The definitive 'how to' guide to data blending for business analysts

Data is at the heart of today’s interconnected world. It is captured in almost every aspect of our lives—television shows we watch, websites we visit, groceries we purchase, and opinions we share. As the growth of data continues to expand, so does the use of this data by organisations to better understand their customers, optimise their supply chain, and much more. Data analysts in the line of business have become the main driving force tasked with utilising this data to answer more complex business questions. 

These analysts know their business and understand what data is necessary to get the results they want—whether it pertains to sales, marketing, operations, or even finance. It is up to them to use this knowledge and business insight to make sense of all of the disparate data that resides in a data warehouse or data mart, or perhaps within a CRM system or a marketing automation system. It may even be social media data or spatial data—data that is becoming more prevalent and necessary to fulfil organisations’ business requirements. 

Many analysts find it difficult to address this new data challenge because their traditional tools and approaches are not robust enough to handle this environment. Utilising spreadsheets, manual processes, and custom scripting or relying on data scientists to build analytical datasets are all too time-consuming, expensive, and complicated in the face of the number of ad-hoc requests that analysts receive each day. 

An analyst must be able to extend their natural capability and creativity with genuine business insight. They must focus their strengths on high-level business questions, rather than the minutiae of spreadsheets and low-level SQL queries. Data blending helps today’s analysts take full advantage of their expanding roles, as well as the expansion of data needed to make those critical business decisions. 

In fact, the life of a business analyst should be a fun one. Less munging, and more doing. Nevertheless data blending is one of the stages where real magic happens when the analyst gets it right. 

It’s important to get the right process and the right tools to make the job as quick and easy as it should be, so analysts can spend their time drawing insights from the data and delivering actionable insights and offer real business decision-making value. 

Data blending defined 

Data blending is the process of combining data from multiple sources to create an actionable analytic dataset for business decision-making (such as retail site selection or multichannel profiling) or for driving a specific business process (such as packaging data for sale by data aggregators). 

Data blending is needed when an organisation’s data management processes and infrastructure are insufficient for bringing together analytic or specific datasets required by line-of-business groups. It can, for example, readily bring together disparate data, such as customer information from a cloud sales automation system (e.g., Salesforce.com) with clickstream web data stored in a Hadoop file system and segmentation models from Microsoft Excel. This is important, because while organisations aspire to have a completely integrated data management system, the majority of data required to make strategic business decisions still resides outside their IT-controlled data environment. 

Data blending differs from data integration and data warehousing in that its primary use is not to create the single unified version of the truth that is stored in a data warehouse, data mart, or other system of record within an organisation—and is typically conducted by a data warehousing or business intelligence professional. Instead, this process is conducted by a business or data analyst with the goal of building an analytic dataset to assist in answering a specific business question. 

Implementing data blending into the line of business can deliver greater benefits and deeper insight in hours—significantly faster than the weeks typically required for manual processes and traditional IT approaches. This time savings can be realised in the myriad business situations in which data analysts find themselves. Let’s look at a few examples where data blending can positively impact business decision-making. 

Fulfilling the requirements of data blending

Data blending empowers data analysts to incorporate data of any type or source into their analysis for faster, deeper business insight, but how do organisations enable a data analyst to perform data blending? Many line-of business analysts have abandoned spreadsheets and custom work projects in favour of a data analytics platform because it fulfills the needed blending requirements or a modern organisation. 

Understand the progression of data 

To self-service, analysts need a drag-and-drop workflow environment that allows them to build out analytic data sets the way they think. These let the analyst understand how data progresses through the process without any ‘black boxes’, and quickly identify where issues may lie. Drag-and-drop allows analysts to focus more on the data and less on the technology by eliminating the need for coding or programming - meaning more time for analysis. 

Enable direct access to data 

Make sure analysts can use whatever data they want, without needing to rely on overworked IT staff or data specialists. There is a huge range of data available for use, including: 

  • Local data (spreadsheets, user device generated data, enterprise data warehouses, etc.)
  • Third-party data (Dun & Bradstreet, Experian, Tom Tom, census, etc.)
  • Cloud/social data (Twitter, Facebook, Salesforce.com, Marketo, etc.)
  • Other analytics platforms (SPSS, SAS, Google Analytics, etc.)

Simplify blending of data 

Users need complete flexibility in joining multiple datasets. The tools or platform employed should allow data of any type or level to be brought together. Data should be joined at both the record and field levels, and it can even be expanded to include multiple key fields with the right know-how. 

Automate and repeat processes 

With the amount of ad-hoc analysis required by today’s analysts, what if there were a way to make this process easier, faster, and repeatable? Analysts should ensure they create and use workflows that can easily be saved and repeated for further data blending, processing, updates, and analysis. Updating the analysis or report should be as simple as updating the data input(s). 

Output data easily 

Once the heavy lifting of data blending is completed, analysts need to implement this data into the right processes of the business. Resulting outputs should be pushed back into a database, incorporated into an operational process, analysed further using statistical, spatial, or predictive methods, or pumped into static reporting or a visualisation software. 

While traditional data analysts use traditional IT tools to generate reports on historic data, today’s analysts must extend that capability with their business insight and natural creativity to find information their organisations really need. With improvements in information technology and the constant influx of big data, a flood of new opportunities for business insight has appeared. Empowered by next-generation tools, today’s analysts can now do what previous generations of analysts could only dream of doing. These analysts are able to perform data blending to create the analytic dataset they need to deliver the deeper business insights they require. 

Chiara Pensato, Director - EMEA, Alteryx Inc. 

Image Credit: Alexskopje / Shutterstock