Friday, September 29, 2006
IT professionals face the dichotomy between server sprawl and low asset utilization.
Between production, failover, disaster recovery, test, and development, many enterprises face 100% year over year server growth. For every server in production, another four or more are in place to support the deployment and development of a given application.
At the same time, server utilization remains anemic, often less than 15%. What drives such waste and inefficiency? Why the continued sprawl if existing resources remain available for consumption?
The problem lies in the inability to easily repurpose servers from one application to another. Repurposing is shackled by three constraints: 1) software constraints, 2) LAN constraints, and 3) SAN constraints.
Operating systems and applications are hard to change independent of powering down the server, servers are bound by their LAN IP addresses and require reconfiguring and often recabling to be available to other network resources, and servers, via HBAs, are hard-bound to certain LUN segments on the storage network.
Accordingly, it is often quicker, easier, and cheaper to simply add a net new server in the data center then it is to repurpose a server from one application use case to another.
Thankfully, Scalent, a Hummer Winblad portfolio company, provides a solution that frees servers from the three shackles noted above and allows IT to instantly repurpose existing assets - servers, LAN connectivity, and storage access - for alternate use.
Scalent virtualizes all the assets required to deploy a business system and by providing a virtualized abstraction eliminates the physical constraints to changing server A from running application B to application A on the fly.
Repurposing servers eliminates the driver behind server sprawl and allows IT organizations to dramatically increase utilization, and hence return on assets.
Scalent recently received an outstanding and in depth review from InfoWorld's Paul Venezia. The report outlines a set of detailed deployment use cases and provides strong validation of the approach and implementation.
Paul writes," Not many products truly deliver what they promise. Scalent, however, comes as close to keeping its pledge as anything I've seen. Scalent is attempting - and succeeding - at reaching the pinnacle of datacenter management: a truly adaptive infrastructure."
Salesforce, on the other hand, is a pure, read on-demand multi-tenant, SaaS play and enjoys a $4BN market cap, $472m in run rate revenue, and a 8.5x price/sales ratio.
As entrepreneurs architect start-up software companies, it is worth asking the following question: how much of Salesforce's 70% multiple premium is a function of the purity of their software delivery and pricing model?
When questioned why he supports so many delivery options, Rightnow's CEO, Greg Gianforte, answers that he sells the customers what they want. If they want on premise, fine. If they want their own instance of the application on-demand, fine. If the want multi-tenant on demand, fine.
While there is no question that being customer driven is a sound business trait, the complication arises when saying "yes" to customer demands introduces systematic weaknesses into your operating model. These weaknesses tend to frustrate the ability to realize economies in development, pricing, development, and sales force training.
Many of the bootstrapped start-ups that I meet with face this exact challenge - how can you say no to a customer when you need to make payroll? Why not agree to sell customer B what they want, even if it is inconsistent with what we sold customer A?
Where is the fine line between being customer driven and being a custom development shop building one-off products?
The economics of multi-tenant software are well understood:
- lower research and development costs (eliminate the need to support multiple code bases, custom patches and the need to port the software across multiple hardware and O/S stacks)
- lower support costs (eliminate on-premise one-off configuration complexities that complicate root cause analysis, eliminate need to support old versions of the product)
- lower sales costs (standard pricing and delivery options vs complex pricing lists)
At Hummer Winblad we are sympathetic to companies who perform "unnatural acts," ie deviations from their model, to win business. Teams, however, must be very careful that in pleasing customers they don't alienate investors who question the wisdom and sustainability of impure operating models.
All revenue is not created equally, and I posit that "good" revenue that reinforces efficiencies and the scalability trumps absolutely higher revenue. For start-ups, purity is a virtue worth aspiring to.
Monday, September 25, 2006
As an open web widget marketplace, Widgetbox serves the needs of both web widget developers and web personal publishers, including bloggers, web site and profile developers, participants in web auctions and others.
Widget developers using Widgetbox include the very small, independent developers and the very large such as Yahoo!, AOL, and eBay. At launch, Widgetbox is proud to announce partnerships with 38 of the web's leading companies, including Typepad, Meebo, and AOL Pictures. At launch, the directory includes over 290 widgets, a number that grows by the day due to a very popular developers' program.
Central to the power and simplicity of Widgetbox is the Widget Syndication Platform. This technology enables:
- Live Widgets: Widgets are always live within blogs and web pages; they can be re-configured instantly and without touching HTML code.
- Smart Blogs: Widgets can be “tag aware”, meaning a web publisher can make widgets react to the content of their web site. For example, an image widget might display images related to the content of the most recent blog post.
- Widget Panels: Drag and drop placement makes it easy to install and manage widgets within Widgetbox. See the Technorati widget on my blog as an example.
I believe that web widgets represent the lowest common denominator for the adoption and usage of web services. Widgets leverage the millions of dollars invested in web services, the power of syndication - ie the consumption of functionality "off domain," and the ease of use of Flickr.
Like AdSense, it is critical that web companies allow users to consume web services at locations of the consumers' choosing. Widgetbox is a powerful innovation in making that consumption easy and powerful - a rare combination.
For those of you who blog or maintain a web page, I encourage you to sign up and begin to enrich your users’ experience with relevant widgets that add to the mission of your site.
Wednesday, September 13, 2006
Peter Rip posted an interesting analysis of an emerging alternative asset model – the platform strategy, whereby a single firm offers LPs exposure to a full range of public and private equity asset classes – early, mid, late, PIPE, LBO, public, etc.
While it is an undeniable fact that platform strategies are on the rise – Carlyle, Pequot Capital, Farallon, etc – I am not a believer in the approach.
1) conglomerates are a discredited concept
a. history suggests that conglomerates do not in fact allocate capital more efficiently than capital markets
b. platform strategies leverage the very same arguments that conglomerates once did to justify their existence
c. investors perform better when they create return/risk appropriate portfolios than when they outsource portfolio construction to a conglomerate
2) platform strategies create an adverse selection problem
a. the best LPs will want pure play exposure to asset class segments and risk profiles
b. LPs who see value in abdicating portfolio selection are probably not the best nor brightest
3) Platform strategies create GP/LP alignment problems
a. While the accumulation of assets and the leverage of LP relationships across strategies clearly lines the pockets of the firm’s principals, it is unclear that larger, diversified funds lead to better risk/adjusted returns
b. Strategy drift in chase of larger platforms may negate the value in sustained focus and pure play execution in an area of one’s true competence
4) Incentives and management
a. How do you pay people for collaboration across strategies – how do you create systematic flow of information?
b. How do investors, typically not by nature great managers, manage the complexity of multiple markets, geographies, risk profiles, competencies…
5) Who has a bigger d*ck problems
a. Some strategies scale more than others
b. How do you avoid smaller strategies partners being marginalized by the larger strategy partners – see Apax early stage experience
6) Focus and decision making
a. Sourcing deals, syndication partners, service provider partners, terms, competition, etc all change as you move along the risk spectrum
b. The lack of common ground makes investing in ecosystem partners, deciding where and why to invest, and having your “partners” add value to your decision making process a real challenge
My own view is that the best LPs will prefer boutiques and will seek to build return maximizing portfolios across pure-play exposures to risk - this clearly is inverse to investors building platform companies predicated on LPs outsourcing their risk allocation to groups that will build that basket for them.
Tuesday, September 12, 2006
Optimizing click through rate is a critical goal and the end result is traffic that lands at your domain. Then what?
How many of you have searched on a company site for the main product, a product data sheet, an officer of the company, contact information and been able to find nothing? Let's take an example, sorry to pick on Cisco...Go to their web site and type in "wifi access point" into the search box. Hit enter and you will find this is the number one result.
Now Cisco spent $500m acquiring Linksys and I am fairly sure they would rather have customers get to a page about wifi products!!
Luckily, Baynote has an answer for you. Baynote recently received Inc 500's 2006 "Best of the Web for Smarter Searching" award.
Baynote was selected based on its on-demand Content Guidance offering and the company's success rate in helping website visitors reach their objectives by distilling visitor search and navigational behaviors into the Wisdom of Community. Using this Wisdom, Baynote dynamically adapts website search and website navigation for each user, vastly improving the conversion potential and usefulness of any business website.
Baynote customers on average realize more than a 20x increase in search-driven conversions of web visitors, turning viable prospects into leads on their websites. In addition, site navigation is streamlined significantly, with a reduction from 6 to 1 in the average number of clicks needed to find information and complete transactions.Now, if the market is going to spend $5bn driving traffic to their domains, why not spend incremental dollars to ensure that customers find what they are looking for!!
Congratulations to Jack Jia and the Baynote team and John Hummer for a great launch to date and for the award.
Monday, September 11, 2006
Widgetbox, a Hummer Winblad portfolio company, is the leader in creating a Web widget marketplace that provides widgets for use with blogs, social networks, auctions and web pages. Om Malik wrote a wonderful article on the web widget phenomenon and the company, which can be read on CNN here.
The company recently announced a widget contest, with the goal of identifying the most creative and useful web widgets. The prizes and ground rules follow:
The winning widget will be shown in the Widgetbox presentation at the DEMOfall conference the week of September 25. This presentation will be seen by journalists, VCs and many of the movers and shakers of the blogosphere.
The Four Runners Up
Each of the four runners up gets a week as the top Featured Widget on the Widgetbox front door, a Lego Mindstorms NXT kit, and a 100% genuine Widgetbox T-shirt. If you win, we'll ask you for your shipping address.
- All entries must be received by the end of the day on Wednesday, Sept. 20.
- You can continue to make changes to your widgets after you submit them.
- The winners will be privately notified via email on Friday, Sept 22. A public announcement will be made at DEMO.
Things that will influence the judges:
- Innovativeness. What we'd really love to see is a "we didn't know that was possible!" moment.
- Web 2.0-ness. Mashiness, thick clientosity, usability, beauty.
- Usefulness. For example, a screensaver widget probably won't make the cut.
- Widgetboxiness. Whether it shows off Widgetbox features such as Tag Awareness.
- Contestants must be 18 years or older.
- It is open internationally. Contestants do NOT have to be a US citizen.
- It is open to all widgets, even those that have already been registered on Widgetbox.
- Contestants may submit multiple widgets.
How To Enter
Send an email to firstname.lastname@example.org with the name of your widget. If you'd like to point out features of your widget in the email, go ahead.
We'll email you a confirmation that your submission has been received.
Who Are The Judges?
Every member of the Widgetbox staff will weigh in. We're going to have a big Judgment Day party, with pizza and veggie samosas, and stay as long as it takes.
Thursday, September 07, 2006
Do you ever feel a disconnect between the images on TV from Iraq and the daily reality of life here in the Valley, of the price of crude and the latest funding announcement, of the energy and optimism of the time with the stories of unfunded pensions and skyrocketing household debt?
With out sounding alarmist, what does the possibility of a real estate/consumer-debt driven recession mean for the sustainability of that disconnect? I recently sat down with a smart hedge fund investor who reeled off a series of disquieting statistics that suggested that the end of a debt-driven asset bubble was nigh.
He argued that consumer spending is the engine driving America's economic engine and that the engine is beginning to sputter. For example, economists believe that consumer spending accounts for two-thirds of current economic growth. The market hangs on the monthly consumer confidence index as a predictor of future economic activity.
The VC industry is also banking on the consumer with Internet, device, and semiconductor investment theses predicated on robust consumer spending activity. In my four years in the VC business, I watched the industry move completely away and then back towards the consumer. The question this post addresses is what are the implications of early warning signs of a slowdown in consumer spending activity, a fall in housing prices, and a growing crisis in consumer confidence? Also, if a slowdown does happen, it may pay to ask if the technology in one's portfolio is pro versus counter cyclical.
Market experts are beginning to question the sustainability of economic growth dependent on a consumer facing record high gas prices, household debt, and rising interest rates. The CCI fell from 107 to 99.6 from July to August, or by 7%. The market is beginning to punish companies exposed to consumer confidence and spending ability - Toll Brothers, a home builder, is down 46% from its 52 week high, Downey Financial, a mortgage lender, is down 15%, and Tiffany and Co is down 28%.
What is driving the stock market's concern about consumer-facing businesses. In a nutshell...consumer fatigue.
- From 2001-2004 median household debt grew 34%
- the household debt service ratio hit a record high in Q106 of 18% (ie. $18 of every $100 after-tax dollars goes to service debt), up 15% from Q199
- In 2005, real disposable incomes of private households in the United States increased $93.8 billion, or 1.2%, while their debts grew $1,208.6 billion, or 11.7%.
- Total consumer spending on goods, services and new housing accounted for 92% of real GDP growth
- average household debt grew to $90,000
- a large portion of consumer debt is set to reset in the coming few years
- 22% of the $8.7 trillion US mortgages are ARM based, with 40% of all new mortgages in 2005 being ARM based
- home sales are falling, inventories are rising, and prices are falling below appraised value
The truth of the matter is that I am not at all sure how to think about the data above. It seems clear that a cyclical shift in the economy is underway with consumer spending no longer a dependable engine of economic growth. Counter-cyclical stocks, Costco and WMT, will probably benefit from a shift in spending away from high-end stores.
Similarly, in IT and on the web it is conceivable that technologies that drive efficiency, reduce costs, and deliver WMT-type benefits to consumers (be they enterprise or consumer) will do well. Companies that target discretionary spending (ie vacation travel, consumer electronics, consumer finance) will most likely suffer.
On the web, we all are benefiting from an allocation of spend from offline to online. Will a recession accelerate that allocation - ie even if the total pie of dollars shrinks, will the hard ROI of internet marketing lead to an increase in absolute dollars spent on-line?
Will a recession accelerate the adoption of open source IT solutions, lower TCO and deployment technologies such as SaaS, and virtualization technologies that increase asset utilization?
Ie, will the recent sector bets of our industry, largely shaped by the last downturn in spending, prove to be prescient and counter-cyclical in that economic distress increases the value proposition of solutions that drive out excess margin and increase productivity?
Or will a slowdown not only reduce the total pie of available dollars but also retrench spending towards incumbent vendors and established business processes (think more not less of newspaper advertising and IBM).
If VC returns and start-up prospects are truly uncorrelated to the equity markets then these questions may be moot? However, if our companies and our exits are a function of the health of the US economy then it is worth thinking how prospective investments as well as portfolio companies will fare if the consumer spigot shuts down and the 2/3 engine of our economy feels the pinch of debt loads that crowd out discretionary spending.
To use two public examples...my personal opinion is that GOOG and VMWare (proxies for start-up related activity) will prove to be counter-cyclical. GOOG delivers auditable value and makes marketing a more scientific lever to create value and ROI. VMWare delivers more flexible IT environments, whereby utilization rises, cap ex is reduced, op ex is reduced, and return on assets goes up.
Despite the ominous storm clouds on the horizon, I believe that the best lessons of the last downturn were to focus on companies that deliver value for lower costs.
It will be interesting to watch how counter-cyclical these technologies prove to be and if an economic slowdown accelerates the rate of deployment and allocation of dollars. If we are wrong, it may be a rough couple of years.
Also, his new record, All These Roads, is now available.
Enjoy and hopefully see you tonight. If you cannot make it, check out his myspace page and/or the new record.