You Ask, I Answer: Unifying Marketing Data?

In today's episode, Jose asks, what is your best advice for about collecting data from different platforms? How do you unify data for better reading? Is there any recommended tools? There's a whole family of tools and techniques for this. What you're essentially doing is what's called data warehousing.

Data warehousing is putting all of your different data in one place in some kind of common format that you can then extract out, parse, slice up, and things like that. So there's sort of three layers of data, right? So there's your databases, which is an individual source of data. There are data lakes, which are collections of data that you pull from individually. And then there's a data warehouse where you distill down what's in your data lakes into a single location.

A real practical example, you may collect social media data from all these different platforms, TikTok and YouTube and Instagram and stuff like that, but you may only want a few metrics from each service. You don't need all 200 metrics from Facebook, for example. You just need to know likes, comments, and shares. So using software, you would extract that information into a data warehouse. And ideally, um, the system that you're using probably is going to try and normalize and make it apples to apples so that uh a like on Facebook is equivalent to a like on Twitter from a data format perspective.

Data warehousing can come in a variety of different formats. You can completely roll your own with a system like uh AWS uh Redshift or Google BigQuery or IBM DB2. You know, take your pick of any of the the major technology players that provide these sorts of large-scale data warehouses. Um there are also off-the-shelf packages. Uh, these are typically uh, at least in the realm of marketing, fall under a category called a CDP or customer data platform.

And these are pieces of software, uh like treasure data and telium and stuff that will automate the collection of data from all these different data lakes into a central data warehouse. These software packages, generally speaking, are reassuringly expensive. You're talking tens of thousands of dollars a month to work with them. Sometimes, depending on the size of the implementation, it can be hundreds of thousands of dollars a month in order to get data from very very complex systems into one place. You may also, depending on the size of your company, have some kind of enterprise resource planning software, something like an SAP that has, you know, the SAP R3, I believe is the name of the software, that warehouses not just your marketing data, but your entire company's worth of data in one location.

Now, this can be challenging to work with if you are a marketer, particularly if you're not necessarily a technical marketer, but um it is certainly the way to get all your data into one place. Which avenue you choose, a boxed product or service versus rolling your own, depends on your technical capabilities and your budget. If you have budget, uh a lot of it, uh a boxed product will probably be the least painful because you will be outsourcing a good chunk of the technology, the infrastructure to a vendor to construct that data warehouse for you and make it accessible for reporting. If you have no budget, then you're gonna have to learn and roll your own, right? So you'll learn how to use uh a cloud-based uh data system, learn how to write uh code that can interface with the different systems and pool all that data together.

That would be what you do if you don't have budget. Um, if you don't have budget and you don't have take technical capability, learn the technical capability. It will serve you well in your career above and beyond just the company that you're working at right now because obviously, with ever exploding numbers of data sources, we want to be able to get to our data as quickly as possible and and adapt to the never-ending amounts of change that are happening in the industry. If you're able to uh cobble together some code to put those pieces together, you will be an extremely valuable person at your organization, uh possibly indispensable if you're the only one who knows how the system works. But the platform and tools are only one-fifth of the overall plan for unifying your data.

You've got to have a purpose in mind. What is the the unified system supposed to do? You've got to have people like you who are the talent who will make the system work, no matter which avenue you choose. You have to have good processes inside your organization that put in good data, because if the data going into all these different sources is no good, then obviously warehousing a bunch of garbage is just a warehouse full of garbage. Uh, it will not the tools will not help you improve that.

You have to use have good processes in place. And finally, you need some measure of performance, some way of knowing whether or not this effort that you're going through is actually worth doing. For many companies, a single view of the customer and a single view of your marketing data does have value if you're going to make decisions from that data. If you can look quickly and say, Yeah, we need to spend X dollars more on TikTok this month. That's a decision that if you have the agility from your data to make that decision quickly, you can have a competitive advantage over someone who can't look at the data or is confined to looking at channel by channel data individually and they can't see the big picture.

That's really what you're after is the big picture from all the unified data sources. So that's my general advice. Buy or build, but depends on your resources. Um if you have the money to buy it, buy it. If you don't have the money to buy it, build it.

Uh if it is part and parcel of your company's strategic advantage, if it is part of your secret sauce, I generally recommend people lean towards build because you want to control as much of your secret sauce as possible. You don't want a vendor to own your secret sauce or a substantial part of it. Um but a really good question, very challenging question. There's a lot that goes into this. Um data warehousing projects and enterprise data management is a profession unto itself.

And even for the smallest company, these are large undertakings. You know, at Trust Insights, for example, we're a three-person company, and still unifying our data and getting it to one place required a few months' worth of coding to get all the data and make it visible and make it uh workable. Um, and you need to have really good governance to make it work. But when you do, you're faster than your competitors, you're smarter than your competitors, um, and you can make your customers happier. So, really good question.

If you like this video, go ahead and hit that subscribe button.