Speed, as a certain wild haired patent clerk has reassured us, is all relative. But processing speed of late has come on in leaps and bounds making things possible which were not previously.
The Big Data phenomenon and real time business intelligence are examples of how sheer processing power can makes things which were impossible seem routine.
There are other examples and a recent chat with Accenture here covered facial recognition and other image processing technologies that have allowed it to develop a passport processing system for airports. Already in operation in Stansted, Heathrow and Schiphol, it allows users to enter a turnstile where their electronic passport is scanned, as is there face, and so automate passport control. But the technology goes further so that commodity hardware can be used in conjunction with the clever software for health monitoring for vulnerable people at home. The systems can detect falls, or unusual behaviour that might be a seizure, or other medical symptom.
In a wider arena, Irish based company Profitero specialises in retail price monitoring. The company predicts that large online retailers, especially in the electronics space, will likely in the future be able to implement stock market-like pricing structures. What this means is that systems will monitor the Web under various parameters to determine the best price for wares and adjust them in real time. This would mean that everything from competitor pricing, supply issues, seasonal factors and even popular media trends could be included to ensure that pricing is most appropriate.
A key parameter in this would likely be sentiment monitoring. This is where social media and other public fora are monitored for what people are saying about brands, products, services and the like, and determining how such expressions might affect those companies, products and services. Like a barometer of the nation, or host of people, such measurements have long been used to gauge interest or positions on the likes of major national issues.
However, the gathering of such sentiment data has not always been easy.
The advent of cloud services, hyper-connectivity and accessible platforms with large subscriber bases has accelerated this. Now, Facebook’s billion plus users can be scanned and monitored in ways that marketers, researchers and, indeed, anthropologists could only have dreamed of previously. The scale of the infrastructure to accommodate people online in social media is only matched by the infrastructure and scale now being deployed to monitor and cater to their whims. So, for example, it has been reported by that august IT news source The Register that the YoutTube has made $8 million or so from the novelty song "Gangnam Style" by South Korean artist Psy. By being able to react to what can only be called unexpected popularity of a regional, obscure cross over, YouTube was able to maximise the opportunity. However, the next step is to be able to predict such hits rather than react to them.
That is where Big Data really comes in. Being able to monitor various parameters, whether for the next YouTube hit, emerging technology or must have device, service or accessory, would allow companies to get the drop on competitors, gain competitive advantage and reasonably predict what might be a hit.
This would mean that when a company, for argument’s sake, develops a new tablet that is a bit different from others and has a value proposition all of its own, the business analytics can look at the parameters and the market and make a reasonable prediction as to early sales which can then be used to make an initial order based on projected demand. This may go some way towards allaying the fears of manufacturers who do not want to be left with unsold stocks.
But as I began, speed is all relative, and the march of Big Data it seems is not so all conquering just yet. According to a survey report from Colt in the UK, the general consensus from 360 financial sector professionals was that the speed and accuracy of social media sentiment analysis presents barriers to predicting stock price movement. The survey said that only 7% of professionals would consider using social media sentiment as a primary tool for predicting stock price movement.
The majority of finance professionals (63%), expressed the belief that tracking public sentiment on sites such as Twitter, can be directly linked to the valuation of individual stocks, meaning that there is some value in the sentiment analysis, even as a trailing indicator. But, it appears as if this source is not yet accurate or fast enough to be a primary indicator. A third of respondents said that the speed of data gathering could cause problems while 43% said that they would struggle to respond quickly enough to the data.
This highlights a number of issues in relation to Big Data, business intelligence and their usage in guiding business decisions. The old adage of ‘garbage in, garbage out’ still applies, irrespective of how widely that garbage is gathered and fed.
Now, I am not necessarily saying that social media data is garbage, but rather unless the tools are equally intelligent in determining the wheat from the chaff as they are in how widely and how quickly it can be gather it, the result will be less than desirable.
The conclusion must be that if BI tools are being applied to anything other than most common tasks for which it was developed, namely the likes of transaction data, financial trading and the like, it is really a case of trial and error, until such trials can prove the effectiveness of ‘I’ bit of the acronym.
Ltd
