Wednesday, February 28, 2007

Partying Like It's 2001

Startups seem to be sprouting everywhere these days, like mushrooms. Silicon Valley is back! If you feel like you're hearing about more interesting new startups these days than before, you're right. There's money sloshing around as VC investment in 2006 hit $25 billion, its highest level in five years. Silicon Valley got more VC money in 2006 than at any time since 2001. Then there's the more tangible effects for people who work in the valley.

There's jobs for the taking.
The number of jobs in Silicon Valley increased for the first time since 2001, with a net 33,000 increase in 2006.

Incomes are on the rise. Median household income rose 6.5% in 2006, after falling about a percent in the period 2001 - 2004.

If you're interested in exploring statistics about Silicon Valley, take a look at the 2007 Index of Silicon Valley, an excellent statistical summary of the area.

Sunday, February 18, 2007

Diagnosing Startup Problems

One of the classic issues that every startup is faced with is identifying causality in the face of uncertainty and noise. This is true for both positive and negative outcomes. For example:

  • A large bank just bought your software - great news! But can you isolate the specific reasons why they bought and make it repeatable? Or is it a special situation or relationship driven sale?
  • If sales are stalling, what's the issue? Is it the product, the market or sales execution? If the organization is heading towards dysfunction, this often leads to finger pointing and a revolving door among the executive team.
I recently came across one of the best posts I've ever read on the subject. Will Price of Hummer Winblad, has a post called "Isolating Causality: Bad Market or Bad Company." This should be required reading for entrepreneurs and employees in any early-stage startup.

Wednesday, February 14, 2007

Hardwired for the Web

I recently came across the wonderful "Hole in the Wall" project, run by Sugata Mitra, a computer scientist in New Delhi. PBS has an eight minute video segment you can watch here. It's well worth watching. Yes, the sociological implications are significant, but its also just a wonderful, heart-warming story of some impressive young kids realizing their unlimited potential.

Tuesday, February 13, 2007

The case of the missing address book

Tim O'Reilly blogs about the missing Web 2.0 address book. He says:

What really needs to be done is not just to connect the various social networks that do exist in internet network-of-networks style, but also to social-network enable our real social network apps: our IM, our email, our phone. Where, I keep asking vendors, is the Web 2.0 address book?
Spot-on Tim! Why is it that, in spite of all this emphasis on community and collaboration, managing my address book is still painful? The various elements of our communications tools need to be linked together effectively, and the best place for that link is in an address book. My address book is still largely an offline entity, despite some rudimentary efforts at linking it to my online world (e.g. the Outlook Toolbar from LinkedIn). I can import my Outlook contacts into Gmail, but I can't keep them in sync. Plaxo was a promising start to solving this problem, but they became yet another social network, not the linking mechanism between networks.

What should a web 2.0 address book look like? Here are some characteristics I think are key:
  • Always available. It goes without saying - I should have my address book available online and offline, on all my computers and mobile devices, and they should always be in sync. This has been an area of active investment, and solutions are getting better, but its still cumbersome to keep things in sync.
  • Easily portable. I own my address book, and no vendor should use their functionality, no matter how cool, for lock-in. Ideally the address book should store data in a human-readable format to begin with. If not, I should always be able to export into a variety of standard formats.
  • Proactive. Today's model requires my active participation for every contact that gets stored in my address book. A good address book should anticipate some of my needs and act for me. For example, if I've exchanged a couple of emails with someone who is not in my address book, it should try and find that person's contact info (e.g. from LinkedIn, or some other source) and add it to my address book (perhaps asking me for confirmation first). If someone updates an email address on orkut and that person is connected to me, my address book should know about it and act appropriately. Plaxo tried to do some of this, but their fatal flaw was that they were not neutral. Which brings me to my next point.
  • Neutral. The provider of this functionality must be neutral in the areas of social networking and collaboration. If LinkedIn provided this functionality they would attempt to lock my address book to their network. If Yahoo IM provided this functionality, they would likely restrict access to Gmail, and so on. The address book provider must not have a stake in the network, but only in the links between them.

  • Mediate trust relationships. I already have multiple places in which I've placed indications of trust. Anyone linked to me on LinkedIn list is a trusted relationship. Anyone on my Yahoo IM list is probably also a trusted relationship. Ditto for people I've had multiple email exchanges with. How can I aggregate this information in one place and manage trust centrally. One way in which this could be useful, for example, is that recommendation sites could use my trust information to appropriately weight the recommendations I got. would recommend stuff not only based on their collaboration filtering techniques (essentially statistical analysis) but also factor in my trusted relationships. If I was looking for an Italian restaurant in Palo Alto, I'd much rather hear what a few trusted people have to say about it than a mass of strangers.
  • Geo-aware. The contacts in my address book usually have location information stored in the record. Why is this information not used for a visual representation of my address book? For example, if I'm going to Chicago for a weekend trip, wouldn't it be great to map my friends in Chicago, so that I might get in touch with them while I'm there? And wouldn't it be even better if I could find restaurants in the vicinity so that we could meet for dinner at a mutually convenient location.
  • Presence-aware. My address book should have a notion of where I am, and take appropriate action based on my preferences. E.g. point out a good used-book store in the neighborhood (I'm always up for browsing in a good used-book store). Or alert me that a friend is nearby and I might want to get in touch.
How would this ideal address book provider make money? I leave that as an exercise for the reader. :-)

Wednesday, February 7, 2007

First there was HTML...

Evolution of the web in less than five minutes by an anthropologist at Kansas State. This is one for the ages...

(via John Battelle)

Hire talent, not skills

Following up on all the HR/recruiting related posts last week, I want to point you to a great article on the subject by Nick Corcodilos. It's a two-parter (Part 1, Part 2) but well worth the read. He makes the point that there is no talent shortage; there is a shortage of the management depth required to nurture and grow it. Weak managers hire skills, but strong managers hire talented people and teach specific skills.

I agree with the general point, and it has helped me crystallize some of the issues I see with the recruiting industry today. Weak management may be an issue, but a broader issue is that the recruiting pipeline has been set up to screen for skills, not talent. Matching algorithms are keyword-based - an indication of skill. Even if a manager is strong and understands the need to hire for talent, the process is stacked against her finding the right person. The need for talent over skills is only growing stronger as work becomes global and mobile, and rapid technology shifts require adaptability above all else. This is such a central issue in most people's lives: finding the right position in which to grow and thrive vs. a non-challenging job in order to collect a paycheck. Why is there no urgency around finding better solutions?

Monday, February 5, 2007

What's your blood commit?

There is a great article in the always informative Infectious Greed blog by Paul Kedrosky. The article talks about putting together a board package, but it provides great insights into the challenges of forecasting/planning at software startups. What metrics are appropriate to track? How can you create the right incentives for the field sales reps? How do you get visibility into the sales pipeline - what is real vs. what is wishful thinking. The underlying issue, apart from accurate forecasting, is about ironing out revenue lumpiness. Lumpy revenue is valued significantly less than smooth revenue and any planned exit needs a plan to decrease lumpiness.

The question of metrics is an interesting one and one that needs careful thought through in a start-up environment. I''ll save that issue for a future post, though, and just focus on the sales process. Here are some rules of thumb that I've seen work well at start-ups.

  • Hire sales reps familiar with how your product is bought, not how it works. What I mean by that is the specific domain knowledge of the sales rep is much less important that their familiarity with the selling process that your company requires. For example, lets say that you're selling a specific type of data security product, one that is useful for CFO's and the financial department. You have two possible sales reps in the hiring queue: one has sold a financial app to CFO's in his past job, and one has sold security software to the CIO in her past role. Who should you hire? You're much better off hiring the person who is familiar with your buyer - the guy who sold the financial app, not the woman who sold security software. Domain knowledge can be learned pretty easily during the course of a sales-kick-off meeting or two, but understanding the modus operandi of your buyer, their concerns, their quirks and sensibilities - that's invaluable and it can't be taught.
  • Fear is not a substitute for metrics. Often, sales leadership will compensate for a lack of metrics by ratcheting up the pressure on sales reps. The title of this post, "What's your blood commit?" is an actual question I've heard asked. I love the term: its very evocative of the near-desperation at the exec level when management doesn't have a good handle on forecast. While the intent is good (to understand what sort of discount factor to apply to over-optimistic forecasts), the approach is almost always counter-productive. Sales reps will respond to pressure by telling you what you want to hear. Use whatever management tool works for you, but it cannot be a substitute for hard numbers.
  • Be the ant, not the grasshopper. The old fable, about the importance of hard work and preparation, is very relevant in sales territory development. A common scenario with a new sales rep is that they come on board and start working their Rolodex. It may work for a couple of quarters, but once its done, like the ant, they find they have no pipeline to work with. Reps should be plugged into the marketing, lead generation and follow up process as soon as possible. This is the only sustainable way for them to develop their pipeline. This means that sales reps should allocate some time for cold-calling every week, develop local relationships such as those with industry trade groups and participate in trade shows and other events. These activities need to be built into their compensation plans.
What are some other sales lessons you've learned from your own start-up experiences?

Friday, February 2, 2007

HR Innovation: The Story So Far

I've written about some of the issues with HR/recruitment this week. How can we do a better job connecting people with their passion? I've seen a couple of innovative approaches to the solution, which I'll mention here. These approaches are not standardized or wide-spread, but at least it's an acknowledgment of the issue.
  • Align incentives ( The folks at took the referral model and extended it. The idea behind referrals is good - your current employees are a good gauge of what it takes to be successful. If they refer people like themselves (and we almost always do this) then odds are those people will also be successful. However, some people better embody the ideal characteristics of success than others, so Amazon recognized and rewarded those people who not only referred people, but referred people who went on to perform well at the company. In other words, if I referred person X, and person X went on to be successful at the company, I was recognized as having brought on a star performer. This created the right incentives, and in fact often led to mentoring relationships.
  • Statistical analysis (Google): If there is an algorithmic solution to this problem, we would expect Google to solve it! And in fact, that's exactly what they are trying to do. According to an article in the New York Times a few weeks ago, Google has recognized the shortcomings of the current approach to recruiting and is trying to change their recruiting practices to set up a deeper match between job and job seeker. They found that a factor they considered essential to success at the company - GPA - was actually not a good predictor of performance.
  • Psychological profiling (McKinsey). Anyone who has worked at a large consulting firm or dealt in group dynamics is familiar with psychological testing tools. The best-known example of this is the Myers-Briggs test. The test essentially tries to figure out how you express yourself (extrovert vs introvert), how you reason (analytical vs intuition) and how you organize yourself (planned vs improvised). There's no "good" or "bad" result on the test - just a better sense of you as a person. At McKinsey, employees generally take this test for the purpose of group dynamics - to ensure that a project team is well balanced. I've not come across any company using this in the recruiting process. Have you?
There have been so many innovations in technology and sociology that can be applied to the art and science of recruiting. It's time they were!

Thursday, February 1, 2007

Daily Product Fix: Zink

The only reason for this post is that I recently came across a fantastic product.

Inkless, portable printing in a beautifully designed device - how cool is that! The company responsible for it - Zink - was started by ex-Polaroid folks. Endgadget has a nice writeup here.