Thursday, July 31, 2014

Failed Entrepreneurs Find More Success the Second Time (BusinessWeek)


Failed Entrepreneurs Find More Success the Second Time
Photograph by Phil Augustavo
Given the slight chances of success, it’s a marvel anyone ever starts a business at all. One-third of new ventures close within two years, half within five years, and so on: only one in four is still around 15 years after opening day. But all that failure may offer its own reward, according to new research from a pair of economists from Stanford and the University of Michigan. They found that failed entrepreneurs are far more likely to be successful in their second go-around, provided they try again.
The entrepreneurship studies that grab headlines tend to focus on investor-backed, technology startups. Those types of firms aren’t the norm. Most new businesses are still small, local retailers. To understand how these enterprises fare, Francine Lafontaine and Kathryn Shaw studied the successes and failures of retail entrepreneurs in Texas from 1990 to 2011. Over the 21-year-period, 2.4 million retail businesses opened and 2.2 million closed. Three out of every four were founded by first-time business owners.
Lafontaine and Shaw found that the Texas retailers were less successful than the national average for small businesses: One in four closed after a year; half after two. What happened next was telling. Of the first-time entrepreneurs whose businesses closed quickly, the overwhelming majority—71 percent—didn’t bother to try again. But the tenacious 29 percent who did were more likely to be successful the second, third, and even tenth time around. Somewhat paradoxically, their success rate increased with their number of past failures.
The researchers argue that experience, even when it’s not positive, is invaluable—that entrepreneurs learn effectively from mistakes as well as from successes. They even found that serial entrepreneurs are successful in new types of businesses. Experience owning a hair salon translates into more success at running a clothing store. (There’s one important exception: First-time-restaurant owners, no matter their business background, tend to fail; serial restaurateurs are more successful.)
This paints a different picture from previous research, which suggested that failed entrepreneurs are more likely to fail in subsequent attempts. Writing in the Harvard Business Review, a trio of researchers presented findings that, among 576 entrepreneurs in the U.K., those who take on one project after another are less capable of learning from failure compared to entrepreneurs who pursue multiple businesses at the same time. They wrote:
Serial entrepreneurs’ greater propensity to remain overoptimistic may be due in part to the deep pain, even trauma, they feel when their projects fail—pain that is especially acute precisely because they involve themselves in only one business at a time. Psychological research suggests that strong emotions often prompt people to blame others or external events rather than themselves so that they can maintain some semblance of self-esteem and a sense of control. This ‘attributional bias’ appears to make serial entrepreneurs less capable of learning from failure.
Looking at a different population, over a different time period, in a country with different regulatory structures, the Texas research concluded that serial entrepreneurs were in fact quite able to learn from past mistakes. Among people who are willing to try again, the odds of success rise. Don’t count the failures out quite yet.

Allison Schrager is an economist and writer in New York City. Follow her on Twitter: @AllisonSchrager.

Wednesday, July 30, 2014

NASA request for information summarizes the datacenter market (ZDNet)

Summary: A request for information from NASA reflects the current mindset of datacenter operators.

By  for Five Nines: The Next Gen Datacenter |

The NASA Goddard Space Flight Center (GSFC) finds itself in the same situation that many businesses and data center operators find themselves: It’s time to start considering the upgrade path for their datacenter operations.
But unlike most businesses, the GSFC has taken the approach of a public request, under the name Data Efficiency and Containerization, for information on three main approaches that they believe can solve the datacenter growth needs.

Special Feature

The 21st Century Data Center
More than ever, data centers run the world, but many of them need a 21st century reboot. Today’s data centers have to be more efficient, redundant, and flexible than ever. We examine when and how to best run your own data center versus when to outsource to the cloud or a service provider, and when to take a hybrid approach.
The GSFC has defined three solution paths — short-term, interim transitional, and long-term — that they believe best suit the facility’s needs. They are asking for vendors to respond to the request for information with solutions to these models that not only serve their needs but also meet the Federal Data Centers Consolidation and Green mandates.
The basic strategies that they want to address are:
  1. Datacenter retrofit technologies and solutions with high return-on-investment (ROI) as short term, interim measures;

  2. Individual stand alone, high ROI containerized DC solutions that meet criteria for two distinct use cases;

  3. Modular containerized DC solutions, again with high ROI, in order to establish a longer term strategic DC plan
In a nutshell, these three strategies accurately reflect the choices facing most datacenter operators. But the GSFC isn’t just letting vendors decide how to do this; they’ve provided very specific guidance in what they would like to see in each of the data scenarios. The detailed scenario considerations can be found here.

While each scenario has its own specific requirements, the savvy responder will see that they effectively need to mesh well, defining a growth path that starts with improving the existing infrastructure and evolves to the modularized replacement of the current datacenter. Solutions that discretely address each component, with associated reference material to real-life metrics and case studies, are a large part of what is being sought. But, an organized approach that walks through the entire lifecycle process seems to be the underlying demand that NASA is making in this request for information.
With more than 20 years of published writings about tech, as well as industry stints as everything from a database developer to CTO, David Chernicoff has earned the term "veteran" in the technology world.

Tuesday, July 29, 2014

Power-hungry SSDs: Hotter than disks (ZDNet)

Summary: Making flash SSDs look like disks isn't easy. In fact, advanced high-performance SSDs use more power and run much hotter than disks. They aren't your father's thumb drive.

By  for Storage Bits |
Anyone looking at how flash SSDs have revolutionized mobile computing could be forgiven for thinking that all SSDs run cool and sip power. But they don't.
Flash technology isn't ideal for high performance apps. Each flash die has limited bandwidth. Writes are slow. Wear must be leveled. ECC is required. DRAM buffers smooth out data flows. Controllers run code to manage all the tricks required to make an EEPROM look like a disk, only faster.

Special Feature

Storage: Fear, Loss, and Innovation in 2014
The rise of big data and the demand for real-time information is putting more pressure than ever on enterprise storage. We look at the technologies that enterprises are using to keep up, from SSDs to storage virtualization to network and data center transformation.
So the number of chips and channels in high performance SSDs has risen to achieve high bandwidth and low latency. Which takes power and creates heat.

The price of performance

In a recent Usenix HotStorage '14 paper, Power, Energy and Thermal Considerations in SSD-Based I/O Acceleration, researchers Jie Zhang, Mustafa Shihab and Myoungsoo Jung of UT Dallas examined high-end SSDs, those with multiple channels, cores and flash chips. Fast, robust SSDs need all the help they can get. 
They found that high-performance SSDs exhibit characteristics uncommon in lesser SSDs.
  • High power. 2-7x the power, 282% higher for reads, up to 18w total.
  • High temperatures. 150-210% higher than conventional SSDs, up to 182F.
  • Performance throttling. At 180F the many-resource SSD throttles performance by 16%, equivalent to hitting the write cliff.
  • Large write penalty. Writes at 64KB and above in aged devices caused the highest temperatures, likely due to extra garbage collection and wear leveling overhead.
Performance throttling was not limited to the high-end SSDs. A mid-range drive slowed down at 170F, probably due to thermally-induced malfunction as the drive had no autonomic power adjustment.

The Storage Bits take

I found these results a little hard to believe, so I looked at some high-end enterprise SSD specs. Sure enough, I found an Intel SSD spec'd at 25w - twice the power required for a 15,000 rpm enterprise disk - and higher than any the researchers tested.
Should enthusiasts be concerned? Maybe. The performance throttling from high temps of even mid-range SSDs could affect a game at its most intense. If your system gets sketchy at high temps, this could be a cause.
For enterprise users this is a reminder to check specs and understand power and cooling requirements for high density SSD installations. Free-cooled datacenters may also be at risk for SSD-induced slowdowns.

While not a huge problem for most people today, the insatiable demand for performance will move more of the market to these high-performance SSDs. The time to think about the impact is now.
Robin Harris is Chief Analyst at TechnoQWAN LLC, based in Sedona, Arizona. He has over 30 years in the IT industry, including DEC and Sun, and degrees from Yale and the Wharton School.


Thursday, July 24, 2014

Self-driving cars...Part Two (Cnet)

Why self-driving cars won't be a mass reality anytime soon

Commentary: As Google reveals its latest self-driving car, a bold future seems to have arrived. But we've been down this disappointing road before.
googleselfdriving.jpg
Google's latest self-driving prototype.
Driverless cars are coming. On Tuesday, Google unveiled its latest prototype -- one so amazing it lacks manual controls.
But after decades of waiting for smarter cars for the masses, will they really arrive?
Believe me, I love the promise. With a teenager about to learn to drive and another only three years away, this parent wants smart, driverless cars for the general public now. Plus, what adult wouldn't love a car that's always your "designated driver" to get you home safely after drinking at a dinner or concert?
I've even been in one of Google's cars a few years ago and was amazed as it zipped me around a test track at high speed. I didn't think twice about my kids climbing in and going for a spin without me, since it felt so safe.
But much as I'm a fan of the cars and astounded by the technology, I can't help but feel that when 2020 comes around -- despite what GoogleNissan and GM predict -- such cars still be in limited use, in limited areas, in limited cases.
That's largely because I keep thinking back to a Los Angeles Times article that I found so gripping nearly 25 years ago. When it came out in March 1991, "A Smart Way to Unclog Roadways" documented a less-ambitious future than the fully autonomous cars we're being promised today.
Planners then were hoping that our cars and roadways would both get smart, so that once a car entered a freeway, it would go into auto-mode. It would then "platoon" in a link-up with other cars, effectively creating a train.
Hook your car into a platoon of self-driving cars, then use your new free time to sip your coffee, listen to music, have a snack, make a call, or read a book.
Hooking your car into a platoon would theoretically free you up to read or make.It sounds like a great idea, but the idea has yet to arrive. In fact, despite all this time, platooning is still in the testing stage and is considered future tech. To be fair, when the LA Times article was written, platooning was seen as something that might take up to 50 years to achieve. So we've got about another 25 years to go, I suppose.
But then again, that same article made mention of predictions from the 1939 World's Fair, where GM's Futurama exhibit anticipated automated highways with "radio-control" keeping cars a safe distance from each other. Wired offers aretrospective on that, and Wikipedia has an entry as well.
Below, a video from the time talks about how motorways of 1960 might work. (Jump to 14 minutes, 28 seconds for the key part.)
So 1939 promised us automated highways by 1960 that never materialized. In 1991, I was dazzled by the article that suggested the future was really about to arrive.
Yet here I am living in that "future" and still finding myself driving my car manually, on crowded freeways, with another promise that automation is just around the corner to make it all better.
Certainly some things promised in terms of car tech have arrived. This part of the 1991 LA Times article makes me smile:
"For example, if a driver wanted to get from downtown Los Angeles to Santa Monica, the usual route would be the Santa Monica Freeway, or Interstate 10. But what if an accident near the La Brea exit occurred while the driver was just a few miles from the scene? Before he could get caught in the jam, his car would be told of the incident by the traffic command central and would automatically exit the freeway, finding the least congested alternative route based on the calculations performed by its computerized mapping system for both the freeway as well as adjacent major streets."
My GPS works exactly like this today, and it can be a huge time saver. But ultimately, it's still me behind the wheel, doing the driving.
Part of me feels that even if our cars are getting smarter, the roads they drive on still seem pretty dumb, which means the cars still must drop into a lowest-common denominator mode.
Part of me also feels like self-driving car engineers are too ambitious. While Google's recently posted of big progress in city driving situations, it ultimately can only have its cars work where it has created detailed maps -- effectively tracks -- for the cars to follow.
How about focusing on just getting all the legal and technological challengestackled so these cars work on freeways, even if only in platooning mode? I'd be happy with that start, if it meant I'd get a self-driving car that I can actually use sooner.
When the next decade arrives, will it be commonplace that cars are just driving us around? I don't think so. One recent study suggests 2035 as a more reasonable date, with nearly all cars being autonomous by 2050.
Still, I want to be optimistic. I'm certainly glad that so much progress has been made with autonomous vehicles recently. I hope the promises being made indeed come true.
I'm still holding out hope that the jetpacks we were promised will materialize too.
Danny Sullivan is a journalist who has covered the search and internet marketing space for over 15 years. He's founding editor of Search Engine Land and Marketing Land, and writes a personal blog called Daggle (and maintains his disclosures page there)

FBI: Driverless cars could become 'lethal weapons' (CNet)

In a report, the agency sees the vehicles as opening up new possibilities for law enforcement, but also for criminals.

BMW Highly Automated Driving
Carmaker BMW envisions autonomous cars handling high-speed driving situations. Here, a BMW 6 Series manages to automatically counterturn to avoid a spinout, as the driver keeps his hands off the wheel and his feet off the pedals.
Driverless cars may one day be able to save lives by cutting down on accidents. But the technology also has the potential to be used to create deadly weapons.
At least, those are some of the findings in an FBI report obtained by The Guardian via a public-records request. The report, created by the Strategic Issues Group within the FBI's Directorate of Intelligence, says the technology "will also open up greater possibilities for dual-use applications and ways for a car to be more of a potential lethal weapon than it is today."
Further, the report says "bad actors will be able to conduct tasks that require use of both hands or taking one's eyes off the road which would be impossible today."
The Guardian's story, published Wednesday, imagines nightmare scenarios such as tailed suspects being better able to shoot at police, and terrorists potentially packing explosives into a self-driving car aimed at a specific destination.
Google has been running prototypes of driverless cars, and several automakers have been testing them as well. As the technology continues to advance, safety and legal issues have naturally been raised. With no person in control of the car, who's responsible in the event of an accident or other disaster? And human beings invariably find a way to misuse a technology for the worse, so the FBI's own concerns shouldn't be discounted.
But the FBI's report does see an upside to driverless cars. The agency expects the technology to reduce the number of collisions and other accidents involving emergency personnel who may be speeding to reach people in need of urgent care.
"The risk [of] distraction or poor judgement leading to collision that stems from manual operation would be substantially reduced," the report says.
Law enforcement may also more easily be able to track suspects surreptitiously.
"Algorithms can control the distance that the patrol car is behind the target to avoid detection or intentionally have a patrol car make opposite turns at intersections, yet successfully meet up at later points with the target," the report adds.
How long does the FBI expect it to take for driverless cars to be OKed for use in this country? The agency says that could happen within the next five to seven years.
Journalist, software trainer, and Web developer Lance Whitney writes columns and reviews for CNET, Computer Shopper, Microsoft TechNet, and other technology sites. His first book, "Windows 8 Five Minutes at a Time," was published by Wiley & Sons in November 2012. 
     

    Wednesday, July 23, 2014

    How Ford plans to win the future like a software company (TechRepublic)

    By  July 23, 2014,

    Ford has used technology to recreate itself as a 21st century car company. Now it wants to build new experiences and innovate like a software maker.

    ford-headquarters-flags-600-v1.jpg
    Ford World Headquarters in Dearborn, Michigan
     Image: Jason Hiner/TechRepublic
    When Don Butler came to Ford Motor Company at the beginning of 2014 and stepped into his new office at Ford's Product Development Center in Dearborn, Michigan, it was kind of like a star athlete slipping on the uniform of his cross-town archrival.
    The move was a huge win for Ford's journey as a technology company, because Butler came from General Motors where he had combined his expertise in engineering, marketing, and business development to lead two of the auto industry's premier technology projects -- GM's OnStar system and the Cadillac CUEinterface.
    Meanwhile, over the past five years Ford has generated its owntechnology mojo with the spread of its SYNC and MyFordTouch systems, which despite a few hiccups and imperfections, have successfully integrated information technology and consumer electronics into millions of new cars and trucks.
    With Butler on its team, Ford has the potential to take its technology strategy to the next level. And, Butler hasn't wasted any time figuring out how Ford needs to think differently to get a jump on the next evolution of the auto industry.
    "[This role] was tailor-made for me..." said Butler in a recent interview with TechRepublic. "I'm passionate about being able to help this company and this business... My role has been, first of all, to synthesize our vision around connectivity. What does connectivity mean for us? Then after that, what's the strategy to accomplish that vision?"
    Butler and his Ford colleagues have broken down vehicle connectivity into three buckets:
    1. Brought-in connectivity - Butler refers to this as "leveraging smart devices and the capability of those smart devices," including their apps, media, and communications capabilities, as well as 911 Assist and vehicle health reports; he said Ford is already in a leadership position in this area.
    2. Beamed-in connectivity - This involves external connections such as satellite radio, Wi-Fi, and Bluetooth.
    3. Built-in connectivity - The final piece is what Butler describes as when "the vehicle itself, independent of any other devices, has its own built-in data connection through on-board data mode."

    Ford has obviously done good work with No. 1, but there's still more work to do in making the interface easier for users to navigate, and Butler brings fresh experience from working on this problem with Cadillac CUE.
    In terms of No. 2 and No. 3, Butler brings all of his experience in telematics from his days helping to launch OnStar at GM. The fact that Butler gave equal weight to these three indicates that Ford is likely to get a lot more serious about connecting vehicles to the internet beyond just tethering to smartphones.
    While he didn't divulge the details of Ford's future connectivity plans, Butler did talk about the two big changes Ford is going to make to its long-term strategy:
    1. Learn how to innovate like a software company
    2. Move to long-term relationships with customers

    As straightforward as those goals may sound, pulling them off will involve deep cultural changes at Ford and that's where Butler is using his considerable energy to drive change.

    "At a high level in terms of business, thinking and acting more like a software or technology company is really what we need to be about," said Butler.
    ford-butler.jpg
    Ford's Don Butler
     Image: Ford
    Of course, that makes perfect sense since software is destined to power more and more of the systems inside Ford vehicles, but Butler is also thinking bigger in terms of the way Ford builds its vehicles, creates a platform to innovate, and delivers a next generation of services to the people who buy Ford vehicles.
    "When it comes to thinking like a software and technology company, [we need to make sure] the vehicle is updatable over time, and we want to plan on a certain number of software updates throughout the year," said Butler. "Device makers have been doing it for a long time. Automakers haven't been doing it for a long time... Enhancements on an ongoing basis need to be thought about and planned... There's some fundamental changes in terms of how we need to organize business."
    One of the biggest obstacles remains the product development lifecycle of a new automobile. In most cases, it's five years or more. That means decisions have to be made about technology at an early stage in the process, and many of the technologies involved will become obsolete by the time the product comes to market. That's the kind of product rhythm that carmakers are used to. That's what Butler and team have to change.
    What Butler wants to do is to create a "platform for innovation" where Ford doesn't have to predict the future so much in making its decisions about tech during the product development cycle. Instead, it can create a platform that solidifies a few key elements while leaving the door open for evolution during the process of developing a vehicle, as well as afterward when the vehicle goes to market. He cited four elements that would make up this platform for innovation:
    1. in-vehicle hardware and software
    2. IT back-end infrastructure
    3. processes
    4. people

    The end game: Move a lot faster and be more adaptable.
    Butler said, "We need to enter a culture of rapid experimentation -- really rapid, small-scale experiments to learn from, and then if something doesn't work, don't go forward with that. More of a beta testing kind of environment."
    And Ford isn't wasting any time getting started. It recently announced a partnership with Intel to bring next generation user interfaces -- including facial recognition and gesture UI -- into vehicles. It's called Mobile Interior Imaging or "Project Mobii." It could be used for recognizing the person that sits in the driver's seat and adjusting the car's controls, seat position, and music playlists to that person's preferences. If the person isn't recognized, then the car could immediately send a photo to the car's primary owner to verify the unrecognized driver has permission to use the vehicle. The system could also use gestures and voice commands to control the car's systems, such as waving a hand in a certain way to open the sunroof or speaking a command to change the temperature of the vehicle.
    "Mobii is a great example of this culture of experimentation," said Butler. "I don't know what the future of UI is going to be; I don't know what the user experience inside of the vehicle is going to be. I know that increasingly I've got sensors, cameras, technology, [and it's] trying to mash some stuff together and see what might happen. The use cases that were portrayed are things that we think might make sense."
    And partnering with Intel was a perfect fit because both Ford and Intel already had researchers working on the same issues, Butler said. In fact, Butler sees Ford working a lot more closely with partners in the years ahead.
    "This is an area where you literally have to create the future together," he said. "We might have some ideas of where we want to go, but if [Ford] can partner with the right people, I think we'll come up with solutions that neither of us on our own could have come up with. Whether it's with Intel, or NVIDIA, or Google, or Apple, we talk to a number of different companies. We have to be comfortable with inventing the future together."
    And, it's not just partners that Ford needs to build closer relationships with, according to Butler. If the company is going to evolve to run its product more like a software company, then it also needs to change how it relates to its customers.
    "Once someone has decided to begin a relationship with Ford, we want to earn their trust and their loyalty by delivering a safe, superior, connected experience. It's that simple," said Butler. "And the challenge that we've had -- not only Ford but automakers in general -- is we haven't thought of it as a relationship. Nor have our customers, quite candidly... And we want to change that."
    That's the other big opportunity that Butler is chasing: building deeper, longer term relationships with customers so Ford can add value to their lives and to their vehicles.
    "[What] connectivity does for us is it enables us to literally connect and know the customer better," he said.
    That translates into big data. Ford has been an early pioneer incollecting and processing big data and using it in ways to enhance its business. But now, it wants to use big data in the same ways that companies like Google, Amazon, and Facebook use it -- to streamline and customize the user experience.
    Fortunately, Butler realizes that before Ford launches into that arena it needs to think very carefully about some of the critical currency of the digital age: trust, privacy, and security.
    "Our philosophy is one of stewardship. The question is who owns the data, whose data is it. It's the customer's data. They are the owners," Butler said. "This notion of stewardship is both one of protection and creation of value... Steward is a word not used very much, but in ancient times a steward is someone who a land owner put in charge of their property. The landowner still owned the property, the steward didn't, but they were in charge of taking care of it. And not only taking care of it, but stewards made sure it was in better condition when the landowner returned than when they left. Our notion is if customers trust us to take care of, manage, and use their data, we will then leverage that to improve their experience in the vehicle."
    That means Ford has to put systems in place to allow users the right-to-be-forgotten and the ability to opt out and to remove or anonymize their data.
    "We want to be in a position that a customer has control over any data that is associated with them, can be personally identifiable to them. They need to be able to say 'I'd like you to extract all of that from your database.' ... Now, how many people do I think would actually do that? Not many at all. Because if that's your inclination, you probably wouldn't have given us permission to use the data to begin with. But if you know you can do it, I believe you'll be more trusting. You are in control, and we want our customers to be as in-control as possible."
    But, for those who buy Ford vehicles and entrust Ford with access to some of their data, Butler believes Ford can add a lot of new value to the equation of owning a vehicle.
    He gave the example of the oil light blinking in your vehicle. Today, you see that and you know that you need to go schedule an oil change, and you'll typically procrastinate for a while until you have time to check your schedule and call and set something up. Butler said Ford could turn that into a "seamless, almost painless experience" if it used the information it knows about its customers. Ford services could detect that your vehicle is low on oil or due for an oil change and could then use the information it knows about your preferred dealership to automatically check that dealership's calendar, look at the window of time when you've scheduled past appointments, and then automatically schedule something in an available slot and email you a calendar notification. Obviously, you could reschedule if needed, but getting it on your calendar right away makes it more likely you'll actually go and do it on time.
    As important as those services are -- and as much as it also means an additional revenue stream for Ford -- Butler still keeps his eye on the ball about what the whole process means for Ford.
    "Nothing happens in our business today until we sell or lease a vehicle," he said. "We are primarily in the business of designing and selling vehicles. So the first thing connectivity has to do for us is make our vehicles more attractive, make them more desirable."
    Jason Hiner is the Global Editor in Chief of TechRepublic and Global Long Form Editor of ZDNet. He is an award-winning journalist who writes about the people, products, and ideas that are revolutionizing the ways we live and work in the 21st century.