Predictive analysis pays dividends

Companies are being drawn to a growing number of specialists who comb the internet for patterns of information that help them to make strategic moves. Samsung has been among these, in its attempt to usurp rival Apple.

Samsung's advert during the Super Bowl helped it to generate $3bn in revenues on top of initial forecasts. YouTube

It was 2011, and Samsung was looking to make a move on its rival, Apple.

To help it do so, Samsung turned to a company called Networked Insights, one of a growing number of businesses around the world that specialises in what is known as predictive analysis.

Using computer software, Networked Insights and its competitors in this space digitally mine through sites such as Facebook, Twitter and YouTube, as well as blogs and forums, while looking for patterns of information that may help clients make strategic moves and garner more market share.

For Samsung, Networked Insights conducted an assessment after spending two weeks combing through seven million conversations and opinions about various smartphones discussed on different sites.

It then employed a separate technological tool, homing in on "a broad, tech-savvy audience that craved a device that would rival the iPhone," according to Networked Insights. Part of the company's plan: speak directly to this audience to limit the "Apple-bashing backlash" other tech companies such as Microsoft have experienced in their own marketing campaigns.

In the end, the different data points that were collected helped shape an advertising campaign that launched over Thanksgiving as well as a subsequent commercial that aired during the Super Bowl. According to Networked Insights, it helped Samsung generate US$3 billion (Dh11.01bn) in revenues on top of the initial forecast.

"Companies that are drawn to us today are companies that have fast decision cycles [such as] movie studios and consumer electronic brands," says Jason Kapler, the director of marketing for Networked Insights, which is based in the United States. "Those are the ones looking for competitive advantages."

There are plenty of other ventures seeking a competitive advantage as well. Some are trying to dig deep into past sales data to get an idea of what new product might become a best-seller tomorrow. Others are scanning customer data and using their historic purchases to pinpoint seemingly unusual account activity and prevent acts of fraud from unwanted intruders.

Predictive analytic specialists and conference organisers are now growing their own business by reaching out to these kinds of future-obsessed ventures.

Predictive Analytics World, which is hosting an entire series of conferences this year in cities such as San Francisco, Chicago, London and Berlin, is telling potential participants that they can learn how to harness streams of data from their companies for prediction purposes.

Major technology companies that boast predictive analytic tools include SAP, IBM and SAS, which notes that it provides "a wide range of software for exploring and analysing data to help uncover unknown patterns, opportunities and insights that can drive proactive, evidence-based decision making within your organisation".

Smaller firms in this space include Networked Insights, TrendSpottr and Recorded Future, whose tagline reads: "Unlock the predictive power of the web".

The challenge for all of these firms, however, is that it is difficult to predict exactly when a client's competitor may be releasing its next consumer electronic device or expanding into a new market, for instance.

Or, for that matter, accurately estimating which day of the week a protest is most likely to erupt and interrupt business in a region like the Middle East.

Still, this has not stopped the firms from trying.

"I wouldn't say today we spit out the answer," says Matt Kodama, the vice president of product for Recorded Future, based in the US.

"It turns out a lot of information is out there in accurate forms and we know where to get the information," he adds.

"A great deal of what we do is help [a financial] analyst get out of the business of scouring the web to gather, collate and plot all this on a timeline and just handle this for them."

One of the areas Recorded Future specialises in is analysing text from sites to determine whether an event has already transpired, is currently happening or set to occur in the future. It can then determine patterns and break out the information visually for clients.

Take car makers or consumer electronic manufacturers, for instance.

Companies from both sectors seek to gain competitive intelligence, such as when Ford's rival plans to open a new manufacturing plant or Dell's opponent aims to launch a new laptop.

It turns out there are many vendors tied to their supply chains for these kinds of products, so studying the vendors' financial reports, social media messages and any mentions on techie blogs becomes increasingly accurate as the actual release date nears.

"There are a lot of people, and as you get closer to the release date the word gets out," says Mr Kodama.

"Loose lips sink ships."