Insight ON How MERGE Turned Their Own Data Problem Into a Product Their Clients Needed Too

Audio transcript:

How MERGE Turned Their Own Data Problem Into a Product Their Clients Needed Too

Jason Dittmer (00:03):

The whole idea of Drink Your Own Champagne is that we want to make sure that the technologies that we're bringing to our clients, we have already done internally. So we're not trying to ask them to invest in some sort of vaporware that people are coming up with. We can come to them and say, "Look, we've already done it ourselves. We've stumbled through it. We've figured this out. We've been successful here. We fumbled down there and we bring this to the table so that you don't have to go through those."

Jillian Viner (00:30):

Drink your own champagne that deserves to go on a poster. Welcome back to InsightOn. I'm Jillian Biner. My guest today is Jason Ditmer. He's the SVP of TechOps at Merge, a marketing and technology agency that lives, as they put it, at the corner of health and wellness. Jason has a background that spans nursing, development, and UX design, which kind of tells you something about how he thinks about problems. And the story he's going to tell you today starts with a data mess that every organization will recognize. Fragmented sources, manual normalization, slow insights, no way to ask a question and actually get an answer. But what Merge did about it, Entirely and Google Cloud turned into something much bigger than an internal fix. And the philosophy behind it is something that any organization can steal starting tomorrow. All right, let's go. Jason, it's so good to have you on the podcast.

Jason (01:24):

Thank you very much for having me. There are so many fun things that have been coming up here. I'm excited to talk about all the things that are going on.

Jillian (01:32):

Before we dive into anything, let's just lay the foundation here. Tell me a little bit about the company you work for, what you do, and maybe a big decision that's on your desk right now.

Jason (01:42):

Yeah. So the company I work for is Merge. Merge is a marketing and technology agency. We like to say that we're built different. That's sort of our tagline. So as part of that, we say that we live at the corner of health and wellness. We like to take storytelling through technology and with that we can create more engaging experiences for our brands, which connect in a much more effective way with their users. So most of what we do is really pushing more on the technology side, the marketing technology space, which Google plays a huge role in. My day-to-day is supporting all that backend technology. So whether it's the end users from a help desk perspective, helping put together different systems within the agency if we're connecting through GCP. We built some products recently that just made it onto the marketplace. We're excited about that. But if you're asking me about what I think the biggest decision on my desk right now is really trying to work through a lot of the AI governance portion.

(02:56):

I know I heard a lot of it here at Next, but trying to work through that and create the guardrails that are going to allow us to execute some of the things that have come up as part of our humanity suite. I think we mentioned that before. That is going to allow us to continue to support the AI technologies, not only for us internally, but also for our clients.

Jillian (03:17):

In a previous conversation you and I had, you shared an expression that I love. I've heard it in other places, I love the way you described it. You said we drink our own champagne and your C-suite is very bought into this idea. What does that actually mean at Merge?

Jason (03:32):

Yeah, so that is a mantra of our new CEO, Stephanie Trunzo. She's fantastic by the way. The whole idea of Drink Your Own Champagne is that we want to make sure that the technologies that we're bringing to our clients, we have already done internally. So we're not trying to ask them to invest in some sort of vaporware that people are coming up with. We can come to them and say, "Look, we've already done it ourselves. We've stumbled through it. We've figured this out. We've been successful here. We fumbled down there and we bring this to the table so that you don't have to go through those." And I think that has been a big differentiator and has actually resonated with a lot of our clients in terms of how they are now bringing us to the table more to talk about AI technologies. And it's not even just about AI.

(04:19):

I mean, anything we're doing, whether it's a strategy that we're executing. For example, we just revamped our whole website as I mentioned, we launched the Humanity Suite. With that, we put in all of the technologies that we are asking clients to do, whether it's SEO, whether we're doing marketing campaigns through a 360 through that website. All of that we're doing ourselves so we can use that as a case study to walk into your client and say, "Look, we're drinking our own champagne. We think you should have some too."

Jillian (04:49):

I wish we had some champagne here with us today.

Jason (04:51):

I would take it.

Jillian (04:53):

I want to ask you about the humanity seat, but before we get there, talk to me about maybe the hardest or a problem that you faced internally that you solved and where that led you.

Jason (05:06):

Okay. I think the big one I'm going to talk about, if you don't mind, because it actually goes with the marketplace offering that we put together.

(05:22):

I am at heart a technologist. I do have a background in nursing, so I do have that empathy healthcare portion of it, but at a heart, I'm a technologist and as part of that is all about data. And in marketing, data is really a huge driver. We're seeing it more and more. Everybody talks about Web 2.0 being social. Web3. I actually think Web 3.0 is about data and our ability to leverage that in a way that we can message better to not only ourselves internally, but also to our clients. So again, I'm mixing in a little bit of the drink your own champagne as part of this as well. For us, trying to wrangle all of that data that's coming up, there's just so much of it. And to be able to normalize it, transform it, dive into it and get insights from it was a huge challenge that we had.

(06:20):

And talking about the website work that we did internally that we were generating so much data, we had to figure out a way to feed it. Being part of the Google Sphere, we went to Google right away. We know we can pump this into Bigfory. We know we can use Looker. We've got Gemini on the backend. Maybe we can start to use a litle bit of that to ask some questions. And I think that I'm not a think, I know that that was the impetus to start us creating these data pipelines so we could take data from a bunch of different sources, quickly feed it in and get contextually aware responses to that data that allow us to execute against that change campaign tactics, quickly respond to something we're seeing in the news or something that happens in the news that we can be relevant in our postings.

(07:08):

With that, we saw the need or we were hearing about this similar need on clients. So we took that back and we said, "Hey, can we use this to leverage to solve this problem for our client as well?" And that's where our automated data operations product came out of. And we actually have one client on it and it'll already, I can't name the client unfortunately for confidentiality reasons, but with this product, we are now able to very quickly take from about 30 different disparate data sources for this healthcare client, none of which the data looks exactly the same. And in very short order, pump it in, normalize it, transform it, and we can already start asking questions about that data within minutes or hours of it being actually in the system than you could have before. That time to normalize that and load it in would've taken weeks at a manual process.

(08:07):

Or you had to have so much money to throw out a problem like this to make it to fix it that it really pushed a lot of people out of the market. I think this new product is actually going to help us get to that type of level with these clients in a way that you were never able to connect with them before.

Jillian (08:27):

There's so much to unpack in that. The data sprawl and the richness of data that companies are sitting on is such a common pain point. Every organization has a wealth of data that is untapped because they can't access it, organize it, consolidate it. The fact that you were able to solve that to solve an internal problem and then turn it around and turn that into a product for a customer is phenomenal. What was the hardest part in that? Because again, this is a problem every company has. So what made this possible for you to solve?

Jason (09:00):

So the hardest part was figuring out, Franco was figuring out how to architect it. The easiest part came out of that because we decided to do it in Google. The whole product is 100% built in GCK from the data ingestion point all the way through to the presentation end, which is in Looker. I mean, we've got BigQuery in the middle. Laying on top of it, we've got Google SecOps as our security platform overseeing all the data that's coming in. So we can do DLP on this. We can look for mask data. We can look for PHI, PII, kick that out or segment it in a way or mask it if we need to in very quick order.

(09:45):

I don't think we could have done what we're doing without Google technologies because Google, and I may be wrong about this, but I think Google is the only full stack available that is 100% contextually aware of all the data that is within Google. A lot of the problem you run into when you're dealing with disparate data sources, outside data sources is that you lose that contextual value. So if I'm tapping into another source, I don't want to name names, I'm sure we can all think about different large data sets, data lakes that you could tap into. We know the players that are out there, right? You miss that contextual portion of it that you get from the full Google stack. And that's because we had all of that in on place, we were able to then build that within GCP and have all that contextually aware data and that's what makes it so much quicker.

(10:39):

The fact that we can bring it in and we have that information, the turnaround time and the ability to then visualize that at the end, that's what shaved days, weeks, hours off of the whole process.

Jillian (10:50):

How long did it take you?

Jason (10:53):

So the first round where we built it for ourself actually took about, I would say three months, believe it or not, which I still think is fairly fast because you figure we were figuring out how to architect it, what data we actually wanted to bring in. You had to figure out all your data normalization of which frankly Insight helped us with. We contracted with you guys as a third party to help us with the data normalization portion of it. And with that, then we could actually go to the next step. Our DYOC, Drink Your Own Champagne version of this was much smaller scale than what we would need to do for a larger provider, whether it's a consumer side, whether it's a healthcare side and the healthcare client that we worked for is definitely much larger than what we were doing. So we had to figure out how to expand that and even within there, that's when we added in a lot of the Gemini layer, which allowed us to start to ask questions of the data as it was coming out.

(11:51):

I'm not just visualizing it now in Looker. Now I can start to ask those questions like, what am I missing? What is the opportunity that isn't readily visible to me by just looking at the dashboard? These are the questions that having contextually aware data brings to the table that I think is harder if you're pulling it from so many different sources into one place.

Jillian (12:14):

Was there an insight that you can share that surprised you or something that you would not have picked up on before

Jason (12:20):

This? So again, I can't talk about the client side stuff, though I can tell you there were a couple of aha moments where we're like, oh, that's interesting. I don't think we would've necessarily gone down that route. But when we're looking at the data that we were pumping out of our particular website, we saw that by changing some of our messaging around how we were talking about AI in terms of the human factor really resonated with the base of people who were coming and checking out and hearing our message. I think not that was the impetus, but that helped feed the further discussion around the humanity suite in terms of our focus on it being a very human led endeavor from a marketing perspective versus just 100% data ladder.

Jillian (13:13):

Okay. So give me the, you mentioned Humanity Suite again. I want to make sure that our listeners understand what you mean by Humanity Suite. So give us that pitch.

Jason (13:20):

So this is our way of thinking about how you need to rethink marketing and marketing technology. A lot of agencies have been running the same models for a very long time. I've worked in advertising for 26 years now, so now I'm dating myself. Jason, you

Jillian (13:41):

Don't look old enough to have all these

Jason (13:42):

Different

Jillian (13:43):

Careers.

Jason (13:44):

That AI, and I know I hate using this word disruptor, but AI is a bit of a disruptor. It's definitely a game changer for marketing. It's one of the things that I actually love about marketing is that there are so many new things that come to the table and we have to embrace them quickly or you fall behind. And that's where the humanity suite comes in. We realized that we had to embrace AI as part of this, but because we're at that intersection of health and wellness, there's that human portion to this and it can't just be a mass. The data says this, so I'm just going to push a mass message out to people. What AI allows us to do now is have more individualized conversations with people. I think we like to call it infinite individualism where in the past, like I said, large messaging tried to hit a bunch of people all at one time, but now with AI, I can actually target in same number of individuals, but individually meet them where they're at, have some of that anticipatory empathy available that really draws them in and makes them part of your brand.

(14:54):

So as part of the humanity suite, AI is a part of every portion of it. It's divided out into two different sections, which is a foundational core and then an impact core. And each one of those has three portions underneath it. Foundation, as you can imagine, is about how you look at AI, data, the technologies that you're using across your marketing stack. And that's not just us as an agency, that's how our clients who are also investing in marketing technologies are utilizing those and then looking at the impacts. I bring this in. Now I have the data. How am I looking at it? What am I deriving from this? And we have three other different portions underneath that impact section as well. But the human factor always being a large part of it with AI helping that, not displacing it.

Jillian (15:43):

So connect the dots for me. The work that you did earlier to solve the data disparity problem led to this humanity AI platform.

Jason (15:52):

I think some of the data that we got out of the website information that we gleaned from that helped. I'm not going to say there was an output, but it definitely helped us formulate this a little bit more and get us to this point like, this is where we need to go.

Jillian (16:07):

Would this have been possible a year ago? Was this on the roadmap?

Jason (16:13):

I think a year ago that quick of a turnaround would not have been possible. I'm not saying we wouldn't have gotten there or somebody else wouldn't have gotten there. I do think it would've taken longer. Some of this, when you talk about the human factor in AI, it really has to come down to the people that you're working with as well and what their vision is. Stephanie, our CEO, Kyle, our CTO, our chief data officer, our chief marketing officer, they all come from that background that they understand it from a human perspective. So you have to have that coming to the table as part of it. That feeds into it. Then you take the data, which is what we're talking about and you bring those pieces together. You have the human factor in the management and you've got the data that we're seeing from people how they're responding and we're making good decisions on that in a very quick turnaround.

(17:06):

So to say, could we have done it a year ago? Maybe, probably, but I don't think we would've got there as quickly and we certainly wouldn't have been able to take the findings from something like that and have probably flipped that into a go- to-market product like we did a year, two, three years ago.

Jillian (17:25):

And what's been the impact of all of this?

Jason (17:30):

Well, it has definitely changed I think how people, at least within our agency, are starting to look at AI. We still have the people who are like gung-ho from the beginning like, "I want in, give me all the toys. I want to get in there." You had your people off to the side that were like, "Ooh, I don't know what this is. " But I think talking about it from that human perspective really starts to level that playing field for everybody. The comfort level starts to settle in. They realize it's not necessarily a job replacer. It's a way to help you as a human do something better or more efficiently or take something off your plate. I think that's probably been the biggest impact is just how people are starting to see AI differently and more comfortably.

Jillian (18:19):

So sticking to your theme of drinking your own champagne, what's the next champagne bottle, the next problem that Merge is going to solve?

Jason (18:31):

We work in some highly regulated industries and I think you're going to see a lot of, and I know we're working on this, I don't want to say too much, but we're working on a lot of stuff on how we help those regulated industries navigate the regulation bottlenecks in a much more efficient fashion. I think for us, that'll be the next big thing that you're going to see coming out of merge for sure. On top of, I think this message, the humanity message is going to start to resonate more and more for people. I think it's going to continue to grow. It's a portion of the AI piece that nobody's really talking about a lot, but I think as more people do and hopefully as our message continues to grow, you'll see that as another coin.

Jillian (19:22):

I love that you've taken an internal problem, turned it into a marketable solution. You're thinking about the connection between AI and the humanity approach. What would be your advice to any leader who is trying to find that transformative solution, that eureka moment with AI?

Jason (19:43):

I would say the funny part is it sometimes is just a eureka moment. You have to be part of what's going on and you can't be afraid of the technologies. You have to start to embrace them because you need to see what they're going to do. And from those, you see how it evolves and then the idea pops up. Maybe there's something we can do with this. You also have to be listening to the signals. There are plenty of signals that are going on out there. If you are in with your clients and you're in a particular industry or vertical, you're going to know what people are talking about and you just have to be aware enough of that to see where you can bring these two pieces together. I can't emphasize enough that people need to sort of get on the AI bandwagon. I know that's an easy thing to say because everybody is on the AI bandwagon, but by that, I mean, you need to get out of the pilot phase of things and we really need to start getting into like, we're making some actual product that we can push out there.

(20:48):

I think I saw a statistic. It's probably about a year old now, but it was something around like only 11% of AI projects that companies were putting out there actually made it into some sort of product. We've got to start to flip that. And if you really push and you really push for something and focus on a few things that you think are going to get you somewhere, that is going to be much more valuable than trying 50 different things that of 49 of which don't go anywhere. So the bottom line for me in this scenario is you really have to go with your gut. You're in this industry, you're working in it, you know you're vertical, your gut knows what you need to focus on. Just trust that and play that out to its extent and commit to it. And I think you'll see good results.

Jillian (21:37):

Strong advice. Jason, thank you so much for your time today. I really appreciate it.

Jason (21:41):

Yeah, no, it was great. Appreciate it.

Speaker 3 (21:43):

Thanks for listening to this episode of Insight On. If today's conversation sparked an idea or raised a challenge you're facing, head to insight.com. You'll find the resources, case studies, and real world solutions to help you lead with clarity. If you've found this episode to be helpful, be sure to follow Insight on, leave a review and share it with a colleague. It's how we grow the conversation and help more leaders make better tech decisions. Discover more at insight.com. The views and opinions expressed in this podcast are of those of the host and the guests and do not necessarily reflect on the official policy or position of Insight or its affiliates. This content is for informational purposes only, should not be considered as professional or legal advice.