Request a quote!

Blog, News &

Case Studies

LearnBlog
What is big data analytics and why does it matter?
By E2E Research | September 30, 2021

Marketers and researchers use the word ‘analytics’ to describe many different things that can be done with digital data. Without a common understanding, it can be easy to misinterpret what a client actually needs and end up assigning project tasks to the wrong people, costing jobs inaccurately, and not meeting client expectations.

 

In this post, we’ll take a deep dive into the different interpretations this word can have to ensure that both clients and suppliers are on the same page when it comes to extracting relevant insights from from myriad datasets about buyers, brands, and businesses.

 

Before we get into the details, you might appreciate this short introduction to data analytics from The Career Force on YouTube.

 

 

 

Types of Data

First of all, let’s look at some types of data that business leadership, marketers, brand managers, and researchers have access to in order to better understand consumer and market enigmas.

.

Primary research data

Primary research data is generally considered ‘small data. They’re easily stored in traditional spreadsheets like Excel and the files are small enough to be emailed without getting stuck in your outbox or flagged as spam. These data tend to represents people’s opinions and perceptions about various topics asked of them in a quantitative questionnaire, or a qualitative interview or focus group.

  • Ad hoc survey or interview data: Often under 1000 records and under 100 variables. Normally focused on one brand or topic. Qualitative datasets converted to quantitative formats may have fewer records but much more, or much larger, variables.
  • Tracker survey data: When gathered across multiple brands or countries, may be up to 50 000 records and a couple hundred variables. Normally focused on one product category though they may shift in focus from time to time.

.

Business data

Business data is often created in passing – as something happens in the company, a physical or digital record is created. Created and stored over years and in many disparate formats, these records are used to fulfill customer requests, manage employees, or keep track of product development. In many cases, these data are left lying around, ignored on servers, collecting virtual dust, and not leveraged for the insights that lie within.

  • Employee data: Records of retention, satisfaction, reviews, salaries, promotions, complaints, departments and more can be transformed and standardized as variables for statistical analysis.
  • Customer data: This is where we start to use the phrase “big data.” Transactional data reflecting purchases, SKUs, prices, times, dates, and more can come in datasets of millions or trillions of records with thousands of variables. Click-stream data gathered from websites can be exponentially more massive as every tiny movement and action made by a finger, pen, or mouse on digital screens is tracked. These data are already collected in standardized datasets and ready to be reformatted or transformed into specialized datasets for analysis.
  • Business data: Executives are often most interested in these data – revenue, costs, finances, operations, inventory, supply chain, & logistical data. These data, also usually available in standardized datasets, are often summarized from individual level data but are even more valuable at the individual level.

.

Secondary research data

Secondary research data is all-encompassing. It can include any type of primary research or business data that were collected for some other purpose, whether by yourself, someone else at your company, or someone at a different company. As such, you might have access to small survey datasets, massive transactional datasets, or compiled and summarized datasets. In addition to the primary and research data already described, it could include:

  • Third party data: A huge range of data types and sizes can be purchased from third parties that create, curate, and collate many sources of data, potentially terabytes of individual or summary level data.
  • Social media data: Originally created to communicate a specific message to a specific person (or persons), social media data can be gathered and used for purposes other than originally intended. These data may include information about brands, people, and companies, date, time, geography, sentiment, and more. It may need to be transformed and standardized but a wealth of insights exist here as well.

 

Types of Analyses

There are three categories of analytics and skill-sets that might be required in the course of a research project. 

.

Standard analytics

Most quantitative market researchers have a broad understanding of the theory and application of statistics. They know when and why to apply certain types of analyses to achieve specific research goals. Specifically, they have a lot of experience interpreting massive data tabulation files and running standard survey analyses to help us identify patterns and understand what happened and why.

They focus mostly on:

  • Types of data: Primary data, usually quantitative survey data
  • Types of analyses: Correlations, t-tests, chi-square, means, standard deviations, ANOVAs, descriptive and diagnostic statistics
  • Analysis tools: Menu driven SPSS, Excel, data tabulations
  • Outputs: PPT reports, static Excel reports

.

 Questions to Find Out If This Is The Goal
  • Will the analyses focus on details from the data tabulations?
  • Do you need insights beyond what is covered in the data tabulations?
  • Do you need anything beyond descriptive statistics like means, standard deviations, and box scores?

.

 Possible Research Questions

 .

Advanced analytics

Advanced analytics are usually conducted by people who have specialized training and expertise in statistics. They are experienced with non-standard and special cases of statistical tests that can’t be determined from data tabulations. Advanced analytics can help us understand what happened, why it happened, and predict what is likely to happen next.

They focus mostly on:

  • Types of data: Primary research data, small business datasets, biometrics data
  • Types of analyses: All of the standard analytics plus linear / logistic / multiple regression, conjoint, MaxDiff, TURF, factor analysis, cluster analysis, segmentation, discriminant analysis, perceptual mapping, special cases of standard analytics, predictive analytics, forecasting, and more
  • Analysis tools: Menu or script driven SPSS, SAS, R, Python
  • Outputs: PPT reports, static or dynamic Excel reports, user-guided dashboards, simulators

.

 Questions to Find Out If This Is The Goal
  • Do you need to segment people or products into groups?
  • Do you need to predict purchases or forecast sales?

.

 Possible Research Questions

 .

Business Analytics/Intelligence

Answering business intelligence questions to improve strategic decision making and create a competitive advantage normally requires advanced expertise in both statistics and data management. That skill set is often described as data science. Of course, for maximum effectiveness, you would also want this person to have extensive experience with marketing and consumer data.

These experts focus mostly on:

  • Types of data: Big data, business data, transactional data, logistics, employee data, real-time or near-time data
  • Types of analyses: All standard and advanced analytics, plus data transformation and manipulation, data fusion, data mining
  • Analysis tools: Python, R, SAS, SQL, machine learning, AI
  • Outputs: PPT reports, static or dynamic Excel reports, user-guided dashboards, simulators

.

 Questions to Find Out If This Is The Goal
  • Do you need to combine different types of data from multiple sources?
  • Do you need to make sales or logistics predictions in real-time?

.

 Possible Research Questions
  • Why are we unable to keep warehouses stocked with the right products at the right time?
  • Where are we dropping the ball with our processes and logistics, and how can we solve small problems before they become big problems?
  • How can we increase our efficiency to improve our overall profitability?
  • When a customer has selected a single product, what other products would they be most interested in?
  • Can we drop a rarely purchased product without causing our highest value customers to switch retailers?
  • How can we ensure optimal inventory for every SKU using existing business data? – A case study

.

What’s Next?

There’s a lot of overlap among various analytical techniques and objectives. One project may require only standard analytics whereas another may require all of them. However, once the research problem and the available datasets are clearly defined (not as easy as you’d think!), your analyst will know which techniques and software are best suited to uncover your answers.

If you’re ready to gather top quality insights about your buyers, brands, and business, please do email your project specifications to our research experts using Projects at E2Eresearch dot com. We’d love to help you turn your enigma into enlightenment!

 

 

Podcasts about Business Intelligence

 

Business Intelligence and Data Analytics Conferences

Tags

Establishing a Strong Data Foundation for Omnichannel Excellence: An IQVIA webinar recap
By E2E Research | September 16, 2021

This post is a recap of a September 16, 2021 IQVIA webinar with:

Please watch the full recording here. Below are a few of the (many!) points that struck me as particularly important.

 

Introduction

Historically, customer, marketing, and business data were siloed and disconnected. Today, however, personalized omnichannel experiences that put customer needs first are increasingly essential. Even though new forms of data (e.g., email, text, Rx, claims, media, social, digital) each with their own data formats are added to the business requirements every day, all of that data needs to be well-connected with seamless accessibility. This is the only way that holistic data analytics and business intelligence can be carried out.

The majority of pharmaceutical of pharma leaders believe that improving their ability to connect, share and analyze data is key to their success. Not only do they need to access their own data, regardless of the department in which it originated, they also need to access data from external sources. Done well, these efforts mean that data is accessed more quickly and with fewer financial and personnel requirements.

 

How to Choose a Data Management Partner

  • Great partners are innovative and capable of influencing project or company goals even when your data requirements aren’t innovative.
  • They easily accept ambiguity in project requirements and, at the same time, deliver extremely exact outputs.
  • Successful partners are always testing new ideas and experimenting at scale.
  • They have extensive, relevant experience working with petabytes of data.

 

Key Future-Proofing Strategies

  • Assume a technology project is never complete. Rather than locking up the code, plan on regular improvements and updates.
  • Build components that can be easily separated and individually upgraded or improved, in anticipation of when the technology will drastically change – you know it will!
  • Know that what worked in the past won’t necessarily work today, and be open to new ways of doing and thinking.
  • Seek to understand what other people are questioning in order to drive collaboration. When you know what your colleagues are doing, you can accomplish more together rather than doing the same thing in two silos.

 

Cloud Technologies Are More Than Storage

  • Cloud technologies are the backbone of agility. They permit all the data to be dropped into the same place using the same formats allowing insights to be generated more quickly.
  • Cloud technologies aren’t just another place to store data. In the case of healthcare, they facilitate faster sharing of clinical trial data. That drives better lives for patients now rather than 5 years from now at the end of the next innovation funnel.

 

AI-Backed Technology

  • Artificial intelligence acts as an objective, third-party that can bring to light issues a person may not have thought of or isn’t comfortable saying.
  • AI is well placed to be an additional opinion or perspective to be considered along side all the other opinions.

 

What’s Next?

Regardless of your company size, every business has data that can be leveraged to generate better insights and grow a more successful business. From survey and focus group data, to transactional and financial data, it’s time to figure out which of your data sources need to do a better job of pulling their weight.

 

 

Learn more from our health care case studies

Learn more from our other blog posts

Tags