Every day, somewhere in the cloud, Bluefin Labs of Cambridge. MA, is analyzing 200 channels of US television broadcasts around the clock. That’s something like an unfathomable 172 terabytes of raw video data (and a goodly bit of drivel one might add). The company then collects programming information including channels, broadcast schedules, and keywords to tag each show and ad. For each group of ads, Bluefin workers make the initial product identification and then their analytics system automatically identifies repeat airings.

But that’s not all: Bluefin additionally monitors 300 million public social-media comments per day for the keywords associated with programs and ads.  Out of that huge number there are on average about 10 million comments about TV content. About 1.4 million of those are made in what Bluefin considers the relevant context for a particular show or advertisement, which is about three hours before to three hours after the actual broadcast. These comments pop up primarily in tweets but there are also public Facebook posts and some other media sites. The company also tracks the online activity of the 9.8 million people who have made online comments about television in the last 90 days.

What does all this add up to (no pun intended)? Based on research originally begun at the prestigious MIT Media Lab, Bluefin Labs is building a business out of creating a context in which television advertisers can understand which ads have more success so that they can make better decisions about where to place their ads in the future. Television producers can also tap into viewers responses and modify content to better please and perhaps increase their audiences.

The effort to better understand audience responses through analyses of social media is still in its infancy. Target marketing is the new Holy Grail for marketing executives. And what Bluefin has been able to accomplish is in some ways impressive.  It has created a context that shows how the same viewers might respond to different ads when watching different shows. It  ultimately provides two measurements as well. The first is “response level,” which counts the number of people commenting on any given show or ad. The second is “ response share,” which indicates the percentage of all responses that a single show or ad has garnered. The company is already attracting some large corporate clients, such as Pepsico, CBS, Comcast, Fox Sports, and Turner Broadcasting, writes David Talbot in the recent print issue of Technology Review.

That’s a remarkable list and it indicates just how important analytics is becoming for broadcasters and manufacturers.TV advertising is big business, after all. It dictates the success or failure of TV broadcasting because companies spend $72 billion annually these days on TV advertising. And  Americans watch a lot of television. According to the Nielson Corporation they spend 20% of their waking hours watching television, some of them multitasking by using their laptops or smartphones at the same time.

Still there are some reasons to be skeptical. For one, as Hill Holiday advertising executive Mike Proulx points out, the connection between social media reaction and the impact of content is still just a theory. There’s no solid evidence to date that proves it is a valid connection. Another major drawback, it seems to me, is that people who actively tweet or post comments on Facebook or elsewhere about television programs are only a subset of the couch potato population, no doubt skewing younger for one thing. What’s more, it’s also becoming more and more common for people to record television shows and watch them at random times, making the correlation between the three hours before and after a broadcast irrelevant for many viewers these days.  

So it remains to be seen if all the effort and advanced expertise that is being poured into advertising analytics and social media will generate better advertising, or whether corporations will still be saying ten years from now what the merchant John Wanamaker said over ninety years ago: “Half of the money I spend on advertising is wasted; the problem is I don’t know which half.”

After several years of debate about the job market, whether we need to create new jobs and save old ones through stimulus or pull back on spending and go for austerity, it’s good to escape that well-worn debate and find a new perspective, one that approaches the unemployment problem with a longer lens. Two MIT professors, Erik Brynjolfsson and Andrew McAfee have takena different look at why the unemployment problem has become such a chronic one. They find a major culprit in the progress of various digital technologies.

In Race Against the Machine: How the Digital Revolution Is Accelerating Innovation, Driving Productivity, and Irreversibly Transforming Employment and the Economy, Erik Brynjolfsson and Andrew McAfee team up to argue that the root cause of the unemployment problem is that we are in the early stage of a “Great Restructuring.” The long-held correlation between job growth and economic growth is no longer valid, the authors contend. It is technology—its widespread use across all industries, especially in the last decade or so—that has precipitated a major “displacement” of jobs. The recession simply accelerated the pace, making the connection between the stagnation in the job market and technology more obvious as the job losses rose and high unemployment has persisted.  GDP, they point out, continues to rise while median income stagnates. And since the end of the recession, companies have increased their spending on equipment and software by 26% but payrolls have remained basically flat.

What’s more, there’s a growing disparity between the rate at which machines improve and the rate at which humans can change, a disparity which will only make things worse in the future. Moore’s Law continues to prevail for hardware, which saw processing speeds improve by a factor of 1000 from 1988 to 2003. During the same period, however, software algorithms improved by a factor of 43,000. Brynjolfsson and McAfee believe that such improvement rates have brought us to the inflection point where huge strides can be achieved very quickly. They cite Watson winning on Jeopardy! and Google’s self-driving cars as prime examples of recent breakthroughs that seemed decades away even a few years ago.

The authors identify three structural changes that have been unfolding for more than a decade and that are creating more unemployment problems:

(1)    Technology is replacing the jobs of lower skilled workers, with everything from robots that manufacture cars to voice-recognition software that answers telephones and resolves customer problems. Even in low-wage countries like China, robots are taking over the jobs of unskilled and semi-skilled workers. The electronics company Foxconn plans to buy one million robots over the next three years to replace most of its workforce. The company currently has 10,000 robots and expects to have 300,000 within the next year.

(2)    Digitization means write once, read many, and this has created the “superstar” effect, where individuals and corporations benefit from the replication of everything from hit songs to advanced intellectual property. As a result single individuals can have a huge impact—and reap equally huge rewards—through their skills and decisions. The superstars overpower some very good competition that just can’t get to the top. It’s becoming a winner-take-all marketplace.

(3)    The division between labor and capital is also shifting.  As the input of human labor decreases in a particular business process, the owners of the capital equipment gain a proportionately larger amount of the bargaining power and the income. As a result, for example, corporate profits have rebounded and risen dramatically since the end of the recession.

Brynjolfsson and McAfee arguethat the digital frontier we have entered represents the third industrial revolution, after steam and electricity. Those who will succeed in this new frontier may well, the authors claim, be the ones who compete not against  the machine but with the machine. Learning how to use computers to improve organizations and to make sure that workers have the right skills for the future are key. The authors also advise people to cultivate the skills that computers will not be able to master, including leadership, team building, complex communications, and creativity. Entrepreneurs should find opportunities that take advantage of cheaper technology and (it is implied) cheaper mid-skilled unemployed workers to create new business models, bringing together people and computers in new and unexpected ways and creating new marketplaces.

The two researchers had originally begun their joint research on a book that would explore the opportunities for innovation in the “Digital Frontier.” The last part of their book does return to that theme, offering visions of entrepreneurial success. They also supply a social roadmap (with no less than 19 points!) for revamping education and government to support innovative and fast-changing organizations. Still even they admit that there are limits to their visions. Not everyone is cut out to be an entrepreneur and not all entrepreneurial businesses create lots of jobs, especially today,  presumably because of the widespread use of more technology and fewer people in those businesses.

After such compelling arguments about the deep economic restructuring that is going on because of technology, I find it hard to be as optimistic as the authors about reviving the job market. The nineteen points are broad ranging. They include some oft-discussed issues such as investing in our educational system to improve it in various ways and revamping the visa system to encourage skilled workers and people with advanced to degrees to remain in this country. Other less common and very good ideas include teaching entrepreneurial thinking at all levels of education and creating databases and sets of standardized business processes for new entrepreneurs to use. Still, the progress of technology has left so many people in the dust at this point that it’s hard to even calculate how long and hard the road to employment recovery might be.