Podcast

A guide to selecting marketing analytics technology

Callahan Agency | June 25, 2018

In the marketing analytics space, especially with technology, there are tons of tools we can purchase. Probably thousands. They’re all designed to handle this marketing analytics realm – that’s what they sell themselves on. But all of them do it in different ways. Some of them do one thing, some do five things and it’s really difficult to make a decision around which technology to use.

Your top two priorities are to choose a combination of tools that complement one another and to have the ability to customize as your business needs fluctuate. Learn how to select the appropriate marketing analytics tools for your brand in this podcast.

Listen here (or subscribe on iTunes,  StitcherGoogle Play, Google PodcastsPocket Casts or your favorite podcast service):

Welcome to Callahan’s Uncovering Aha! podcast. We talk about a range of topics for marketing decision-makers, with a special focus on how to uncover insights in data to drive brand strategy and inspire creativity. Featuring Jan-Eric Anderson and Zack Pike.

Jan-Eric:
Hi, I am Jan-Eric Anderson, chief strategist at Callahan.

Zack:
And I’m Zack Pike, head of analytics at Callahan.

Jan-Eric:
So today on the podcast, we’ve got an interesting topic. We’re gonna talk about how to select marketing analytics technology. We’ve talked a lot about tech stacks, technology that can assist in analytics. Today, we’re really talking about if you’re interested in building your own, what are the types of things that you would need to build, and Zack is extremely qualified to talk about this, way more than me, and he tries to dumb it down for me to understand it, and he was telling me earlier this morning that analytics technology is a lot like a car, and he’s a car guy, so would you care to explain to me how in the world analytics technology is a lot like a car?

Zack:
Yeah, yeah. So think about any car. Doesn’t matter if it’s your Toyota Prius or the dump truck that’s hauling gravel down the road. A car is a collection of parts. It’s a bunch of parts that are put together in a very specific way to accomplish a specific task, and some cars are more flexible in what they can do than others. An F1 race car does one thing really well. It goes really fast, and it turns left and right, and can do all that stuff, and it protects the driver really well, where your dually pickup truck can do a lot of stuff. I can drive that downtown, typically I can park at most places, but I can pull a giant trailer with cars and cows and horses and stuff on it. So I think it’s a really good analogy because in the marketing analytics space, especially with technology, there are a ton of tools we can purchase. There’s probably thousands. There’s definitely hundreds that are all designed to handle this marketing analytics realm, right? That’s what they sell themselves on, but all of them do it in different ways, and some of them do one thing, some of them do five things, it’s really difficult to make a decision around it.

Jan-Eric:
So okay, so I get it, I’m on board. I like the analogy. Thank you for that. So what are some of the components or things that you would want the tech stack to do, or your analytics platform to do? What are some of those components?

Zack:
Yeah, so we have a lot of data in marketing. It comes from a lot of different places. It’s very dirty, it’s disorganized, it’s messy, lots of people don’t know the clear definition of all of the different dimensions, and metrics inside the data, so at its core, whatever the technology is, it has to be able to handle that issue. That’s the number one issue in marketing is the messiness and disorganization of all of this data.

Jan-Eric:
So you need a closet, a pantry, a place to organize?

Zack:
Yeah, something like that. It’s a database. That’s the core, underlying piece of any marketing analytics technology you’re gonna buy, is some underlying database. I’ll relate this to what we did. I mean we have this product called the Intelligence Platform at Callahan. Our clients use it, they get a lot of value out of it, but it is just a big collection of technologies. The most important thing to know about this space, if you’re gonna buy your own technology and kind of build your own thing, is you will not be buying one thing. You will not be working with just one company to do all of this for you, because storage of the data is very different than organization of the data, and logic on that organization is different than those two things, and then visualization is different than those three things I just mentioned, so each of those, to do them right, require different pieces.

Jan-Eric:
Well clear something up for me because I was talking to a guy at a conference a couple weeks ago, and we were talking about analytics, and the intelligence platform, actually, and he said, “Oh, yeah, we’ve got analytics, we’ve got Tableau.” So is Tableau not a one-stop shop for analytics? If I’ve got Tableau, can I check the box, and say I’ve got it?

Zack:
Absolutely not. No, Tableau is like having a windshield to a car and saying you have a car. It’s like the pretty paint that’s on the side of the vehicle. It’s like the graphics you see on Fast and Furious, right? Going down the side of the Eclipse. It’s-

Jan-Eric:
Well a windshield is a very valuable part of a car-

Zack:
Extremely.

Jan-Eric:
So Tableau is a very valuable component, but keyword is component.

Zack:
Right. You don’t have a car if you just have a windshield, and I mean, on the flip side, if your car doesn’t have a windshield, you’ve got problems, right? You’re not driving that thing on the highway, for the most part.

Jan-Eric:
Right. By the way, we, as an agency, advocate for Tableau. We leverage that technology as part of our tech stack.

Zack:
Yeah. Tableau, for us, is our visualization layer, on top of many other layers. So with us, it started with figuring out all of the things that we needed to do from a data perspective. We knew we needed a database technology that could handle all of the messiness, where we could control all of the logic applied to that data, and the two best pieces of technology that are out, today, in my opinion, are Amazon Red Shift and Google Big Query. Those are two of the largest in this space, and they’re really well optimized for reporting and analysis, which is what we use this data for most in marketing analytics. There’s a lot of reporting that happens, but the real value is in analyzing and asking a lot of questions of that data. These two database technologies are designed to do that really well. They’re not great transactional systems. You wouldn’t hook your E-commerce website up to them, but for storage and analysis, it’s really, really good, and they’re relatively inexpensive, which is great as well.

Jan-Eric:
So that’s a base component, sorry to interrupt you, but that’s a base component of the stack?

Zack:
Yup.

Jan-Eric:
Gotcha.

Zack:
Absolutely. So that’s storage. That’s also … we’re applying logic in there. So if I’m organizing data, I’m trying to mash together two separate data points, or buckets of data, I do them inside of that platform. But I’ve got to get the data in there, right? So it’s great to have all of this data, but especially in marketing, we’re always making decisions on the data, we’re trying new things, and we’re measuring the value of those decisions on the back end. So I’ve got to have data flowing, and that’s where it gets really difficult in this marketing space because Facebook data does not talk to Google Analytics data, and Google Analytics data does not talk to your store sales information, and your store sales information doesn’t even … isn’t even in the same realm as your media data, they’re like totally different.

Jan-Eric:
Many data sources.

Zack:
Yup.

Jan-Eric:
The challenge becomes many data sources to get the data from those sources into your storage tech, so the data flow, or the data transfer is another tech, is another piece of the component?

Zack:
Yup, and there’s lots of options for that, and what we’ve chosen to do is use a partner that is kind of a custom shop where they will essentially build any pipe that we need. They’ll manage it for us. We work with them on all of the logic and security and everything that needs to be in there, but they specialize in this data transfer process. This is what they do all day, every day. So if the Facebook API changes overnight, they work overnight to fix it, and our clients, it’s seamless to them, there’s no issues, and so that’s really important. Now, there are a lot of systems out there where it’s self-serve, where you can go in through an interface and set up your own data pipes, and connections. Those typically require a lot more handholding. They’re cheaper, but it’s been my experience that they’re just not as valuable, they’re not as reliable.

Jan-Eric:
But as a CMO looking to build something like this or interested in building something like this, you have the option to leverage whatever existing expertise you have within your own organization. If you’ve got somebody who’s really into this, and knows how to do it, maybe the best option is a self-serve type of thing because you’ve got the expertise in-house. If you don’t have that expertise already in-house, you can find more of a managed service option where you are basically hiring someone else to keep up-to-date and maintain your data flow.

Zack:
Right, exactly. So that covers storage and flow of data, acquiring all of that data. So now you’ve got the data ready, right? It’s in that analysis ready format that we’ve talked about on a lot of these podcasts. Now I need tools to visualize it and analyze it. It’s called, you know, synthesize all of that data into something that a human can actually understand, and so for us, our visualization layer, like we said, is Tableau. We do a ton of analysis inside of Tableau, as well. It’s a really good business intelligence tool, but it doesn’t do everything, and what’s really important to know is Tableau is really well optimized for a lot of stuff but not everything.

So we also employ tools like Esri. Esri is the leading geo-spacial analytics tool. So we’ve had a podcast on this. We do a lot of store-level analysis, geographical analysis of data, and that’s where a tool like Esri, leader in the … it’s the Tableau of geo-spacial analytics. Esri is a really good tool for stuff like that. Then outside of that it’s people to leverage those tools, right? There’s other technologies that you’ll throw in there from time to time, and for varying needs, but back to the car analogy.

I can have the most awesome Ferrari that looks awesome, it’s got a great interior, it runs great, but the driver can make or break that vehicle, right? The driver, if they’re not good at parking in a small parking lot, the car’s not gonna look good, a year from now. But if I put a race car driver in there, someone who really knows what they’re doing, the car is like a different vehicle. So that is a very important piece of this equation, and this is where, I think, a lot of CMOs and marketing people in general get into trouble, is we get sold a bill of goods on a product. So the sales person comes in, they’ve got their marketing analytics technology, in many cases they’re saying that it handles the full end-to-end piece of the spectrum, right? You want your sales data in here, no problem, you want your media data in here, no problem, and oh yeah, we’ve got a visualization piece, as well. It’s an inclusive package.

So it sounds great, right? There’s a couple of very important pieces to that equation. One of them is you have to have someone who understands it inside and out. Even Tableau, which is just one component of our platform, someone has to know that tool inside and out to really be able to leverage it, and that takes years and years of experience. You can’t send someone to the Tableau seven-day training course, or a local two-day immersion course, and expect them to be able to leverage that tool the way it should be leveraged, so you’re getting full value out of it. They’ve got to be in it for years, and be able to experience all of the stuff they’re gonna run into, especially in this marketing space.

Jan-Eric:
So there’s a mastery of how to use the technology in and of itself, and this is kind of managing the entire kit. We talked, a little bit ago, when we were talking about data transfer and flow, about even with our own example, at Callahan, how we outsource that to a partner who specializes in that, and is always up-to-speed on API, and how those things are changing with different sources of data. But what you’re talking about, here, is the need for a person to have the ability to get the most out of the tech stack, and having in depth knowledge of each of the technology pieces, which can protect, then, against the investment should an organization and enterprise decide to invest in and establish this type of technology. You want to make sure that you have someone on board who knows how to take advantage of, and get the most for your money out of the tech that you’ve invested in, otherwise, you run the risk of probably using it to about 10% of its capability, which is way too often the case.

Zack:
These tools are not cheap, and so now you’ve spent a bunch of money, and typically what I’ve seen, when I’ve run into this, is the product sits on a shelf, for the remainder of the contract. The person assigned to use it isn’t allowed to spend the amount of time they need to really learn it inside and out, or whatever the reason is, they don’t quite get it, and then it’s like, well, geez, this tool doesn’t work like they told us it was gonna work, and I’m not happy with it, we’re locked into this contract for the rest of the year, or maybe there’s a six month out, or something, so we just shelve it, and we start looking for another one. Then I start talking to Tableau, and then Datorama, and all of these other companies that are out there selling this type of stuff, when you’ve got to really define the different pieces of technology that you need, go find the best in the industry, in each of those parts, and then put them together.

Jan-Eric:
So let me throw your own analogy back at you, and challenge you a little bit on this. You’re talking about the importance of people, of a human being, to drive this thing. Well you said that the analytics tech is a lot like a car. I don’t know if you’ve seen the trades, or if you watch anything on TV, but cars can drive themselves, these days, so why should I believe that if analytics technology is like a car, why do I need a person to drive it? Cars can drive themselves.

Zack:
Right. Computers are actually really good at driving cars, so far, right? There’s not that many bad examples of what we’re seeing out there. They’re way more effective than humans. The problem, though, in the marketing analytics space, is okay, so let me back up. Machine learning and artificial intelligence are two very hot items, right now, especially in marketing. We’re always talking about machine learning. It is the thing we’re closest to. Artificial intelligence, actually, is still a bit of a ways off, but even in the machine learning realm, it’s still not progressing as fast as a lot would hope, as fast as I would hope. If I could build something that would automatically tell us all these questions about marketing, I would not be sitting here right now. None of us would be. We’d be on a beach somewhere because we sold it for five billion dollars.

But the problem we have, in marketing, is our data is really dirty, it’s really messy, and it’s disorganized. For a computer to do something in an automated fashion, it has to have clean data. It has to know what to do, every time. We don’t have that in marketing yet. When the data is clean, machine learning can start to take over, and start to do things for us. So a machine learning algorithm, at its simplest form, is deciding yes or no. So the simple example is a picture of a dog or a cat. You can teach an algorithm to look at a picture and decide if that picture is a dog or it’s a cat, and the way that that algorithm gets to that conclusion is based on data. So you give it a set of stimulus, maybe 10 pictures of dogs, and 10 pictures of cats. You tell it, “This is a dog, this is a cat, this is a dog, this is a cat,” 20 times.

So now that’s its reference data set, you put the 21st picture in front of it, it has a 50% chance of getting it right. The algorithm then looks at its historical data and says, “Okay, I think that’s a cat.” Then it’s up to the human to tell them, “Oh, you were right, or you were wrong.” You do that millions and millions of times, and that’s how you get these algorithms that actually can predict that stuff. But the data is going in there as clean, right? You’ve got the human telling them yes, no, yes, no, yes, no, the whole time, so it can build that history. We don’t have clean data in marketing, unfortunately. We just don’t. If you talk to people who are in the machine learning space, this is their biggest issue, is I don’t have clean data. I spend all my time cleaning this stuff, and then there’s just not enough to build these algorithms to be able to do what we all want them to do. Will we get there? Yeah, eventually, but we don’t see many self-driving cars in Kansas, yet.

Jan-Eric:
Right, true, true. Well, okay, so from a practical standpoint, I think, in organizations that are trying to build marketing analytics, having someone on hand that can speak it, that knows how it works, is invaluable to, ultimately, the adoption of it, within an organization to getting the most out of the investment of building a tech stack, having someone who’s knowledgeable, can build credibility, and faith, and confidence, and trust that this marketing analytics platform is worth the investment, is key.

Zack:
That person will also be able to know when we can automate things, and when we can’t. Because, and it depends on the business, and what you’re trying to do. Some businesses have just so many variables that it requires lots of people to get in a room, and think through all these variables, and what the data’s telling them, to find the best solution, based on assumptions. Some questions can be answered relatively easily, right? Click-through rate on Adwords. If I’m optimizing to click-through rate, it’s pretty easy to know what keywords are doing the best and where I should optimize to them. In fact, Google does that for us. So I think that person, able to kind of navigate that stuff, will, one, make the technology more effective but will also make the analytics function inside whatever company more effective as well.

Jan-Eric:
Well, I think I’m convinced. You’ve convinced me that analytics tech is a lot like a car. If I could summarize, for what you’ve kind of laid out, there are different components to analytics technology, and it’s really about understanding how to select and where to invest in those types of components to fit your needs, but essentially you’re gonna build around a storage facility, which is a database that stores all your data. You’re gonna need a component that allows you to get data from different sources into that storage facility, and then you’re gonna need some sort of tool that will help you synthesize that data, analyze that data, visualize that data, to bring it to life for consumption, by a marketing team, and the fourth component is having someone, a person, staff, who’s credible, and has the competency, and the understanding of how to get the most out of the tool, and adapt those needs to the organization. Those are really the key components for establishing analytics tech.

Zack:
Yeah. The last thing I’ll say on that is your stack will probably be more expansive than that as you get into it. So our Intelligence Platform, yes, those are the kind of four key components, but there’s a lot of stuff inside of that, and it’s just really important to kind of start there, and then work your way out. Those are the core fundamentals, and be careful, if a company comes in and says, “We handle all of these.” Unless they’re a custom deal, and they’re a consulting partner, and they are managing that stack for you, which is kind of how we do it for our clients. If it’s a thing that I’m buying off the shelf, and it’s supposed to handle all of those, it probably does one, really well, and the rest just okay. So keep that in mind.

Jan-Eric:
That’s a great point.

Zack:
It can be important.

Jan-Eric:
Zack, thanks for your time, this has been great. He’s Zack Pike, I’m Jan-Eric Anderson. Thanks for joining us.

 

Thanks for listening to our Uncovering Aha Podcast. Callahan provides data savvy strategy, and inspired creativity for national consumer brands. Visit us at Callahan.Agency to learn more.