Resources >

Guide to Data Analytics: How to Tell a Story With Your Data

December 28, 2020
-
Peter
Scobas
Download guide

The business intelligence lifecycle involves three major components:

  1. A data pipeline (to connect, extract, and transform all your data sources and integrate your data)
  2. A data warehouse (to effectively and efficiently store and maintain your data)
  3. Data science and analytics (to analyze your data in order to help your business make decisions - using reports, dashboards, visualizations, and other analysis techniques)

You can think of these steps like a necessity and sufficiency clause, seen in logic and mathematics: having a robust data pipeline and a well-organized data warehouse is necessary to effectively use data to make business decisions, but it is not sufficient. To round out the business intelligence lifecycle, you need to invest in data science and analytics.

Here, we’ll cover why data analytics is important for businesses, what question data analysis can help you answer, different types of data analysis, strategies for effective data analysis and more.

The importance of data analytics for businesses

The importance of investing in data science and analytics can’t be overstated. Having a data analysis framework that is aligned with a company’s goals and KPIs is invaluable. 

Once you’ve put together a data pipeline and decided on a data warehousing solution, you’re ready to move forward with the analytics component of the process. Data pipelines, data infrastructure, warehousing, schemas and organization are the foundational pieces. After that, you can focus on data analytics, which can be divided into two major steps:

  • Deciding what success looks like for your business and what metrics you need to track in order to determine performance (this includes things like KPIs, basic dashboards, and making sure you understand the business and important key performance indicators across different areas of your business)
  • Performing more advanced work, like predictive analytics and deeper analysis

Both of these components are important steps of the data analytics roadmap. And ut can be helpful to think of the relationship as a sort of “walk before you run” type of process. 

As a business, you need to take stock of what is important to keep an eye on - do you know what success looks like? What are your business’s most important performance indicators? Once you have done that, you can invest more time and dig deeper into your data and build upon your data foundation. No matter where you are as a company in your data analytics journey, you need a bottom-up approach. Think of it like a pyramid, with your data pipeline and data warehouse making up the base. Then you have your basic data analysis, dashboard, and visualization work in the middle and your more advanced analysis at the top of the pyramid.

The pyramid of effective data analysis

What questions can data analytics help you answer?

Data analytics can help support many different teams and business functions, from marketing and sales to finance, product and even business operations.

For example, your marketing team might be interested in visualizing the marketing and sales funnel. Is there a significant drop off at a specific stage of the funnel? How is the marketing team doing driving traffic to your business? Are many customers converting, and what helps them convert? If you are able to pinpoint where in the process your business is succeeding and struggling, you’ll be able to solve problems, identify opportunities and strengthen your funnel. 

Similarly, data analytics can be incredibly valuable in trying to quantify your company’s marketing or sales performance - tracking traffic sources, calculating conversion rates, attributing sales to the correct sources, and better understanding your customers are all components of marketing and sales analytics.

Business teams can also utilize data analytics for support with decision making. Maybe your marketing and sales teams are killing it - bringing in tons of new leads and customers with a fantastic conversion rate - but converting customers is only part of the customer lifecycle. Are you converting the right customers, and do you know how much they’re worth to you over time?

Two major key performance indicators during a customer lifecycle include churn and lifetime value. Converting and acquiring new customers adds to your growth rate - but if you are a subscription-based (or similarly-structured) company, the rate at which your subscribers are cancelling (or churn) can have a big impact on your ultimate bottom line. You want to make sure you have a high growth rate and low churn rate. 

Related to growth and churn is your customers’ lifetime value (LTV). Customer lifetime value is an incredibly important metric for a number of reasons. While it can be somewhat complex to calculate (correctly and with relative accuracy), once you are able to calculate your customers’ lifetime value effectively, you’re able to make much more informed decisions on how to identify and target the right potential customers (with the highest potential LTV), make sure that your customer acquisition costs do not exceed your customer lifetime value, and project out potential revenue based on your growth rate and your customer LTV. 

Lastly, your organization’s product and inventory teams should also rely heavily on data science and analytics. Understanding which products are most popular, which ones help you attract and retain the right customers, which are most (and least) profitable, and what you need to buy more of or discount to move more quickly are critical to ensuring a successful strategy across your entire business.

Different types of data analysis

Saying you’re investing in “data science and analytics” can mean different things to different people. In addition, “data science” and “data analytics” are two terms that are often muddled together and there’s a significant debate on what type of work is “data science” and what type of work is “data analytics.” 

In general, data science tends to be more technical than data analytics - a higher level of statistical and programming work typically goes into data science. For example, machine learning, regression analysis, and various statistical modeling work tends to fall into the data science category rather than the data analytics category.

Another important distinction with data analytics is the types of analysis involved in a specific data analytics project. In general, there are four major categories:

  • Descriptive analytics: This is the least complex (and in general, easiest to execute) type of data analytics work. Descriptive analytics helps you answer questions related to how your business is performing - like what’s happening in my business? Am I getting more customers? Are my costs increasing?
  • Diagnostic analytics: This level of data analysis is focused on the why - why did something happen the way it did for your business? Why are more users churning? Why are people bouncing from our website? Why is conversion rate where it is?
  • Predictive analytics: Predictive data analytics does exactly what the name suggests - it predicts. What’s likely to happen in my business in the future? Based on historical data, what predictions can we make? Will growth sustain through the summer? Will customers convert faster if we email them more?
  • Prescriptive analytics: Prescriptive analytics is often considered the next level after predictive - it’s about making recommendations. Based on the data and insights we have, what recommendations can we make? What do we need to do?
Different types of data analysis

How to perform data analysis

Becoming a data-driven organization takes time. While there are a few different ways to approach data analytics, depending on your business and your goals, there are some useful guiding principles that all organizations need to follow. 

First, it is important to take a “walk before you run” approach to data science and analytics. Some businesses err by diving too fast into complex data analysis, predictive modeling and trying to connect and parse dozens of data sources at once, rather than prioritizing what to start with. Ultimately, this strategy is likely to fail. As we mentioned before, you need to make sure you have a robust data pipeline in place as well as an organized data warehouse, first - one-off data analysis projects can often take up more time in the end. A data pipeline and data warehouse allows for more reproducible, scalable analytics processes.

With those components squared away, you need to put together a roadmap and start to prioritize what data analytics processes are important to you. Think through the data you have available to you, how you can transform it, what questions you want to answer, and which KPIs you need to focus on. What your specific company goals are, as well as what type of business you run, will determine what analytics and metrics you need to investigate. 

For example, if you are an ecommerce company, your top-line metric is obviously revenue - but that can be broken into smaller segments and analyzed in different ways that can help you determine what’s working and what’s not for your business. On one end, how many customers do you have? What’s your conversion rate? How many impressions are you getting? On the other end, what’s your customer lifetime value? Average order value? Do your customers have significant purchasing frequency? How much does it cost you to acquire a customer?

These metrics differ compared to what metrics a software-as-a-service (SaaS) company might be interested in. Since you’re a subscription business, you may be more interested in monthly recurring revenue (MRR). Broken down further, what about the average revenue per account? Or your lead velocity? Other important metrics include your churn rate as well as customer acquisition cost payback period, i.e., how long it takes to “pay back” the cost of acquiring a new customer.

Other effective data analytics techniques include cohort analysis or customer segmentation, which allows you to segment your customers and glean insights on a more granular level. Beyond the metrics, you can also move more into “data science” territory, where you can spend time putting together statistical modeling, predictive metrics, regression analysis, machine learning work, or A/B testing.

Visualizing your data + creating reports

Once you are working on a data analytics project, you’ll ultimately need to figure out how to create reports and visualize your data. And there are a range of tools you can use for this step. 

On one end of the spectrum, you have out-of-the-box dashboarding tools. In the middle, you have business intelligence tools with some drag-and-drop functionality and some analyst-extension capabilities. And on the other end of the spectrum, you have tools with far more functionality but that require quite a bit more technical skill to operate.

  • Out-of-the-box tools like Glew have less technical skill requirements and a shorter implementation time - if you’re looking for an ecommerce, product or customer analytics in a simple, all-in-one solution, these types of tools can be helpful
  • In the middle of the pack are business intelligence tools like Looker, Mode Analytics, or Tableau. They offer varying levels of drag-and-drop dashboard functionality, but also give data analysts the option to create complex queries and investigate and visualize the data in a myriad of other ways
  • Technical tools like a SQL editor and a Juptyer Notebook allow for complex data science and machine learning capabilities. However, these types of tools require a high-level of technical knowledge, like SQL, Python, and R
Options for data analysis tools


Determining the correct business intelligence tool for your organization depends on the questions you’re trying to answer, what data science and analytics methodologies you are looking to deploy, and whether you have (or want to build out) a data science and analytics team. Turning your organization into a data-driven company can happen in a number of ways, and deciding on the correct business intelligence solution is an important step in that process.

>>> Glew is one of the only data analysis tools available that includes both an out-of-the-box dashboard tool and custom reporting/querying capabilities. Learn more, or start a free trial.

Strategies for ongoing data analysis

Effectively utilizing data analytics to help derive insights and make data-driven decisions takes a big commitment and significant buy-in from the entire company. Being a forward-thinking, data-driven organization means you need to invest in a good data foundation - one that starts with a robust data pipeline and an organized data warehouse. Once those components are in place, you need to focus on your guiding data analytics principles: 

  • What does success mean for your business? 
  • What are your key performance indicators (KPIs)? 
  • What important decisions are being made in your organization that need to rely on data?

Data science and analytics is a critical element of any forward-thinking organization, and while there’s no “one size fits all” solution, there are a few guiding principles you should follow: 

  • Make sure you have a robust data pipeline to bring data in correctly and efficiently. 
  • Spend time developing an organized data warehouse solution. It will save you in the long run. 
  • And lastly, take a “walk before you run” approach to data analytics. Data analysis is a process - beneficial business insights take time, organization, and trial and error

Plus, further reading about other elements of the business intelligence process:

Get started with Glew.

Connect your data for a 10-day free trial today - no credit card or commitment required.