Marketers, advertisers, and agencies face the challenge of scattered marketing data regularly. New marketing channels appear all of the time. This means that more data is available for business decision-making.
It becomes challenging for marketers to handle all this information. The data pool becomes too large and messy, slowing your reporting processes and impacting your decisions. In the age of rapid data-driven business decisions, these delays can be very costly.
So why is this happening, and what can be done about it? Luckily, there are several ways to tackle this problem. Let’s dive into the differences between centralized and decentralized data models, the world of data warehouses & data lakes, and when to consider moving your data to a centralized model.
Fast forward >>
What is the difference between centralized and decentralized marketing data
Centralized data access model
Data centralization is becoming increasingly popular. Cloud-based data warehousing vendors have made it easy for anyone to spin up a data warehouse in the cloud. With just a few clicks and a credit card, you can store and process an unimaginable amount of data.
As budgets grow, marketing and advertising performance become harder to measure. The need for internal visibility of marketing performance is driving companies to centralize their marketing data into a data lake or data warehouse.
Decentralized data access model
While centralized data models are growing in popularity, the benefits of decentralized data access models have been overlooked. Such models don’t require that raw data be stored in a centralized warehouse but rather give users direct access to the raw data they need.
What is a marketing data warehouse?
A marketing data warehouse is a cloud-based destination for storing and analyzing cross-channel marketing data. By consolidating data from multiple platforms in one place, data warehouses allow teams to analyze campaigns, create reports, and improve their targeting strategies all in one place.
Data warehouses consist of structured tables, making it quick and easy to query the exact data you want to include in your report or analysis.
Data warehouses consist of two main elements:
Data warehouses let you store a large amount of data in one place for an affordable price. Instead of relying on your marketing platforms’ retention policies, which can be restrictive, or paying for all the historical data you need from multiple vendors, you’ll have everything in one place and for a relatively low cost. Your storage capability will grow as your dataset grows.
In addition to storing data, data warehouses also support processing large amounts of it. If you want to scale your business by rapidly crunching more numbers, on-premise solutions won’t help you. With a cloud data warehouse, you can scale up and down quickly—which is critical for analytics because you’ll want to be able to quickly query specific datasets.
What are the benefits of marketing data warehousing
The major benefits of using a cloud-based marketing data warehouse include:
Creating a single source of truth
Marketing teams are often slowed down by scattered data because they either don’t have the time to sign in to a dozen different platforms to collect the data they need or spend so much time collecting data that they don’t have time for analysis and optimization.
Data warehouses can help marketers by consolidating their data into a single source of truth. This helps them get a better handle on important metrics like customer acquisition cost (CAC), return on investment (ROI), and return on ad spend (ROAS).
Time to insight
You can start centralizing your marketing data in a cloud-based warehouse without buying expensive hardware or getting access to a physical data center. Simply choose your data warehouse—e.g., Google BigQuery, Azure Synapse Analytics, or Snowflake—and start moving your data with a fully managed pipeline like Supermetrics.
And since getting started only takes a few clicks, you can immediately start pulling insights from your DWH.
To query data with SQL or feed data from your data warehouse directly into a data visualization or BI tool, you can process complex queries in seconds and push the data you need into reporting or analytics tools of choice.
Key analytics tools, such as Google Data Studio, can pull real-time data from your data warehouse instance without extra configuration.
Instead of placing your trust in the data retention policies of Facebook, Google, HubSpot, and other platforms, you can store your cross-channel marketing data in a warehouse.
This ensures that you always have access to historical data about your marketing campaigns, which will empower you to make better decisions about the future.
Cost & scalability
Whether you work for a growing SMB or an enterprise company, storing marketing data in a cloud-based data warehouse is relatively inexpensive. Additionally, having elastic storage means your data warehouse will always be ready to grow with your business.
Cloud-based marketing data warehouses are also known to require low to no maintenance since the cloud provider takes care of the upkeep for you. You just pay for the resources you use.
What is a marketing data lake?
A marketing data lake is a cloud-based solution for storing and consolidating your organization’s unstructured and structured cross-channel marketing data in its raw form— usually as CSV files. In the marketing context, cloud storage solutions such as Amazon S3, Azure Blob Storage, and Google Cloud Storage are often used as data lakes.
In a marketing data pipeline like Supermetrics, you can replicate data from the most popular marketing data sources—like Facebook, Google Analytics, and Salesforce—directly into your data lake of choice.
After that, you can move the data into a data warehouse for reporting and business intelligence workflows and give direct access to your data science team so they can get the data they need with whatever tools they use.
For example, here’s what your marketing data architecture might look like if you’re working in the Google ecosystem.
What are the benefits of storing marketing data in a data lake?
The major benefits of using a cloud-based marketing data lake include:
Better data governance
Managing data from multiple channels and departments in one place is easier with a data lake than it is with a data warehouse. For example, you can store all your Facebook Ads data in one cloud storage bucket and start a new bucket for LinkedIn Ads, Twitter Ads, etc. Alternatively, agencies can have a dedicated cloud storage bucket for each client.
Data lakes are also a great option if you want to keep access to historical marketing data that you may need one day—but you don’t want to clog your data pipeline or data warehouse with a bunch of metrics and dimensions you may never use.
Security and access
Most enterprises with stringent security standards would rather not have a managed data pipeline writing directly to their data warehouse. Even in a smaller company, you may have information in your data warehouse tables that you can’t share with any external parties.
If security is a concern, you can create a data lake architecture that allows you to avoid putting your data warehouse behind a firewall. A managed data pipeline like Supermetrics can automate your data transfers into a dedicated bucket in your data lake, and then you can move data between your data lake and warehouse with a tool like AWS Glue or Google Dataflow.
Similarly to data warehousing, you’ll own all your marketing data once you have moved it into a data lake.
This means that you don’t need to trust the data retention policies of Facebook, Google, HubSpot, and other platforms. Also, this ensures that you’ll have access to data about your past marketing campaigns, which will empower better decision-making in the future.
If you’re used to analyzing your data using SQL, you’ve probably noticed how easy it’s to make mistakes. As a result of a bad SQL script, you might lose access to some of the data you need for your analysis.
Instead of going back to your data pipeline tool to rerun the queries and waiting for the data to reload, a data lake offers a faster way to restore the lost metrics and dimensions. Rather than waiting for your pipeline to back up the lost data, you can quickly restore the data you need from the data lake and pick up your analysis where you left off.
The pricing model of data lakes is mainly built around storage and can have very low costs in most cases. This makes storing marketing data in a data lake an attractive option.
If you need to answer big-picture questions, you’ll usually want a centralized data model in the form of a data warehouse or data lake. This is because the volume of data needed to answer those questions will probably not fit in a single spreadsheet or a dashboard’s local memory.
Basically, you need a broad perspective on your data. To know how different marketing strategies have performed over several years, you’ll need centralized tools to answer that question.
About the author
Pieter is a Sales Engineer at Supermetrics. He works closely together with customers to identify opportunities for increased value return in their marketing data stack.