Posts Tagged ‘software’

In our prior post we highlighted the comments of two well-known hedge fund investors as they describe the shift from a world often dominated by software to the new world of the model-driven business, wherein models power the key decisions in business processes, creating revenue streams and improved cost efficiencies.

The co-authors, hedge fund CEO Steven Cohen  and venture capitalist Matthew Granade, both of Point72 Ventures, and writing in August in The Wall Street Journal’s Opinion section, conclude their essay by pointing out five key implications of these models for business in the future, which we’ll summarize below.

First, businesses will be increasingly valued based on the completeness, not just the quantity, of the data they create.  Companies like Amazon and Tencent (cited in our prior post) don’t just collect massive reams of data about their customers, they know how to act on it.  It’s not just breadth of knowledge in other words, it’s also a ‘closed loop’ in which each recommendation a model makes, based on that user data, is captured and used to improve the model going forward.

Second, the goal is a ‘flywheel,’ they state, or a virtuous circle: Models improve products, products get used more, the new data improves the product (and the model) even more.  It’s a nearly frictionless process of continuous improvement.  Pretty close to a business holy grail if ever there was one.

Third, as incumbents (i.e., entrenched, major, competitive, winning companies) they will be “more potent competitors in this battle relative to their role in the software era,” insofar as they will have a meaningful advantage (operationally and profit-wise) this time around.   They have the trove of data competitors don’t, and they’ve learned how to monetize it.  And they keep getting better at it.

Fourth, just as the best companies have built deep organizational capabilities around managing people, tools, technology and capital, the same will now happen for models.  Just as software has become an ‘agile’ process in its delivery, the data-driven companies now on the rise are creating a new discipline of model management that affect the same domains: people, tools, technology and capital.  The models can now deliver, and that creates a critical competitive edge.

Fifth, companies will face new ethical and compliance challenges, according to Cohen and Granade.  With data becoming more comprehensive and important, consumer concerns over use and abuse are bound to rise – in fact, they already are.  Facebook lost over $100 billion in market cap in a couple of days in July because investors became concerned about its data assets in the face of increasing regulation.  That’s not likely to abate, and examples will only multiply.

Still, while software continues to “eat the world” (prior post), yesterday’s advantage becomes, in the authors’ words, “today’s table stakes.”  In the hunt for competitive advantage, the model-driven business will become ubiquitous.  For the average investor, software has been a great place to make money since Marc Andreessen’s famous “Why Software is Eating the World” essay seven year ago.  But in the next seven, the serious money is on the model-driven business.


Read Full Post »

In a famous essay in 2011 Netscape founder and venture capitalist Marc Andreessen notoriously wrote about “Why Software is Eating the World.”  He was right.  Companies he identified including Amazon, Netflix, Spotify and others did just that – they used software to eat their industries.  Today, newer companies are repeating the feat, including Airbnb, Stripe, Uber and others.  All are digging in.

It is said that today, all companies are software companies.  Or they had better be.  Many have adopted software to maintain or extend their competitive dominance, even in industries as arcane as pizza to name just one, where Domino’s has achieved a dominant position thanks to its software focus.

A recent article co-written by hedge fund CEO Steven Cohen and venture capitalist Matthew Granade, both of Point72 Ventures, and writing in August in The Wall Street Journal’s Opinion section, begs the question: What’s next?  Their answer?  Models.

This new paradigm is defined as a shift from a world often dominated by software to one driven by ‘models,’ which power the key decisions in business processes, creating revenue streams and improved cost efficiencies.  It requires a mechanism (usually software-based) to collect data, processes to create the models from the data, the models themselves, and finally a mechanism to deliver or act upon the suggestions from those models.  To be specific, quoting the authors on the latest paradigm shift:

“These companies structure their business processes to put continuously learning models, built on ‘closed loop’ data, at the center of what they do.  When built right, they create a reinforcing cycle: Their products get better, allowing them to collect more data, which allows them to build better models, making their products better, and onward.  These are model-driven businesses… being created across a range of industries.”

While there is plenty of hype about big data and AI, the models, they state, “are the source of the real power behind these tools.”  They go on to say… “A model is a decision framework in which the logic is derived by algorithm from data, rather than explicitly programmed by a developer or implicitly conveyed via a person’s intuition.”

Get it?  It’s the data-driven model’s ability to learn from itself and its own mistakes – at a rate much faster than mere humans can do – that sets it apart as an increasingly rapid prototyping tool for building the better business.  Humans, it seems, need not apply.

A Chinese company called Tencent provides a brilliant example.  They have customer data across social media, payments, gaming, messaging, media and music, and information on hundreds of millions of people, all of which they put into the hands of thousands of data scientists in order to make their products better.

That unique data helps Tencent to power a model factory “that constantly improves user experience and increases profitability – attracting more users, further improving the models and profitability.”  That’s a model driven business.

Closer to home, Amazon used software to separate itself from physical competitors but it was their models that helped them pull away from other e-commerce companies, Cohen and Granade point out.  By 2013, over a third of Amazon’s online revenue came from its recommendations, the result of model-driven data usage, “and its models have never stopped improving,” as Bezos & Co. continue to find myriad ways to use machine-learning models.

The authors go on to extol the model-based leveraging changes now taking place across the business spectrum.  Key industries include agriculture, services and logistics.  The implications are vast, and Cohen and Granade sum it up by describing five key implications for the future of business … and we’ll cover those five in our concluding post.  Stay tuned…




Read Full Post »

It was nearly 30 years ago that a Dutchman by the name Guido van Rossum set out to create a new programming language that would adhere to three principles: First, it should be easy to read; braces and dangling punctuation marks would be replaced by indented white space surrounding fairly readable code words.  Second, it should let users create their own modules or objects that would be usable by others to form the basis of new programs.  And third, he thought, it should have a catchy name.  And so he named it after the English comedy troupe, Monty Python.  The code package repositories became known as “the Cheese Shop.”

Little did von Rossum know that one day his creation – Python — would become the most popular language, said to have received more Google search hits in the past year than Kim Kardashian, according to the editors at The Economist.  Today, Python is used by nearly 40% of professional developers (another 25% wish to do so according to a programming forum).  It’s also popular with ‘ordinary folk’ as well, notes The Economist, and is snaring youthful adherents as well, noting that about 40% of American schools now offer computer programming (compared to about 10% just five years ago) and among them, about two-thirds of 10 to 12 year olds have an account on Code.org’s website.

Mr. van Rossum recently stepped down from his longtime role as curator and supervisor of Python (“benevolent dictator for life,” he once called it) saying he has long been uncomfortable with the fame.

Like all languages, Python is not perfect.  The more traditional C and C++ languages offer a broader suite of lower-level functionality (lower-level usually meaning that you are closer to the machine’s own language, and thus have more control over commands and overall language functionality).  Java is popular for large and complex apps, and JavaScript is the language of choice generally for apps accessed by a web browser.

But Python’s simple syntax makes it easier to learn to code, and share.  It’s used everywhere from the CIA for hacking to Pixar for producing films, in Google web page crawlers and in Spotify’s recommended songs feature.  It’s also recently said to have become a language of choice for AI researchers, which means it’s truly here to stay.

And even the non-coders of the world have taken it up in non-technical jobs, where marketers use it to build statistical models for their campaigns and teachers can see if they are properly distributing grades.  CitiGroup has introduced a course in Python for its trainee analysts, and it’s a haven for those who have long relied on spreadsheets for data analysis.

While no computing language can ever be all things to all people, specialization of languages matters, and so Python just keeps growing.  Over the history of computing, any number of languages have grown, dominated and then faded: think Fortran in the early days, then Basic’s many versions throughout the PC’s heyday, and even Pascal, a presumed lingua franca of the PC-era computing lexicon.  Presumably, Python’s day will fade one day as well, but now nearly 30 years after its birth, that day remains anyone’s guess – and probably in the distant future.

Read Full Post »

patentableThe answer to the question above may become a lot clearer next year, according to an article in the Wall Street Journal (Dec. 7, 2013).  The high court has agreed to hear an appeal of a case that “has tied the lower courts in knots,” according to the Journal article.

Basically, there has long been a lack of clarity with respect to whether – and when – computer code should get patent protection.

On one side of this digital divide are “new” tech companies like Google, Facebook and Intuit.  They believe too many patents have been issued lately, and would like to see courts “apply a more exacting standard when reviewing them.”  These firms believe that the pace of technology and innovation are improved when patents are harder to secure.

Opposing them are “old” tech firms like IBM that have spent decades amassing troves of patents.  As the Journal sums it up: “Smaller players worry tighter standards on software patents could hurt innovation by making it harder for them to get legal protection for their ideas, while the larger ones fear new standards will trigger attacks on their portfolios.”

According to at least one patent attorney, Matthew Moore of Latham & Watkins LLP in Washington D.C., “this is the biggest patent case we’ve seen in years for the technology sector, and probably the biggest we’ll see for the next decade.”

The suit being heard involves two international banks over a computer program that helps foreign-exchange buyers and sellers settle their trades.  One side says that the way it’s done now – with patents that describe a process using an intermediary to settle trades – has been around a long time and doesn’t deserve protection.  The other says that their patents did a lot more than just bring an old idea into the computer age, and the software serves as “an electronic intermediary in a particular way”.  Interested parties are divided then on how much scrutiny a court should apply to software contracts.

Whichever way this goes, the lack of clear rules in the software industry begs for certainty.  But it won’t be easy, according to legal experts.  As patent lawyer Ed Reines sums up: “You basically have a ‘you-know-an-innovation-when-you-see-it’ standard.  Articulating a test that divides what’s patentable and what isn’t is extremely difficult.”

No date for the high court hearing has been announced, but expect to more about this in the news next year.


Read Full Post »