Last night I slept incredibly well. Why? Because my youngest son and I were sleeping on our sailboat (hence also no post this morning as we had breakfast together). The cabin is tiny, the cushions hard, the sleeping bags hot. Still slept like a baby (as did he). My theory is that was due to not having blinking crazy colored lights everywhere. At home, it seems that every device these days comes with one or more LEDs to announce its presence. Between a couple of laptops, some cell phones, some cordless phones, several powerstrips, a TV and a cablebox our bedroom looks a bit like Times Square. Now I realize that the idea behind (at least some) of the LEDs is to provide some status indication (new email received, battery charging, etc). But do they really have to be so insanely bright? And could they not add a tiny switch or software configuration to turn them off? I am going to start applying duct tape!
The top link on techmeme right now is yesterday’s post by Jason Calacanis titled “Yahoo committed seppukku today.” It is worth reading because it powerfully (and hilariously) lays out the position that handing over search to Microsoft is a huge mistake for Yahoo. It is a position I am very familiar with, because I made much of the same argument (although in a less entertaining fashion) last May. But it is more than a year later and I am not sure the logic still applies.
First, Yahoo has lost a ton of talent along the way. Some of that talent went to Microsoft. Most notably, Microsoft recruited Qi Lu, one of Yahoo’s key search engineers in December of last year.
Second, despite some slight gains in search share (from hovering just above 20% for most of 2008 to slightly above 21% in early parts of 2009), Yahoo’s search monetization has been falling. In Q2 they reported a 15% decline in search ad revenues at a time when Google reported a 3% increase.
Third, search may simply not be a winnable category for some time for anyone other than the leader. It is worth remembering that the arms race with the US was a key contributing factor in the demise of the Soviet Union. So if you are Yahoo and you are behind and slipping (see first two points), do you really want to keep plowing money into search? Might be smarter to let someone with deeper pockets wage the fight on your behalf and wait it out.
Fourth, the economics of the deal with Microsoft might be attractive. On one hand, there was no upfront, but on the other there are a number of guarantees, such as “Microsoft will pay traffic acquisition costs (TAC) to Yahoo! at an initial rate of 88 percent of search revenue generated on Yahoo!’s O&O sites during the first five years of the agreement” and “Microsoft will guarantee Yahoo!’s O&O revenue per search (RPS) in each country for the first 18 months following initial implementation in that country.” Both of these are difficult to interpret since not enough detail is provided, but it sounds intriguing.
Bottomline is that in May 2008 I shared the same point of view that Jason published yesterday. But at the end of July 2009 a lot has happened and now I am thinking this could actually turn out to be a smart move. The initial reaction clearly suggested the stock market did not think so. But this seems more like a knee jerk reaction to me. I think it will be a couple of years before one can properly judge this decision.
I have been participating in a series of meetings on how to ensure New York City’s future as a (new) media hub. A persistent theme of these meetings has been that there is not enough external awareness of how many great companies already exist in the city, including many that do some fairly hefty engineering! There is still too much of a perception that New York is good at content but not at innovating in delivery or anything else that requires some heavier technical lifting. At yesterday’s meeting, I proposed that it would cost the city next to nothing to lift its profile here. All that would have to happen is for Mayor Bloomberg to do a tech tour where every week he visits an existing tech business in the city for 10 minutes. This could even be enroute to another meeting. Just get someone with a video camera to be there and then put it on blip.tv and distribute it out to the world! Someone at the meeting offered that if the community put together a list, they would bring the idea to the mayor. I agreed to help with that. Suggestions for companies to add to this list via comments to this post would be great. Of course feedback on the idea itself is welcome also!
I have been mulling over the ferocious rumors of an Apple Tablet (sorry, no links as I am on BB, but if you missed these - congrats, you were off the grid!). Really not sure what to think about it. The only device that I would consider carrying regularly outside the house in addition to a smartphone would be an eInk style reader. The reason is simple: weight. A traditional display uses a lot of power and that requires a big battery which makes the whole device heavy (and requires remembering to charge yet another device). So that suggests I would use an Apple tablet primarily in my home, unless they pull off something amazing (and I am not ruling that out). At home, however, I am hard pressed to come up with a lot of (any?) good use cases. I already have my laptop with me at home and as a fast typer really love having a keyboard. So will be interesting to see what Apple comes up with here - if anything!
I just ran across an article titled “Skilled Immigrants on Why They are Leaving the U.S.” in BusinessWeek (via Techmeme). It is not surprising to find diminished job prospects in the U.S. as a key reason, but I was shocked to read that the wait time for a Green Card for Indian and Chinese citizens is now 10 years (that’s a decade!). That is a very long time for anybody to wait and live with the uncertainties faced by visa holders. Just entering the country when you are here on a visa can be a challenge, especially if you are coming in through a major airport. And that was true even before 9/11/2001 when I was flying frequently into JFK and BOS on a student visa.
It would be a mistake to think of this problem as a nuisance to handful of foreigners. For starters, according to this article from earlier in the year, the total number of skilled professionals (doctors, engineers, etc) waiting for a Green Card had already reached 1 million in 2006 (wonder what it is now!). More importantly though is that immigration has been a key aspect of entrepreneurial and high tech activity in the US. From the same article:
Despite the fact that they constitute only 12% of the U.S. population, immigrants have started 52% of Silicon Valley’s technology companies and contributed to more than 25% of our global patents. They make up 24% of the U.S. science and engineering workforce holding bachelor’s degrees and 47% of science and engineering workers who have PhDs. Immigrants have co-founded firms such as Google (GOOG), Intel (INTC), eBay (EBAY), and Yahoo! (YHOO).
I was very lucky, in 1997 — just as I was starting my first company — I married Susan Danziger and received a Green Card shortly thereafter. I have since become a US Citizen and as such now sincerely hope that we can achieve some kind of immigration reform that will stop this brain drain. With every one who returns because they are tired of waiting we are potentially losing a chance at the next breakthrough company (which might very well be targeting the Indian or Chinese markets!).
I am a fan of using puzzles in interviews, especially for developers. But it is important to do it right and to draw the right conclusions. Here are some things that I have learned over the years.
1. Puzzles should be just one of many different parts of the interview.
2. Try to find a set of puzzles that works for you and use the same puzzles with many candidates to get better comparability.
3. Focus primarily on approach and behavior and only secondarily on outcome.
4. Unless the position specifically calls for ability to think under stress (e.g. certain ops positions), try to put the candidate at ease. Being in an interview is stressful for many folks.
5. Ask for a vocalization or visualization of the thought process.
So what are some of the things that I believe can be learned from puzzles?
- Is someone naturally curious?
- Do they think before talking?
- Can they explain how they are thinking about a problem? Verbally? Visually?
- Does someone get easily frustrated?
- Do they ask for help if they get stuck?
- How does someone react to getting an answer wrong? Give up or persist?
If applied right, I believe puzzles can provide a lot of insight along these lines into how candidates will approach problems they encounter. I was reminded of all of this by the Amazon booth at OSCON. Every day they provided a new code puzzle. I enjoyed solving them, but possibly even more enjoyed observing the reactions and approaches of others!
I am sorry to see Amazon buy Zappos. Not because I have bought a lot of shoes at Zappos (still liked this as a title for the post though), but here are three reasons:
First, I felt that Zappos could be a real competitor to Amazon for online retail. And I love competition. It is great not just for customers but also for suppliers.
Second, Zappos was an IPO-able company based on size and apparently profitability. It takes IPOs to build confidence in the financial markets. IPOs beget IPOs. Now we have to scratch one candidate off that list.
Third, I am an Amazon shareholder and don’t think this was a smart move (as so often, the market disagrees with me as AMZN is up 5%). This reminds me of eBay buying half.com and then never really figuring out what to do about the two brands.
There is a rumor going around that Tony Hsieh did not want to sell Zappos and that this was a Sequoia Capital move. Bijan points out, it’s not so easy for an investor to sell a company unless management wants to go along. On the other hand, it may also not be easy to IPO a company if the VC doesn’t want to go along. Be that as it may, I sure would have love to see Zappos stay independent and go public.
To us, this appears to be one of the great constants of the web. It is taking power away from existing large institutions and pushing it out to smaller entities and often all the way to individuals.
In the case of Covestor the institutions in question are large investment management firms and the individuals are people who are investing their own money. In its first phase, Covestor made it possible was for individuals to publish their verified track records to the world through the covestor.com site.
We are excited that Covestor is now launching its second phase with Covestor Investment Management (CVIM), which offers the world’s first Multi Managed Account (MMA).
Here is how it works in a nutshell: You take some money and put it in the MMA. CVIM then allows you to subscribe to models representing different individuals. As those individuals trade in their own accounts, the trades are replicated in your MMA (appropriately scaled). CVIM ensures that you can only subscribe to models that fit your risk profile and permits fine-grained control, such as excluding trades in the stock of the company you work for. Since the MMA is based on a regular brokerage account, you can see your actual positions at all times and CVIM lets you track your performance for each model.
Unlike traditional asset management firms, CVIM is an open and transparent platform that brings together model managers (individuals executing trades with their own money) and investors who want to follow those strategies. As an investor you know at all times what your positions are and which trades were executed as a result of which model. Your assets always stay in your own account. As CVIM adds more model managers to the platform, your choice increases and you can pick which models to subscribe to based on the performance of those models, knowing that they represent trades the model manager is making with his or her own money.
A big congrats to the Covestor team, which has worked incredibly hard and has overcome numerous technical and regulatory hurdles to bring the CVIM platform to market. Now to find out more and sign up for your own MMA, visit the new CVIM website.
I believe that manned space exploration is essential. There I have said it. Not irrational, not an expensive gimmick, but essential. Why? Some day, leaving Earth may be the only way the human species can go on. Let’s hope that’s not the case, but hope is no substitute for preparation. Species go extinct all the time as their habitat changes dramatically or disappears altogether (these days often due to human impact), but we are the only species that can pro-actively explore new habitats.
The objection that money spent on manned space exploration would be better spent on protecting the Earth is flawed. This is not an either or decision. There are many other things we could cut if we really had to and the sums involved are small compared to those in the financial bailout that is just taking place.
There is also the objection that space exploration would make us appreciate Earth less. That has always struck me as odd. Is there anyone not moved by pictures of Earth from Space? My own appreciation for how marvelous the human brain is has only been increased by our attempts to create machine intelligence. Similarly, our struggles and certain setbacks in space exploration would likely highlight just how amazing a place Earth is.
Then there are the objections that space exploration is simply impossible due to the vast distances and technical/physical problems such as shielding. But nobody (in their right mind) is suggesting that we immediately try to travel beyond our solar system. Mars, however, seems eminently within reach. We have also been making amazing progress in manufacturing technologies (e.g., 3D printing) that make it easier to “live off the land” once we get there. Mars with its lower gravity would make a great basecamp for further expeditions once we figure out how to have a permanent base there.
So 40 years after astronauts stepped on the moon — let’s boldly go where no human has gone before. On to Mars!
So I am late to this party (because I don’t blog on weekends), but James Altucher’s column “The Internet Is Dead (As An Investment)” is just too tempting an invitation for rebuttal. Fred already provided his take here, stating our (at USV) belief in the transformative nature of the Internet and taking on Altucher’s comments about business models. My own point of departure is this quote from the column:
The days of infinite margins, 1,000% productivity gains, and growth of market throughout the universe are long over. Internet companies now should be treated, at best, like utility companies that get bought at about 10 times earnings and sold at 13 times earnings. Even then, I’m not sure I would give the Internet sector the same respect as the monopoly-protected utility sector.
Let me take the points here one by one:
- The “infinite margins” are of course hyperbole, but we see companies in our portfolio that have gross margins above 80% and EBITDA margins of 30% and better. This is a result of having mostly fixed cost of operations with virtually zero marginal cost for additional users or customers.
- The productivity gains misses the mark as most interesting Internet business do something entirely new (e.g., Facebook) as opposed to being more productive at something existing.
- Growth of the market is still very much happening. There are now about 1 Billion folks on the Internet and that is likely to double over the next decade. There are still way more ad dollars offline than online relative to where attention is actually spent.
- The utility analogy applies only to a few companies that operate at the plumbing level of the Internet where there is in fact fierce competition and businesses are capital intensive. For instance, the CDN market has proven to be a tough place to make money. But again, the interesting businesses, such as Amazon, operate primarily at the application level. They deliver value directly to the enduser and are beneficiaries of the competition at the utility level.
- To be regulated as a (local) monopoly is a highly restrictive form of profit protection and one of the reasons why utilities trade at those multiples. For Internet businesses there are dynamics that can allow for both rapid growth and protected profits. The most powerful of these are network effects and data assets, which can provide not only sustainable competitive advantage but often have a winner-take-all effect as well (in which the number 1 in any area is much bigger and much more profitable than the rest of the field).
Now I come to think of it — I hope that many readers will follow Altucher’s advice. I sure hope that I get another opportunity to buy more Internet stocks than I already own. So please ignore all of the above and dump your holdings, especially GOOG and AMZN.
Much of the discussion of the real-time web has focused on content generated by humans, such as tweets. But the bigger growth over the next decade is likely to come from m2m interactions. One big source of data will be smart meters. For instance, the UK government recently unveiled a plan to equip every home with a smart meter by 2020. The real-time electricity usage data from these meters can be used for better management of consumption by individuals, companies and utilities. In addition to price, thanks to our portfolio company AMEE, those decisions can now also be based on the real-time carbon intensity of the UK grid. It will be interesting to see how quickly smart metering takes off in the US. Both Google with PowerMeter and Microsoft with Hohm are already building consumer facing services that can be fed with smart meter data. At the beginning of this year, fewer than 5% of meters in the US were smart meters. So there is room for some very fast growth over the next couple of years.
Apparently, Twitter had a lot of confidential documents stolen via unauthorized access to gmail and google docs (writing on BB, so no links). This brings the security of cloud computing / web apps very close to home, especially as we are contemplating moving all of USV to gmail and google docs. The threat of access by a third party increases exponentially with the move to the cloud, because the machines that now contain the documents and the links to those documents (as sent by email) are accessible to the Internet at large. With anybody with an Internet connection potentially being able to access, a simple Username/password scheme is clearly insufficient for authentication. This is especially true given password reset mechanisms based on “canned” questions with easily guessed answers. So here is a modest proposal. Give users the option to secure with a second factor. Two ideas come to mind (not novel - just saying now is the time to get serious about these). The first is SMS. Just enter your cell phone number during registration to enable the second factor. As you log in with username and password you receive an SMS with a code that you need to enter also. This will admittedly slow things down a bit and might be a total nuisance if you are on a plane, but it is a nearly universal solution. The second idea is simply a twist on the first one. Instead of SMS, use an app downloaded to the phone. The app would display the second factor on the phone to be entered along with the password. The app strategy might be a way to get back to what seemed like a promising idea from early web days: client side certificates. Instead of the cert being in the browser it would now be on the phone. To log into a web app you fire up your phone app, which talks to the server and gets you a secure one time password. I am hoping that nothing worse than the Twitter breach has to happen before providers such a Google and Microsoft will offer stronger authentication as an option.
This morning in scanning Techmeme while having my early morning cup of coffee, I noticed two separate items about TLDs. First up, a post from Techcrunch on how .cm domains are opening up for pre-order. Second, an entry from Slate on how ICANN is planning to further expand the available TLDs. Last year, I wrote that I believe “New TLDs Are Highway Robbery” — they are simply a tax on the system.
To some extent search is solving this problem as few people these days type domain names directly into their browser bar. So one might argue that companies should not worry about typos or squatters, but that ignores other ways in which extra TLDs could be manipulated. For instance, there is the problem of trademark dilution — if you let folks register your domain in other TLDs you are may be reducing your ability to enforce your trademark. There is also the issue of phishing. Additional TLDs make it easier to create links that look like the real thing but go to a phishing site.
So, ICANN, let’s please stick with the TLDs we already have!
For a while it seemed that Android would be the Google OS, especially with new Android devices such as netbooks appearing on the market. But as I recently wrote, Android seems native app centric at a time when Google is showing with Wave that the Web can be the app (see http://continuations.com/post/133482897). Now Google has taken the logical step of putting a fast-booting version of Linux under its Chrome browser to create a complete OS (http://bit.ly/1ae7vz). While their own post tries to explain how this has only some overlap with Android, it is pretty clear that the two will compete instantly in the netbook market and from there it’s only a small step to smartphones. I am impressed by Google’s willingness to let internal groups with different approaches but similar targets compete like that. Something that Microsoft never did well and now they are being outflanked in two key future markets — netbooks and smartphones — as a result (as an aside: when will the EU competition commission finally wake up to the fact that they are a decade late wrt MSFT?).
Andrew Parker and I recently wrote a post on the USV blog about “The Mobile Challenge” — looking for apps that make use of the unique capabilities of the latest (smart)phones to deliver experiences that were previously impossible. Since writing that post, I have become fascinated with augmented reality apps, such as Wikitude by Mobilizy (Austria), Layar by sprxmobile (Netherlands) and Sekai Camera by Tonchidot (Japan).
I have no idea how well any of these actually perform, having only experienced them through Youtube videos, but they sure look like they might provide exactly the kind of unique experience we are looking for. This leads me to two important questions: Has anyone actually tried one of these apps? Also, are there any US-based companies working on augmented reality apps?
A lot of the work of a good board takes place outside of the board meetings (see Tip #1), but the meetings are critical to overall effectiveness. It is the time when the team has the (mostly) undivided attention of the board members and when important decisions will be made (or at least influenced).
1. Schedule your board meetings well in advance, ideally for the entire next year. Most VCs are on multiple boards and have hectic schedules and the longer you wait, the harder it gets to have everyone attend. Insist on in-person attendance and don’t try to reschedule last minute (for large boards that is true even if it is at the request of a single board member, unless there is a super compelling reason to do so).
2. If you are an early stage startup, monthly board meetings are a good idea. A lof of stuff can and does happen in 4 weeks. Also, it often takes up to three board meetings for a major issue to go through the cycle of being noticed (board meeting 1), options being evaluated (board meeting 2), and a final decision being made (board meeting 3).
3. Start with a good board package. See Tip #2 for what should go into it and how to prepare it. If you followed the suggestions from that post, try to stick to the package and don’t let a director highjack the meeting. After all, you will have picked a big topic that you want to get through. So if someone wants to go off in a totally different direction, see if you can enlist the other board members in returning to plan. If everyone wants to go to the other topic then that will tell you that it is more important than the one you had picked.
4. Have team members who are not board members present. It is highly motivating to have direct exposure to the board. As the team grows, you may want to rotate folks through depending on the topic at hand instead of always having every team member present their area (meetings can run on too long if you do that).
5. Actively solicit input on decisions from every board member. Some people are oddly quiet in group meetings. Don’t assume that means they are in agreement with everything that is being said. Yes, it should be the directors responsibility to speak up but that doesn’t mean they will. So don’t be afraid to call on a director and ask them what they think. That is especially true for anybody dialing into a meeting.
6. Don’t be defensive when a board member makes a suggestion. You run the company and you cannot/shouldn’t do everything one board member says, but you should always listen to it and take it as one input to your decision making. Of course, if the entire board decides something, then it means you have to really work towards that (or you may find the board executing its ultimate responsibility).
7. Leave enough time for formal business so that you don’t need to rush through things like option grants. These may require a bit of discussion, especially early on before you have bands established. As the company matures and the board grows some of these should be discussed by a separate compensation committee and then presented to the larger board.
8. Encourage your board to occasionally hold an “executive session,” which despite the name means a board member only session without the management team and founders. You should assume that board members talk to each other outside of board meetings, but this facilitates an independent discussion of the your and the team’s performance, which is a critical role of the board.
9. Provide refreshments and have one or several short bio / phone / email breaks. If you announce these ahead of time, you may even succeed in reducing blackberry usage.
You will generally know if you had a productive board meeting or not, but it never hurts to ask for feedback. Often small changes can result in vastly improved meetings as no list of tips can account for the idiosyncratic preferences of your particular board.
The thought of putting together a board package seems to instill a lot of apprehension in founders/management of startups. That’s a shame, because — if done right — the board package will create hardly any additional work and will be an incredibly useful tool not just for making the board effective but also for managing the company. Here are some suggestions that may help:
1. Choose a format and order and stick with it. Good to start with a “what’s working and what is not”, together with or followed by KPIs.
2. Pick no more than one big topic to go into detail on. This could be the biggest problem you face, the biggest opportunity, etc. Trying to cram too much into one package virtually guarantees it won’t get enough attention.
3. Have each of your key management team members contribute a section to the board package on their area. They too should pick a format and try to stick to it.
4. The KPIs and team member sections and possibly even the “what is working and what’s not” should be part of your day-to-day management process. The only extra time for board package should be putting these into a single file.
5. There are no points for production value. These should be working documents not Powerpoint Extravaganzas.
6. Send the package out about between 48 and 24 hours before the board meeting. Much earlier and folks will forget to read it (“I have plenty of time to get to this”). Last minute and folks won’t have time to read. In either case dangerous, because most directors will have an off-the-cuff reaction anyhow in the meeting (certainly no shortage of opinion among most VCs, myself included). But the off-the-cuff and the having-slept-on-it opinions may be quite different.
7. Avoid big (negative) surprises. If board members find out about things going poorly from the board package you are not communicating enough directly. See Tip #1 for this (sorry, no link - writing on my BB).
8. Provide spreadsheets in .xls format in addition to any summary that you may have included in the deck. Often I want to draw a chart for myself or calculate some ratios that you might not have. Don’t want to have to retype to do that.
9. Don’t make the deck a data dump. Attaching backup data is great (as per #8) but the board needs to know what you believe the data means or if you don’t know at least what the key question is that it raises.
10. Include one unexpected item somewhere in the deck. This could be a chart with a surprising metric or an odd quote from a customer or competitor. That may sound like unnecessary extra work, but I would be surprised if in the course of one month of running your startup you don’t come across at least one item that makes you laugh (or cry) or at least scratch your head. Share that with the board.
If despite doing all of the above you have a board member who regularly has not read the board package before the meeting, you will have to challenge him or her on it (obviously not in the board meeting). Much as you might want the responsibility for that to rest with someone else, it is not likely that another board member will take this on and you may discover something useful in the process.
The post on the USV blog about the challenge of creating native mobile apps (meaning ones that can on exist on the mobile platform) has created a lively discussion with by now over 60 comments. A good chunk of that discussion was about the relative merits of just sticking with the mobile browser versus developing apps. For instance, Fred observes:
I’m a big fan of the mobile browser too but it does seem like the prevailing delivery package today is the mobile app, not the browser.