English
Perfection – the ultimate ambition of every business leader and marketing chief when laying out their big data solutions. Mathematical formulae which deliver attractive predictions, comprehensive facts and figures demonstrating improbably perfect correlations, and personalized ad campaigns with 100% conversion rates. Big data – when algorithms become goldmines.
Big data case studies
What do beer and nappies have in common? No idea? Don’t worry – an American supermarket has the answer. Through a clever use of big data and data mining, the supermarket was able to deduce that fathers who are sent to buy nappies on a Saturday often reward themselves with a six-pack of beer. Consequently, they decided to position the beer next to the nappies and sales went through the roof!
In production, the use of big data is known as industry 4.0. Sensors placed on machines are combined with other data sources to gather new insights to optimize production quality and speed.
An example from the automobile industry
Production line analysis carried out by a well-known German car manufacturer showed that errors and delays were more common when car type A production was followed by car type B. The cause was quickly identified: the process of changing tools and parts to suit type B production was cumbersome and awkward for workers when production was scheduled in this order. To rectify the problem, production was scheduled in such a way so as to ensure that type A production was only rarely followed by type B production, thus keeping difficult and error-prone tool switch-overs to a minimum.
Another good example of the use of big data can be found in the use of personalized customer targeting in marketing. Research has shown that people are more likely to book exotic holidays and trips when stuck indoors on rainy days. By combining this research with meteorological geo-data, travel companies can target their advertising more effectively.
(https://makeameme.org/meme/i-use-big)
Overcoming real-life hurdles
As wonderful as this brave new world in which machines assume responsibility for marketing and optimization undoubtedly is, the day-to-day reality is often somewhat more problematic. CEOs and marketing bosses are often confronted with considerable barriers when it comes to implementing big data solutions.
Before an organization can analyse and evaluate big data, it needs to collect it first – but loyalty cards, sensors and the necessary IT infrastructure don’t come cheap and need to be tailored to the individual needs of each business. Medium-sized businesses are often immediately put off by the initial costs – but larger concerns also encounter problems. Most large organizations already use existing IT systems which, although old, have been designed to suit the company but are rarely suited to the implementation of new data handling processes.
The acquisition of the necessary hardware for data gathering is one thing, but the selection of the right software for data processing is another. Software providers offer an unfathomable range of systems from high performance tools for every eventuality to specialist software for individual branches and tasks. In addition, any new software must be seamlessly compatible with existing systems whilst also fitting within a company’s budget. Executives quickly begin to realize that their dreams of fully automated big data goldmines come at a cost …
Even having overcome software and hardware issues, one hurdle still remains – the human factor. Data processing software can only achieve so much – the most informative analyses still need to be carried about human beings with outstanding statistical knowledge and a flair for data management. Data mining experts don’t grow on trees and such capabilities are understandably in high demand. So when switching to a more data-driven marketing strategy, don’t forget the less easily measurable human component.
Conclusions and recommended approaches
The possibilities and scenarios offered by big data know no bounds but when it comes to actual implementation, the initial hurdles are not always easily overcome. But it’s not all doom and gloom. There are an increasing number of tech start-ups who specialize in niche hardware and software solutions for small to medium sized businesses who want to profit from big data.
So-called big data labs exist to enable businesses to carry out case studies and test data-processing on a smaller, less risky scale before making a larger investment.
External data-processing experts and advisers can also be contracted to provide specialist guidance on the implementation of big data projects.
