Big Data

Digital Marketing - Study Notes:

What is Big Data?

Big Data describes the huge volume of data – both structured and unstructured – that can inundate a business on a day-to-day basis. This is because big data involves datasets that are too large to store on standard computers, and it requires multiple computers or servers to work together to process the volume of data. In turn, you can use big data outputs to predict better results and forecasts than would be possible with smaller datasets.

Big Data comes from meshing, or combining, different datasets to unearth core insights about your customers and business environment. It’s the coming together of several different tools and techniques. 

Big Data analysis involves several key components and technologies: 

•      Data quality software: This cleanses your data by de duping and cleaning up the data sets, making it more reliable when considering what insights can be used.
•      Integration tools: These tools merge data from external environments with your own internal data in order to create one dataset to glean insights from. 
•      Data visualization: The way you present your data can affect the success of a strategy pitch. Often, using visuals like charts and graphs to represent data can help the various stakeholders easily grasp the key insights. 
•      Distributed file stores: Think carefully about where you store your data. It can be stored internally on your own servers, or through a cloud-based server depending on your set up.
•      In-memory data fabric: Big Data can hold longitudinal data – in other words, data that is timely. You can span it over a number of years. But there are limitations to this, and you need to make key judgments about the data you choose to store or not.
•      Stream analytics: Stream analytics tools can dissect and share the core insights you wish to extract from data collected. 
•      NoSQL databases: NoSQL databases collect information and analyze it for you. Your technical colleagues will be able to advise you on the best option for your organization

Some of the key technologies in this area are focused on predictive analytics. In other words, they enable you to use the data to understand how things may happen in the future.

Predictive analytics

Predictive analytics is software or hardware solutions that enable firms to discover, evaluate, optimize, and deploy predictive models by analyzing Big Data sources to improve business performance, or to mitigate risk. Effectively, this is about making core predictions based on previous behavior to evaluate what the likely future outcomes could be.

This is very important when considering buyer behavior. In the marketing world, you need to be intuitive about what exactly is going to happen, and understand the variables that have changed so that you can make certain assumptions about what might happen in the future. It’s also worthwhile using commonsense and intuition when examining the data to add the benefit of your own experience to the metrics as your own business knowledge will add depth to the numbers, making them more useful and insightful.

You can think of Big Data technologies as being like a jigsaw puzzle. It’s all about meshing technologies together. There’s an element of overlap, as well as linkages, because all of these technologies will need to come together to create a robust Big Data strategy. And if you don’t have these linkages, your picture might not be complete.

The Four Vs of Big Data

You can use various approaches when thinking about Big Data. One very useful framework involves the ‘Four Vs’:

•      Volume: the size or scale of the data
•      Variety: the different forms of data
•      Veracity: the trustworthiness of the data 
•      Velocity: the frequency of incoming data

Volume

Big Data, by its very nature, comes from a large number of data sources, and can provide greater volumes in terms of insights for individual customers. It’s the aggregation of the information that creates the richness within the dataset. In other words, it’s not just a single dataset. It’s about meshing different data sources together at scale to create the necessary volume.

Variety

Big Data includes a wide variety of data types and sources. For example, the data could be a rich combination of structured, semi-structured, and unstructured data.

Veracity

Veracity refers to the need to have robust and reliable data. Essentially, you must have some form of integrity within your data. Make it trusted, make it clean, and remove duplicates , because any anomalies that you have within your data sets will eventually lead to inaccuracies that you might not notice until after your initial analysis when it’s too late. Remember, robust inputs lead to robust outputs, and this applies to Big Data as well, so veracity is key.

Velocity

This is the frequency of new data being entered into the data set. It’s always better to work with fresh data and data sets with high levels of data velocity. 

The Four Vs is a useful framework for structuring your approach to Big Data, including how you collect and analyze it.

Statistics and advanced analytics

When interpreting Big Data, you should consider statistics and advanced analytics. 

Statistics

Statistics focuses on collecting, analyzing, interpreting, and presenting data. From a marketer’s perspective, the results are often interpreted to assess how likely it is that something will occur. Statistics are also a tool to help understand correlation versus causation. Basically correlation will show how one object or event relates to another, though it didn’t necessarily cause the other event to happen. On the other hand, causation can also show the relationship, but it will also show that one object or event caused another event to happen. At a more general level, the strength of the relationship and causing factors between data, actions and events can help you understand and interpret the data in a more meaningful way.

Advanced analytics

Advanced analytics focuses on analyzing data using sophisticated quantitative methods, such as statistics, both descriptive and predictive data mining, simulation, and optimization. It can produce insights that a traditional approach to business intelligence is unlikely to uncover.