Data lake versus data warehouse

Technology is changing at an exponential rate, and in such a rapidly evolving landscape, it is easy to get confused by the seemingly endless terminology and acronyms that come with those changes. So in this post, I have attempted to explain the difference to between two big buzz words being thrown around the traps. 

Data Lake and Data Warehouse are now commonly used terms within the world of big data and are often used as if they are interchangeable. However, they are not, and in fact, it’s the core function of each that differentiates them. 

Data Lake

A data lake is for the storage of unstructured, unorganised raw data; and its purpose and audience is undefined.

Data Warehouse

Data warehouse stores processed, structured data for a specific purpose and audience.   

Because of these differences, it is essential to understand why you are storing data to decide which is right for your business. They do not need to exist in isolation or as a single repository; however, most organisations will have a single data lake but might have multiple data warehouses.

Why Data Lake

• Uncertainty- if you are not yet sure what you want from your data, a data lake requires no structure or organisation to your data. Meaning you can store your data from numerous sources and in various formats until you have decided its purpose. 
• Machine learning - if you are using machine learning, then unprocessed raw data is malleable and can be easily analysed.
• Multiple business units - if different business units are accessing the same pool of data to meet different objectives. Keeping the data in its raw form allows it to be used by multiple Data Warehouses and manipulated to serve various purposes.

Why Data Warehouse

• Actionable data - because data warehouses store processed data, it can be structured to deliver more actionable outcomes. 
• Cost - with a specific purpose a Data Warehouses by design will filter out information that does not meet the business requirements and therefore will store less data, reducing storage costs. 
• Accessibility - perhaps the principal reason for a Data Warehouse is that by being designed, structured and organised to meet a specific use case, the data can be presented and understood by a broader audience.

Unfortunately, the strength of each is also generally the weakness of the other and as the world shifts to a customer experience economy, most organisations will discover they need both to meet the current and future demands of their internal and external stakeholders. 

However, as access to cloud storage and computing becomes more affordable, even small businesses have the opportunity to leverage the power of almost infinite storage and compute power. 

The key is to make sure you understand the “Why”. if you know that then you can articulate what you need for your business and that is what will guide your decisions.