken on the web. It was immediately obvious that the data buried inside everyone’s personal spending patterns could be an incredibly important asset. It was equally obvious that it was very personal and sensitive data and any company that wanted to aggregate that data would need to have a great deal of respect for their users privacy and for their ownership interest in that data.
Our portfolio company, Wesabe, made two announcements yesterday that demonstrate both how valuable this data can be and how much they respect their user’s rights to the data.
I’ll talk about the second announcement first. The Wesabe Value Engine is a turbocharged Tips tab that incorporates anonymous, aggregate consumer spending data to help you make better decisions. How often have you stood in line waiting to pay the auto dealer before picking up your car and wondered just how much more you are paying than you should be? All you really want to know is how much you are paying relative to everyone else in your neighborhood, but until now, there was no practical way to aggregate that information and make it available to consumers. Wesabe anonymously aggregates spending data from thousands of “neighbors” so you can now know that, on average, a service visit to the Weatherford BMW dealership costs $700 more than a visit to a visit to Bavarian Professionals and yet users rated Weatherford 17 on a scale of 100 compared to the independent shop’s rating of 96. Aggregating data on spending patterns and satisfaction for the hundreds of merchants and service providers we depend on in our everyday lives, and making that data available to consumers will empower consumers to make better choices with their money.
Looked at from a distance, some might say that we already have thousands of review sites on the web that let us rate everything from grocery stores to plumbers, so why is this fundamentally different? The difference between user contributed reviews and actual user spending data is in some ways obvious and in other ways subtle, but profound. On the obvious side, the first thing Joshua Schachter, the founder of del.icio.us said to me, when I mentioned the idea was “how cool – you can’t spam it”. There are lots of reasons why someone might slant a review, but how many folks would buy more shoes just to promote a shoe store. The subtle distinction is more interesting. Someone could give a fancy, expensive restaurant a five star review after visiting only once. That review will be helpful to some, but others might find it a lot more useful to know that the anonymous reviewer of the five star restaurant ate there only once, but visits the unpretentious Italian place down the street five or six times a month. Wesabe freely admits the tips are not perfect, but they are already useful in the many markets where the company has a concentration of users and it will get better and better as more users contribute more data.
I could go on and on about how valuable this data can be. There are a number of other in Wesabe’s post, and there will be more and more as the site grows, but I want to mention the other announcement they made yesterday, the Wesabe Automatic Uploader. This new service makes it possible for a user to input their credentials once and have their accounts automatically uploaded into Wesabe. Now, whenever a user logs into Wesabe, their account data will be there and up to date. Users have been asking for this level of convenience for a while. It has taken a long time because Wesabe’s commitment to user privacy and control made it impossible to use a third party aggregator. Only by investing the time and money to build their own aggregation technology could they meet the commitment outlined in their Data Bill of Rights and assure their users that no other provider would store a copy of their data and that if they ever chose to leave the Wesabe service, their data would leave with them. Building their own uploader also allows Wesabe to offer an API, so that the consumers who choose to contribute their data won’t be entirely dependent on Wesabe for innovative new services based on that data.
With the announcement of the new Wesabe Tips tab, the company has enabled consumers to anonymously share spending and satisfaction data, shifting forever the balance of power between merchants and consumers in favor of consumers. By waiting until they could offer the convenience of automatic uploading without compromising their users ownership of their data, they have reinforced their reputation as a trusted partner.
April 24, 2008, By Brad Burnham
Comments ()
| Tags: automatic engine uploader value wesabe
AB Meta
Over the last two years, our portfolio company Adaptive Blue has become known for its ability to automatically recognize things like books, movies, stocks, and wines on the web. By using semantic analysis to figure out the difference between The Da Vinci Code, the book and The Da Vinci Code, the movie, they are able to provide useful shortcuts that accelerate the experience of surfing the web.
Last night, Adaptive Blue announced an initiative that compliments their top down approach with a drop dead simple bottom’s up approach that can be used by any publisher or blogger who wants to make their content more discoverable and more useful. AB Meta a simple and open format for annotating pages about things. AB Meta complements existing approaches like RDF and Microformats and is eRDF (a flavor of RDF) compliant. Because the format annotates the header and does not require any modification to the page, it is light weight. Because it is less ambitious and not as general as existing formats, it is simpler and more accessible to a blogger who may be familiar with HTML but has never written a line of code.
The Web is changing. It is no longer just a playground for the digerati. As more and more people use the Web every day for everyday things, its power and limitations are both becoming more obvious. Adaptive Blue developed AB Meta in the hopes that, by making annotation simpler and more accessible to the growing group of publishers and bloggers on the Web, it will make the Web more valuable to users.
April 21, 2008, By Brad Burnham
Comments ()
| Tags: ab adaptive blue meta semantic
This is Nuts
Yesterday the New York Times ran an Op-Ed piece that led with the collapse of the effort to create a hand-held device for managing the 2010 census.
The latest problem is the Census Bureau’s failure— after nearly four years and almost $600 million — to develop a reliable hand-held computer system for counting millions of Americans who are not counted by mail. Census takers will now have to use far less accurate paper and pencil.
The full NYT Opinion piece is here. The AP story is here.
This is more than a technology problem. It's a colossal screw up. But there is an underlying technology problem. There is no easy way to create a purpose built device and integrate it into a new or existing process. The current method requires that the entire device be designed from scratch, all of the components or subsystems sourced anew for each new design. Finally you have to write custom software from scratch to stitch all the components together. It can take months to get a prototype to boot and years to integrate everything into a working product - electrical, mechanical, industrial engineering, manufacturing process engineering, QA, and support. And this is just the device. You then have to integrate that device into a business process and software applications environemnt. Half the time, by the time you are done, the process has changed and the technology embedded in the device is obsolete.
Contrast that to the way web services are built today. Start with the open source LAMP stack, modify slightly for your unique requirements, cut and paste a little Java script to mash up two or three other services on the web and than spend a couple of weeks hacking in a light weight scripting language like Ruby or PHP and presto you have a service that can serve hundreds of thousands of users. With a little more work it can support millions. The foundation of open source software and standard interfaces makes it much easier to create an innovative service and get it into the market quickly and cheaply.
Many of the folks who follow Bug Labs are really jazzed by the potential to scratch their own itch - to create a quirky device that meets a personal need. I wrote about my Bug here. It is great that Bug has captured the imagination of hardware hackers, but Bug is so much more than a modern day Heathkit.
Bugs goal is to create an architecture that would allow anyone (even the Census Bureau) to quickly build a device and integrate it into a service. I am not going to try to design a device for the Census Bureau here, but you could have fun mixing and matching a video camera, a GPS, a touch screen, WIFI and any number of other components to create a device. That, however, is the easy part. The hard part is making all of those components work together in an application. What if you wanted to be able to extract census data from a video interview? How would you pick out an address on a mailbox or a front door and cross check that address with the on board GPS, how could use voice recognition software to create a transcript of the interview and identify key elements of the census questionnaire and correctly populate the a database. Maybe this is a really dumb way to automate a census. That is not the point. The point is that the Bug Labs architecture lowers the cost of failure for the Census Bureau, not just because they can iterate cheaply as they define and then refine their device, but also because all the work they do on the device and on the software is reusable. So even if they end up deciding that their first hardware, software, and service implementation is fatally flawed, they can tweak a little here and there, evolving all three together until they get it right.
That makes a heck of a lot more sense than authorizing another $600mm and sending them off to try again the old way.
April 10, 2008, By Brad Burnham
Comments ()
| Tags: bug census failure hardware labs
I (May) Have a New Platform
Google has unveiled App Engine and unlike some other recent Google efforts this one feels fully baked and well thought-out. For starters it includes all the elements that I had identified in my post calling for a new platform, such as a highly scalable DB using a simple storage model that maps well to the common use cases and even support for email processing. They even provide a fully featured local development environment. There is a lot here to digest and there may be issues hidden in the terms of service but this is a major milestone in making cloud computing a reality and potentially in breaking down the scalability barrier for startups.
April 08, 2008, By Albert Wenger
Comments ()
| Tags: google platform technology
Covestor
Access to financial markets is fairly open. All it takes is a bit of money to set up a brokerage account and one can start buying stocks. Over the last decade the Internet has further opened up access by letting people trade from home (more recently even without commissions) and putting financial data, SEC filings and more at investors’ fingertips. The result of this accessibility of the markets is that many people try their hand at investing. Some of them turn out to be exceptionally good. Probably the most extreme example is of course Warren Buffett who has become the world’s wealthiest individual and accumulated much of his wealth by trading in the public market.
Most of us (myself included), however, are fairly lousy public market investors. For every great pick we make, we tend to have more than an offsetting number of stinkers. Of course in talking to friends who would mention anything but the winners? So how can one tell if one is talking to the next Warren Buffet or simply someone grossly exaggerating their actual track record? Thanks to Covestor this is now easy. Covestor publishes the returns achieved by individual investors based on actual trades in their brokerage accounts. So next time a friend brags about a great trade they made, just tell them that if they are really that good they should go ahead and prove it on Covestor.
Creating a publicly verifiable investment track record with Covestor is really easy. After signing up for an account with Covestor, all one has to do is provide login credentials for one’s brokerage account (Covestor takes extensive security measures to protect those credentials in transit, does not store them and has only read access to account information). Covestor does all the rest by automatically extracting holdings and trade data from the account. Covestor then does all the number crunching necessary to let other investors judge the track record, for instance by graphing it, comparing it to benchmarks and examining stats such as the Sharpe ratio.
One might ask why anybody would want to share their great investment ideas with the world in the form of a detailed track record. Probably some are doing it simply as a competitive sport. Many participate because Covestor offers a vibrant community of active investors who learn from each other. Others, however, do it because they realize the value of an established and verifiable track record. For now they would have to “monetize” this track record elsewhere, in the same way that great coders don’t get paid for their contributions to open source projects, but the recognition of the community can translate into a promotion or a paid speaking engagement or book contract. We are excited about our investment in Covestor because of the possibility for going much further by creating a platform to enable talented individual investors to compete directly in the trillion dollar market for money management. Instead of having to join an existing firm or go through the difficult and expensive process of trying to form their own firm, Covestor will handle asset management, and these individual investors will be able to concentrate on picking winning investments. We are convinced that opening up the asset management market to talented individuals will be as fundamentally disruptive to this huge industry as blogging has been for journalism (what would Mike Arrrington’s or Om Malik’s influence be if they were working for the San Jose Mercury News?).
Covestor will use the funding to build out the “asset management” side of its platform which will permit the successful investors to collect fees. By doing so, Covestor will fall squarely in the category of new marketplaces that we believe will at least partially supplant what currently happens inside of firms.
We are excited to be investing in Covestor together with Spark Capital and Amadeus Capital. We look forward to working with them and the Covestor founders Rikki Tahta, Perry Blacher, and Simon Veingard.
April 06, 2008, By Albert Wenger
Comments ()
| Tags: