What You Need to Know About Marketing Data Lakes
Fun fact: I took this photo two years ago while exploring Colorado. :)
In the wise words of Seth Godin, “Marketing, more than a lake or a forest, is the landscape of our modern lives.”
I’m almost positive he meant that in a poetic way, but his statement could not be more true. Marketing data lakes are up and coming.
What are marketing data lakes?
Well, I suppose “up and coming” wouldn’t be an accurate way to describe them. TDWI states that data lake architectures have been around since 2015. The buzzword is just making its way around the sun.
With a quick Google search, you can find hundreds of articles on the origin stories of data lakes, and if you’re a bit more savvy on the programming or data-driven application side, you can even dig into how open-source software networks like Hadoop spurred its evolution.
But this blog is a quick-and-easy, bite-sized snack to help you digest what you need to know about marketing data lakes. And trust me, if your organization is focused on being a data-driven entity, you’ll want to.
The Purpose of Data Lakes
Data lakes allow organizations to store a large capacity of data — raw, unstructured data. This is valuable, especially when data-driven marketing campaigns that have changing objectives or time durations need this special flexibility when analyzing parameters. This flexibility with owned data gives companies the power to truly be data-driven. The data layering possibilities are endless.
This, however, should not be confused with data warehouses. Both systems do store big data as repositories, but data warehouses hold data that is structured. It is filtered and refined for a specific purpose. This is useful if you already know what you’re aiming to study, but more often than not, as companies are moving forward into adopting data into more and more decision-making processes, having fluid data sets may be more valuable for certain business objectives. For example, storing customer data in data lake systems allows marketers to pull from raw data and target by different parameters, such as geo-location, devices, social networks, and more.
If you wanted a 5-year-old-friendly analogy: Data warehouses are structured — each data piece is a particular, unchangeable piece that will be used to make a specific picture. Data lakes are unstructured — each data piece is a Lego block, waiting to build a masterpiece.
Check out this data lakes infographic by the EMC, and shared by I-Scoop.
So, Do I Use Data Lakes or Data Warehouses?
This is not to say that one is better than the other, or that you should only use one. Whichever data system you choose will depend on your team’s objectives and what your role is. If you’re more on the business operations or development side, you may not need unstructured data. It will save you time reviewing structured data from data warehouses, as the findings are already organized into a dashboard or spreadsheet. These resources will still reveal insights needed to make educated marketing decisions, in spite of being nonmalleable. (Oh, and unless they have a background in data systems, you should not give your CEO data lake reporting results. They’d probably be too busy to take on the added weight of sitting down to filter the data…and they’d look at you with a face that says, “What am I paying you for?”)
On the other hand, if you are a data analyst who can water ski beautifully along a data lake with navigating ease, sail away on that pool of unprocessed, raw data. You will have more leeway in adjusting your data repository, and you’ll be able to make tweaks relatively quickly. Last few notes: Data warehouses, due to their architecture, are usually more secure than a data lake. But with this, they are typically more costly and time-consuming compared to data lakes. According to Talend, sectors like Education and Transportation usually opt for data lake systems to be able to make forecasted predictions and problem solve, whereas the Finance sector often stores data as a warehouse model due to its security and cost-effectiveness.
Did your eyes glaze over with information overload? I hope not — we skipped a bunch of technical terms, haha. But now you can explain the basic concepts of a data lake and its beneficial use cases! Consult with your business development or IT team to see how data pools are being collated. It’s imperative to be aligned in enterprise adoption.