Software-as-a-Service - Salesforce.com

Technology Review
October 2005
The Customer is Sometimes Wrong

BACKGROUND:
1999, Oracle executive Marc Benioff founded Salesforce.com. The idea was to rent access to business software running on another company’s server. The important lesson may have been to stay with one’s convictions and go against the advice of investors and customers.

SYNOPSYS:
Salesforce.com is taking advantage of a new wave of software, software-as-a-service. Their idea was to sell CRM services over the web. The idea being that it’s amenable to the hosted model.

Salesforce.com was built to address the problems of Oracle’s customers in that they did not have expertise or the time to deal with constant upgrade cycles. Essentially, their customers only wanted to deal with the data, and not the backend, which many larger suppliers, Siebel and Oracle, were ignoring. Upgrade cycles often cost end customers weeks in lost data, as it never went smoothly. Old databases had to be ported and converted. The software was painful and unwieldy. Salesforce.com built a better uniform platform with the web as its backbone.

The biggest drawback was that customers were not initially willing to let go of control over their databases. Investors and customers suggested an in house option, but the executives decided against that option. Marc Benioff argued that the control issue was simply emotional and not grounded in logic.

It took pricing to really make customers seriously look into the Salesforce.com model. Through their model, they were able to provide pricing almost half that of traditional CRM software (including hardware, and staffing, on top of the license).

RESULTS:

• Customers increased 40% in 2004
• Y-o-Y revenues doubled with profit margins at 5%.
• +200% stock price increase from 2004-2005
• Siebel has joined them rather then try to beat them

Better Cancer Detection

Technology Review
October 2005
Better Cancer Detection

BACKGROUND:
Tools for screening and diagnosing cancer are crude at best. Research is focusing on more sensitive tests based on specific molecules called bio-markers.

SYNOPSYS:
In the field of bio-marker based cancer screening and diagnosis tools, protein based solutions dominate research. However, tools based on a phenomenon called DNA methylation are looking to reach market first.

DNA METHYLATION:
Methyl groups – carbon atoms surrounded by 3 hydrogen atoms each – attach to a gene without changing its sequence, but can alter a gene’s behavior.

• Aberrant patterns of methylation are involved in almost all types of cancer
• Occurs early on in the disease process.

What makes them interesting:

• Current methods – examination under a microscope – are inaccurate with a 70% success rate.
  
• Non-invasive - examines urine, saliva, and blood, rather than through biopsy.
  
• Available 2 years earlier than others

Where the challenge is:

• Not completely accurate.
• Good at determining that one doesn’t have a problem.
• Needs to be cheap and fast enough to make interesting.

The Starving Actor

Technology Review
September 2005
The Starving Actor

BACKGROUND:
TiVo launched in 1997 as a provider of digital video recording hardware and software. TiVo enabled users to continually capture incoming tv, and experience them on their own schedules.

SYNOPSYS:
Tivo launched in 1997, and it's founders thought they had a product to change the paradigm of television. Truly, they had a killer app that would turn its name into part of the American lexicon. To Tivo, is to record a show to be watched for later. Tivo's business plan was to obtain a dominant position by selling their software to equipment manufactures, selling their hardware directly to the consumer, get a subscription fee for their services, and selling ad time.

Early on, Tivo's advantage over its competitors was in its UI. However, they were unable to parlay this into an increasing customer base. It was difficult to market Tivo, as people seemed unwilling to pay for a $300 dollar box and an additional subscription fee to watch something they already had access to. By 1999, Tivo had already decided to stop selling the boxes. Tivo realized it needed access to the customer base of the service providers. In 2000, the largest service provider, Comcast, balked over a partnership because Tivo wouldn't budge on who owned the consumers. Tivo knew that without that data, they would have a difficult time signing up advertisers. It wasn't until 2005 that Tivo relented and was able to sign Comcast. But some believe it is too late for Tivo to spark the demand it could have in 2000.

Why they will FAIL:

• Poor execution. Tivo has not lived up to it's potential.
• Eroding customer base. Tivo's largest licensee, DirecTV, was bought out and it was announced they would go away from the Tivo software.
• Increased competition. Digital cable providers developing their own server based DVR.

Why they might SUCCEED:

• IP. Tivo has 85 DVR patents and 115 patents pending.
• New functionality. Downloadable internet content to the TV. TivoToGo mobile/portable television.

Commentary:
There's no denying the effect Tivo has had on television. It's unfortunate that they have been unable to capitalize on such successes. I think they're IP dominance, and the new functionality could be another breakout for Tivo. Combined with some sort of peer sharing network, that could be killer. However, I would not bet on this company. They have a history of being unable to execute.

The Willing Partner

Technology Review
July 2005
The Willing Partner

BACKGROUND:
Research In Motion’s (RIM) Blackberry handheld wireless email device was introduced in 1999. It became an important tool for increasingly mobile and pressed workforce. It succeeded where others failed by providing the entire solution. Not only was the device physically slick, RIM provided the software, the servers to push email from the wired network to the unwired network and the airtime to use it.

SYNOPSYS:
This article is about the evolving of a technology company in a burgeoning technology space (wireless e-mail) as it tries to maintain its market leadership. In June 2000, one and a half years after the initial release of the blackberry, RIM decided to allow the wireless carriers to sell the blackberry directly to the customer; removing the need to lease airtime from them. This move pushed sales of the devices from 360,000 in 2002, to 2.3 million in 2004. As the wireless networks evolved to handle both data and voice, with the introduction of GPRS (general packet radio service) and GSM networks, at 10 times the speed of previous systems, RIM found it challenged again. RIM needed to decide if it wanted to keep the software to itself, or make it open in an effort to turn it into the industry standard. In November 2002, RIM partnered with Nokia to design handhelds that could run blackberry software. By 2002, RIM no longer leased the minutes, and it was well on its way to not making the devices, RIM had become a middleware player. There is competition in the middleware space, but through these arrangements, RIM has given the big players a reason to keep them around.

Competitors in the Middleware space:

• Microsoft: Plans additions to its PocketPC operating system to allow for pushy email. Microsoft also plans on giving the push email servers with its server software.


• Good Technologies: 5,000 organizations use the rival GoodLink system.

The Results:

• 42,000 blackberry enterprise servers installed worldwide.


• 3,000,000 blackberry service subscribers.


• $57 million in licensing revenue, up from $5 million.


• Stock is up 800% since 2002.

Commentary:
RIM thinks it is safe in the middleware ground because of customer loyalty. They’re banking on the fact that their customers rave about the technology, and secondly that their customers dislike the other offerings. I think that is a weak argument. Companies will go with whatever is most profitable. I think the Microsoft offering will be a serious challenge to RIM’s dominance. The article is a good read if you are to read it in full.

CVS is out, Subversion is in

Red Hat Magazine
August 2005
CVS is out, Subversion is in

Synopsys:
In this months’ Red Hat magazine there is an article that from the title would suggest that it would compare the revisioning systems CVS and Subversion. The article claims that CVS is dead, and its newer, leaner cousin, Subversion, is light years better. Really the article can be summed up as such. Subversion has that familiar feel that CVS has, so switching to it should be relatively pain-free. The real power of Subversion is its ability to version control metadata, or file properties and history. For example, Subversion allows you to move around files in the repository, and still keep the version history in-tact. Overall, the article is a bunch of fluff, and not worth the read. I would like to have seen information on Subversion’s performance compared to CVS, and evidence of its much cleaner code base.

Subversion Pros:

• Familiar command line interface.


• The ability to version control file properties (metadata).


• The .svn directory, analogous to the CVS directory, is hidden and will not come up in listings.


• The ability to copy and move files in the repository.

Subversion Cons:

• After each commit, the entire tree must be updated to sync the repository and the working directory.

The Next Transistor

EETimes
08/15/05
The Next Transistor

SYNOPSYS:
The planar transistor is hitting its physical limits. The knobs which can be tweaked to ensure proper behavior are all close to their limits. As transistors get smaller and smaller there are a variety of issues that come up (leakage and performance loss). At dimensions below 65nm, we will probably find that planar transistor technologies will diverge; one performance sensitive, and the other power sensitive. At 45nm and 32nm technologies, novel techniques will be needed, such as depleted SOI / silicon-on-nothing, and multigate structures, to keep the planar transistor alive. Below 32nm, non-planar transistors may start getting some use.

Issues:
* Short-Channel effect: As you get farther away from the gate, the electric field used to turn off the gate is not sufficient. Leakage increases.

* Thin gate oxides: Tunnelling current between gate and channel increases. Leakage increases.

* Parasitic capacitances: coupling capacitance increases as elements move closer. Performance is lost.

Techniques to minimize issues:
* High-K dielectrics: Improve electric field strength at smaller dimensions.

* Strain engineering: The silicon structure is strained to improve carrier mobility (increasing drive current). Limits at 45nm (silicon fracture can occur)

* Depleted SOI / Silicon on nothing : Channel built on an insulator or on top of nothing to separate the channel from the substrate.

* Multi-gate devices: Channel surround by multiple gates to improve electric field. Difficult to manufacture.

Mobile Music

EETimes
08/15/05
Mobile Music

This article describes the music industry's approach to dealing with digital content. Their challenge being to "respect copyrights while embracing that technology is where [the music industry] needs to go."

Revenues from the sale of CDs have slumped by 2.5 billion since its peak in 1998. The music industry blames piracy for this slump, although one analyst believes that "some of the industry's decline is about the end of the CD-upgrade cycle." The article has some interesting statistics. In two years, the number of adults who admit to illegally d/l music dropped from 15 to 8 percent, while the number of young adults held steady at 31 percent. 31 percent!

To stem the piracy, the music industry is experimenting with several copyright technologies. Sony has released 2 CDs that only allow 3 backup copies and transfers of songs to 5 other devices. Their research shows that consumers accept copyright protection, but I know "accept" is a very loose term. Other research has shown that sales of copy-protected CDs might slump as much as 75 percent vs. unprotected CDs. So it seems as if the article points out that it is inconclusive how consumers will react.

The music industry is also experimenting with digital music sales, which have generated $350 million annually. This is compared to the $11 billion generated by CD sales, and it is pointed out that it will be sometime before digital music sales makes up for the $2.5 billion gap from the peak.

The environment is so murky, that progress in digital media will be slow at best. The recent Supreme Court case of Grokster vs. MGM studios only made things less clear. The case did nothing to "define the scope of existing legal liabilities for so-called vicarious and contributory copyright infringement. Instead it added a third possible liability: inducement."

One of the most exciting technologies for delivering digital music is through "over-the-air" downloads, downloading songs on cellular networks. Brad Duea, president of Napster, predicted that music-ready mobile phones will dwarf the sales of portable audio players this year. Analysts believe that the market for music-ready phones, will be in the low-end, flash based mp3 player market. Similar to what camera phones did to the low-end digital camera market. A study done on 6 you people, showed that only 1 would be willing to switch, as it eliminates the need to carry an iPod around. The others were concerned with poor sound quality and the lack of battery life.

My take on the article is that, it will take some time for a standard to be worked out, so in the mean time the best thing for me to do is get it free off the web.

Technically Productive, Technically On-line

I'm always interested in learning new things, but I simply don't have the bandwidth to keep up-to-date on all the things I'm interested in. If I were to try, I'd be reading Wall St. Journal, Harvard Business Review, Wired, EETimes, Technology Review, the latest white papers, Business Week, Time *pant* *pant* 24 hours a day 7 days a week. I'd never get to sleep!

This blog was started as a way to increase the productivity of us like minded individuals by pooling our intellectual resources.

The idea behind Technically Productive is for individuals to summarize the articles they read, distill the information, and freely offer it for the benefit of the community.

Another use for Technically Productive would be for individuals with technical problems to post them to this site for the collective brain to churn on.

My hope is to have a large contributing body from around the world, generating pages and pages of knowledge 24 hours a day, 7 days a week. How freaking cool would that be?!?