Big data for marketers: Why, how, and where to start

Marketing is getting more technical, and it requires marketers to work well with different tools and platforms and use data to make decisions. But, working with big data seems impossible if you’re starting from a more traditional, non-technical side.

The secret is if you’re just getting started, rather than jumping head-first and learning different languages, it’s better to familiarize yourself with the basis of data management.

In this post, I’ll discuss the best way to start working with big data.

Navigate this post:

• What’s big data?
• Why should marketers be comfortable with using big data?
• How can marketers work with big data?
• 8 steps to managing big data
• Additional resources to help you learn about managing big data

Watch the full session.

What’s big data?

Big data includes larger, more complex data sets that can’t be managed by traditional data processing software.

One of the biggest marketing data management problems is its complexity. There’s too much data in different formats coming from multiple marketing channels. Marketers are dealing with a large amount of data generated from different sources, and the question is how they can work with big data effectively and find meaningful insights.

Why should marketers be comfortable with using big data?

According to the analytics maturity model, every business will go through a data and analytics evolution. And at different stages, you’ll need different data and tools.

You may start with two or three marketing channels, and in that case, the default reports provided in the platforms or spreadsheet reports may be sufficient for campaign optimization and tactical reporting. But as your companies grow and your analytics capabilities mature, you’ll move towards more complex analysis, such as historical analysis, cross-channel budget allocation, customer journey analysis, etc., and you’ll need to work with big data.

How can marketers work with big data?

Many marketers think that to be able to work with big data, you need to master multiple programming languages and tools.

In fact, if you break it down, what you actually need to know are:

• Where to store your data
• How to explore your data
• How to feed data into your marketing tools

How to store big data for marketing

One of the first things marketers need to know about big data marketing is how data can be stored and structured. Efficient data storage is key to building a strong data strategy that can be used alongside business intelligence, analytics, and machine learning tools.

Typically, there are three ways you can start storing data:

• Data lakes
• Data warehouses
• Data lake houses

But what’s the difference between them?

Data warehouse

Data warehouses are used for storing, organizing, and managing structured and semi-structured data.

It typically stores historical and real-time data from multiple sources to serve as a single source of truth. There are two ways to feed data into a data warehouse—batch or streaming. Batch means sending data in batches, for example, once per day. In the stream mode, you constantly stream data between your data warehouse and your marketing tool, which helps you analyze data in real time. 

This solution is suitable for businesses that have a well-defined data model and focus on structured data analytics.

Data lake

A data lake can store almost everything —from images to PDFs to CSV files. It’s a good option if your business wants to store and analyze large amounts of structured and unstructured data in its raw forms. 

You would typically store data that will be transformed in the data warehouse. For example, if you were to use a machine learning (ML) service, you could tag images with a particular phrase, and these tags will be stored in one column with a link to the images in another column. This turns unstructured data (a series of images) into structured data (a table of data).

Data lakehouses

A data lakehouse combines the best of both worlds by allowing you to store and analyze both structured and unstructured data. Additionally, it offers the same analysis capabilities of a data warehouse, which makes it possible to gain insights from data regardless of its structure.

How to explore your data

Once you have worked out how to store your data, it’s time to put it to good use. Exploring big data is tough to do manually because it naturally consists of huge amounts of data. That’s why it’s important to visualize your data into insights that you can actually use.

There are many BI and visualization tools you can choose from. I want to emphasize that there’s a difference between BI tools and visualization tools. Visualization tools help you visualize the data, whereas BI tools will offer more sophisticated analysis. Many BI tools have capabilities to let you connect to the data warehouses.

How to feed data into your marketing tools

The next step is feeding data from your data storage into your marketing tools. This process is also known as reverse ETL (Extract, Transform, Load). For example, you can classify your audience into different segments and then feed that data into advertising platforms for better retargeting. 

Connecting marketing tools is usually straightforward, as most well-known marketing tools already have connectors to data warehouses. Some will expect you to give the tools access to the data, and then they take it from there.

It’s important to consider data security. Because you’re working with a lot of data from different sources— your web analytics data, your CRM, social media data—that you connect to your data warehouse and you send this data regularly to different platforms, you want to use the tools that respect the privacy of your customers.

8 steps to managing big data for marketers

Let’s look at the step-by-step process that marketers should use.

1. Define the use cases

For example, do you want to use data to recommend products? Or perhaps you want to use data to create better, more relevant ads for your audience. Once you’ve figured out your ideal use cases, then you can explore tools that make this possible.

2. Think about how to store and consolidate your data

At this stage, you need to think about the type of storage solution that will consolidate your data. Depending on your goals and data format, this could be a data lake, warehouse, or lakehouse. Spend some time comparing solutions that are popular in your industry. 

3. List your data sources

Think back to your use cases—what you want your data for—and list all the sources you want to pull data from. This could include email, CRM, social media, website data, etc.

4. Define how to get the data

Are there existing connectors between the ad platform and data warehouse? Where’s your stance on the build vs. buy decision? Would you rather have your team build the integration, or would you buy a managed pipeline from a trustworthy provider?

5. Clean and organize your data

Depending on the type of data and methods for storing, your data may or may not be in a usable form. If your data isn’t usable and is hosted in a data lake, for example, it will need to be cleaned and prepared so you can use it. You can use SQL for this to let your BI or visualization tools understand the data.

6. Define what reports you need to understand your data

To get the most out of your data, think about what types of reports you want to create. You may want to think about this before you select a BI or visualization tool if you need to create more sophisticated reports. That way, you can choose a tool that has the functionalities you need. 

7. Choose a BI or visualization tool

Visualizing data helps you turn numbers into meaningful stories. With data storytelling, you can present your data confidently and make it easy for your team to digest your reports.

You can choose a suitable BI or visualization tool depending on the type of reports you want to build and your data skills. Power BI, Tableau, Looker, and Looker Studio are the most popular options for visualizing and exploring data. You can find some comparison guides on the Supermetrics blog, for example:

• Power BI vs. Excel
• Looker Studio vs. Tableau

It also helps to have a basic understanding of SQL when exploring your data. SQL isn’t a programming language but a language that helps you communicate with your data warehouse.

8. Feed data from your storage into your marketing tools

When you have your transformed data, you can easily feed it back to your marketing tools. Most tools are ready to accept your data and start providing insights that you can use to fuel your marketing decisions in the future. Let’s say you store a lot of valuable data about product usage in your data warehouse. You can sync this data with your CRM system so the sales team can get full visibility into the customer lifecycle and identify expansion opportunities.

Additional resources to help you learn about managing big data

A good place to start with big data is to read up on all the terms mentioned above and to look at what other marketers do with big data.

For example, look at what tools people are using and compare their features while also checking that they’re compatible with each other and have all the functionalities you require to produce the data reports you need.
Another thing I recommend is to learn some basic SQL so you can better understand how data communicates between storage and other tools.

Here are my favorite resources to help you learn more about big data:

• To work with data warehouses: Exploring Your Ecommerce Dataset with SQL in Google BigQuery
• Connect to your data in DWH and visualize it: Explore and Create Reports with Looker Studio
• Use ML with your data in BiqQuery Data Warehouse: Getting Started with BigQuery Machine Learning
• Activate your marketing with data and ML: Instant BQML and Accelerate & Simplify Advanced Marketing Activations
• How to develop statistical intuition—a podcast episode with Tim Wilson.