Ep 115 | How to Build a Cost-Effective Customer-Centric Technology Stack
This week Tisson Matthew, CEO of SkyPoint Cloud joins Allison Hartsoe in the Accelerator to talk about how to build a cost-effective customer-centric technology stack. From his executive roles at Amazon and elsewhere, he knows that customer data management is a very challenging, messy problem but one that must be solved to remain competitive. Listen in to learn about today’s customer data solutions.
Please help us spread the word about building your business’ customer equity through effective customer analytics. Rate and review the podcast on Apple Podcast, Stitcher, Google Play, Alexa’s TuneIn, iHeartRadio or Spotify. And do tell us what you think by writing Allison at info@ambitiondata.com or ambitiondata.com. Thanks for listening! Tell a friend!
Read Full Transcript
Allison Hartsoe (00:00): This is the Customer Equity Accelerator. If you are a marketing executive who wants to deliver bottom-line impact by identifying and connecting with your revenue-generating customers. Then this is the show for you. I’m your host, Allison Hartsoe, CEO of Ambition Data. Every other week, I bring you the leaders behind the customer-centric revolution who share their expert advice. If you are ready to accelerate, then let’s go. Welcome, everybody. Today’s show is about how to separate the signal from the noise and essentially build cost-effective customer-centric technology. Now, to help me discuss this topic is Tisson Matthew. Tisson is the CEO at Skypoint Cloud, a customer data platform driven by a very powerful, real-time customer technology stack. Tisson welcome to the show.
Tisson Mathew (00:55): Thank you, Allison. Appreciate it.
Allison Hartsoe (00:56): Tell us a little bit more about your background and how you started thinking about your customer-centric technology, particularly in the real-time aspect.
Tisson Mathew (01:06): Excellent question. I was the director of engineering at amazon.com. I was responsible for leading our team. So one of our biggest challenges was to manage our customer data as a very prominent customer-
Tisson Mathew (01:16): Centric organization. We have a lot of technology at our disposal, but the prime now aspect of things was very different than the traditional amazon.com delivery and convenience and selection mechanisms. So we have to build a data platform from its ground up and how we actually manage the data for customers and how their personal preferences are and what they want to buy, how to get to deliver it to them within the same day or same hour. So we were able to build up the platform and scale it and do it in a real-time manner. So the lessons I learned from that whole experience is that it’s a very, very challenging problem, as well as it’s very costly to do it right.
Tisson Mathew (01:54): So those are the learnings that I had from Amazon. Then after that, I was the chief technology officer of a fairly large healthcare company. And I also worked in a private equity firm. I saw the same pattern repeating multiple places. So quite a lot of hype around the subject, the idea of the right type of implementation, and doing it in a cost-effective manner was the roadblock. So the customers were getting challenged with vendors coming up with a variety of solutions, but it was mostly hype and wasn’t able to deliver. So the customers turn around for an internal implementation of a customer data platform, and it became very, very costly. So my thought process was to build a product that is very customizable by the customers, and it’s very cost-effective and gives them the fundamental building blocks of a great customer data platform. And it is real-time, and it is cost-effective, and it is scalable.
Allison Hartsoe (02:41): That’s almost like the Holy grail of having those elements together. I want to circle back to the Amazon side. Did I hear you right in that? You said it was costly to do it right. So I was trying to figure out was Amazon really cost-sensitive or did you just really notice how much it costs there to do it right and then kind of find a better solution the second time?
Tisson Mathew (03:01): Yeah, the latter.
Allison Hartsoe (03:02): Okay. That makes sense. And then I noticed you also defined customer-centric as personalization preferences, almost seamless tech. And yet there are tons of MarTech companies out there maybe 7,000, over 7,000 and over 3000 AI companies not to mention the major platforms like Salesforce, Microsoft, Amazon, Adobe, Oracle, and everyone is saying that they can solve the exact same center, like the exact customer-centric problems of unified, personalize, optimize. It almost seems like I don’t even need to care about what’s under the covers because everybody’s saying the same thing. Shouldn’t I just pick the lowest price, check the box and get that job done.
Tisson Mathew (03:47): Yeah. Great question. I think there is a lot of tools and technologies that are turned in front of customers without understanding their pain points. So it does like a knife versus a fork scenario where the solution to a problem would be different, right? So what customers are doing now is they go purchase a certain variety of solutions and inherently go back to their file system and Excel spreadsheets. And which is unfortunate because you are not getting the adoption within the customers. So the reason are a couple of things. One is all we’re promising on the capabilities and the value the product delivers to the customers. The second is it has been flexible to implement within a company. So implementation of a lot of this cookie-cutter product fall on its face because it’s inflexible to implement them. And the third aspect is it’s too costly and takes too much time. A lot of customers give up on it, they buy a product, try to use it and take out wrong and one at a time to customize and implement the drive for their organization for their use cases like, Oh, right now I’m going to go forward with this going back to Excel or something like that. So, yeah.
Allison Hartsoe (04:47): Do you think they underestimate routinely, not just the purchase side, but the amount of time that it takes to blend technology into the stack and then allow people time to adopt it and use it the training time?
Tisson Mathew (05:00): Yeah, I totally think so, but I think the vendors are under a time crunch of hit growth numbers. So rather than looking at a three or five-year horizon for success for an organization, they’re looking at, can I deliver ROI in six months and publish a case study in two months? So the value prop is not really aligned. For a customer, their value prop is how to build a great customer data is the foundation of an organization. So their value prop is how to build my customer data platform and foundation, right? So I have a long term competitive business. For vendor, they are looking for very quick when from the, can I get a case study out of this? Can I grow this licenses? So their incentives are not aligned. So until these incentives are aligned, it’s going to be very difficult for the pain points to be addressed and have a very longterm thinking on this.
Allison Hartsoe (05:48): You know, I’m so glad you said that because I’ve often felt that way about metrics and that whatever metrics are reported inside a tool, it’s because it’s easy for the tool to do that. And then say, Hey, we have analytics check the box, but it’s only because they can measure something note that that measure is meaningful. In the same way, you’re saying that the incentives are not aligned, and yes, the vendor wants to sell you the tool, but it’s not really about the deep, pervasive customer outcome, as much as it is. Did I solve your problem? Did I solve your problem? Did I check the box?
Tisson Mathew (06:20): Yeah, that’s right.
Allison Hartsoe (06:21): So in order to solve for a customer to create a real customer-centric solution, are we saying that having a variety of vendors cobbled together is a problem, and do I need to automatically get on a platform?
Tisson Mathew (06:36): Yeah, it’s a great question for a certain pain point. You need a platform type solution. So, for example, data management, especially customer data management, is a very challenging, messy problem. And there are so many different use cases that rely on that common platform, whether it could be sales or marketing or customer service or, or your product and what products you should build your business strategy. Quite a lot of it is relying on customer data. So having a platform that you can rely on for a very, very long time, it’s extremely important for that. But for certain use cases, for example, sending emails in a reliable, consistent manner to our customers, I think an email tool, it doesn’t matter what it is. It’s probably more of a point solution. So if you look at the overall technology stack, like for example, your infrastructure like a cloud infrastructure or a data infrastructure by the bow, that is your application. So their applications can be point solution, and it could be SAS and whatnot. The horizontal platforms, it’s SAS or not? Those are foundational. So that’s my viewpoint. So when you’re actually creating a technology stack, it’s very difficult to build all of them in house, but you still have to purchase certain products or services and integrate them and having the right strategy of integration and what should be foundational horizontal platforms versus vertical is the job of a CTO or CIO or an architect to decide and work with the business.
Allison Hartsoe (07:58): This makes a lot of sense. And this reminds me of conversation I had guess it was last summer when we were talking about CDPs. And one of the points that came up was a CDP might make sense, but then how do you integrate it into the larger stack and the larger the company? The more difficult this problem is, but I would almost say that it’s not just about large companies, smaller companies can get really great advantages or economies of scale. I guess if they build the tech right in the very beginning,
Tisson Mathew (08:30): That’s a hundred percent, right? If you think of traditional businesses and when they started doing e-commerce, they stood it up as a separate business or a separate entity where you couldn’t return a product you bought online to the store, right. It was like; they didn’t have any idea of who purchased what on an online store. Now, cause it’s just, I like, so we can buy from the store, you can buy online, but the same product at the same customers in the chain need a holistic, centralized solution. So a lot of use cases in the company started off as a departmental thing. And then when it comes to customers, it has to be a company-wide thing.
Allison Hartsoe (09:02): So are there examples you can share about where the foundation was real time and cost-effective and the kind of benefit that accompany got by dialing it incorrectly?
Tisson Mathew (09:12): Yeah, absolutely. I think a couple of examples, we are working with them, major league sports team. Their challenge is they own multiples sports teams and inherently. They all play in the same arena, so they had to figure it out. What is the cross-section of those customers rather than trading? Does it team number one customers versus team number two versus team number three. What is our common customer base? What are their preferences? So rather than creating three different solutions for this, have one common solution. The second is the in-stadium experience, and out of stadium experience would match. So customers expect that the arena kind of knows them, right? So they have given a, quite a lot of digital input, whether it could be an app or website, the emails that are received, the type of merchandise they buy. So if you have a holistic understanding of a customer, we can personalize experience in the stadium as well as outside of the stadium and if they are aligned. So they build loyalty to the overall brand. So then you have three different teams. So you know who to target. Who the sponsor should be. Am I targeting the right customer that they’re getting value from me? So understanding those problems beforehand and having a common solution across the board is taking them to the right direction. Where in sports and entertainment, so-called analytics or data analytics, it’s becoming very prominent.
Allison Hartsoe (10:27): I love this example because years ago, at the customer-centricity conference, we had a sports owner talk about how they were using customer-centric ideas. Now they didn’t really, I don’t know that they had the technology dialed in, in the same way, but they were looking at things like ticket sales before the event. And there was a philosophic dilemma of do we let people sell tickets that we should have sold to each other before the event. And ultimately, when they had the right customer-centric perspective, they realized that if they just made it easier for people to do that, you know, that these people had honestly bought the tickets. They had honestly tried to go, and for whatever reason, they couldn’t do it. It wasn’t like they were making a career out of scalping. So they enabled these customers to have a way that they could trade in their tickets, and other people could sell them so that they could get more volume out of people enjoying the event. In other words, coming into the stadium and having the experience rather than creating a hard stop, because it was all about the product. It became all about the customer. And in a way, this sensitivity is what you’re talking about with the technology. The technology can sense this in for them.
Tisson Mathew (11:41): That’s right. Yeah. And expose the hidden patterns that was not obviously visible before.
Allison Hartsoe (11:46): Tisson, do you think the word personalization is almost too weak to describe what’s really happening with the technology?
Tisson Mathew (11:53): Yeah, I think so. There is things that are common among all of us. And for that, you don’t need personalization. You need to a customer service understanding the common elements across all customers. Like for example, when you’re refunding customers, if the refund take 30 days, I don’t think any customer would be totally happy with it. So like, look at Amazon so quick that it is a fundamental thing that they’re here to do for everyone. So personalization matters when it comes to quite a lot of preferences, and those preferences can be explicitly given to the organization by the customers either by checking boxes, surveys, and others, but there are so many other implicit from purchase behavior interactions and things like that. So it’s a combination of those, and having that in the customer profile and personalizing their experience is extremely critical. But otherwise, I think sometimes just going after personalization as a buzzword probably puts you in the wrong track and then the right one.
Allison Hartsoe (12:45): Well, and the other thing that you’ve said here is that we shouldn’t be using maybe customer lifetime value as a way to create lower levels of standard service. That there’s a certain level of satisfaction, we need to be giving in across the board to operate as a company. And then maybe the personalization is less about you have a five minute wait time versus a one minute wait time and more about the nuances or the moments when someone needs to have care.
Tisson Mathew (13:18): That’s right. Yeah. To kind of tag onto your example, if it is a five minute wait time, just tell the customer, would you prefer a callback? Amazon is one company. What I learned a lot from that, some of these things extremely well. So there’s some of these things comes through Amazon, anecdotally on customer emails and whatnot. Others we actually learn from data.
Allison Hartsoe (13:36): Nice. And do you think that it was particularly because the data was dialed in and was comprehensive that you were able to get those learnings out of the data?
Tisson Mathew (13:44): Yeah. It has been a very long term goal of the company from its founding from a starting point. This is a company started up selling books where they figured out to store people clicking on which book and what time and who is clicking on this books. And what is the session time on each page, analyzing all that data, provide book recommendations, and then they found out, okay, so I can literally predict most likely what type of books these customers would like? So from their customer data and ingestion of customer data and having the mechanism to consolidate, it has been the foundation of the company from the get-go. So by every product or service that is built is built on that foundation. So there is no bigger assets for amazon.com than customer data. So there is no bigger asset. And then, the service and the reputation and the value was created was based on that foundation. So it’s a very, very long term thinking. I mean, the company grew only 25% year over year, but look at what it is right now. So you don’t have to be hyper growing. So you can have a sustainable longterm growth as a business, and it will be successful.
Allison Hartsoe (14:50): I love that thinking because it exactly lines up with customer lifetime value, right? It’s longterm growth. Now I’m going to come back to the foundational element and the cost-effectiveness of that. Let me just say that. I’ve noticed that many companies who are leaders in this space, Amazon, and others tend to build their own technology. And I wonder if it’s easier today to not cobbled together from different vendors, but to pull from a foundational, a horizontal foundation to get what you need to accomplish better technology faster. Is it possible? Is it easier now than it ever was before?
Tisson Mathew (15:30): I think so. I think because of cloud technologies; it is easier to build solution in house than before, for sure. What matters is the time to market and the reliability and access to talent. So there is only a finite pool of engineers. And if you divide that finite pool engineers to create custom solutions for every company that pool runs out, then what happens in a lot of organizations trying to go build their own higher talent and compromise on the hiring.
Tisson Mathew (15:57): And then there is attrition within the talent pool because there is not enough long-term growth within the company or for an engineers. So they leave the company. So there is a lot of transitions in and out of an organization when it comes to talent, and that creates instability and bad product. So essentially, the infrastructure is becoming easier, more accessible, and then Collin on how to actually create a solution around this. They’re coming more harder. Okay. Rather than hiring a 10 member team can be go into a cloud infrastructure by a cost-effective solution that is flexible enough for us to customize. And I will have a couple of engineers to do that. Probably a more viable scenario than trying to get 10 people. So hiring 10 engineers into a team might take a year to complete. Their come on board. They have to learn your business. They need to understand the problem they need to gel as a team execute. Yeah. So the rookie years of any engineering team is horrible. So, and if you are expecting results in the first year, so that’s when the hybrid model works out much better than going all in house.
Allison Hartsoe (16:58): And I can see where the hidden costs are here. And what we’ve been talking about is with cost-effective technology is I’ve got direct costs, and then I’ve got the hidden costs, and the hidden costs are oftentimes built around the people, whether it’s the ability to adopt what comes out or whether it’s the ability to build and modify what’s there. Yeah. Yeah. Excellent. Okay. Justin, do you have another example for us?
Tisson Mathew (17:22): So another organization that we are working with is in healthcare. So that organization is trying to make prescription drugs more affordable for businesses. So one of the challenges in that space is, as we all know, that costs has been a big problem overall, as part of the medical costs. So this company came to us in two ways. One is what they want to offer is a consumer solution, and employers are a channel for them. So what we liked most is the cost-effectiveness of prescription drugs is something everybody cares about. We’d love to work on it. So the challenge they had is going through employ errors to get their own employees data around their preferences, medication inherence, would they refill the drug. In accord a lot of it would, they actually go for the lowest price. Employers were very sensitive around the subject is like getting into the privacy aspect of employees information.
Tisson Mathew (18:14): So they wanted a platform and a solution, and then continuously learn from the interactions of the tumors. So he can understand the behavior. Will they be willing to actually get a mail delivery of a drug versus to go into the store? The reason why the mail-order delivery is more effective is because it actually comes to your door, and you haven’t, it’s more likely that you’re going to have medication inherence versus you’re going to a store, or you might miss going to a store. Or we have already said, you know, feel like it or whatever else the situation would be. So there is a quite a lot of factors that are customer behavior and medication inherence. So technology allows them to figure out what’s the best location and what the right delivery mechanism for the customer is. Initially, when you start off, you have very little information about the customer, and then as you start creating interactions with them, how do you build out a continuously progressive profile and also report on the cost of drugs where all the last three months worth this.
Tisson Mathew (19:09): And now it is trending towards that and the medication inherent assessed and the medication inherent now going this way and you’ll refill rates or this. And so all of this matters to an employer to make sure that employees are healthy and then report to them to benefits and HR, that this is the actual data that we are seeing in an aggregate. And this is what you can do entirely within an organization in terms of employee communication and others to benefit that. So that is a good example that we were able to work through from the beginning of this organization to where it’s actually going.
Allison Hartsoe (19:39): All based on pinning the data together to really understand. And I always think about it as a U shaped curve. You have people who maybe they fill the script, or maybe they don’t fill the script. And then they take like one, and they don’t adhere. And then you have people who are the other end of the curve, and they’re on it all the time. In the issue of health, perhaps there’s a point of that U shaped curve further to the right, that is indicative of health and creating health. And maybe even the point where they don’t need the product anymore. I think other industries might find a dangerous, anytime you have addiction issues, you’ve got that U shaped curve. And the more you drive them to the right, the more you’re like Oh, like the story we did last year was on big fish casino, which was a gambling problem. And people who couldn’t get off the gambling platform, but health is a little different. Can you be too healthy? Can you over personalize people’s health and make them too healthy?
Tisson Mathew (20:35): Yeah. That’s a good point. As long as they use that. And then thinking at that point, when you’re amazingly wholly, that point, if you can devote your time towards making other people how they’re and yourselves that’s when you probably get better value than, you know, crossing that point of, Oh, I am liking the, there was a 1% and now I want to be getting even better. Right. So that point you get, you should be a trainer, a mentor to others as well. Yeah.
Allison Hartsoe (21:02): Okay. Well, let’s say that I’m convinced, and I would really like to build cost-effective real-time technology. How should I think about this problem? How should I go about pursuing that?
Tisson Mathew (21:16): Yeah, great question. I think the starting point is at least writing down at least a high level, your top one or two pain points that you want to address, but the technology or a platform, and then doing some sort of a pilot or a proof of concept around it, but maybe one or two vendors possibly. And with your internal team, that is the right first step that I would recommend.
Allison Hartsoe (21:39): And here you’re talking about horizontal vendors. Yeah. So you’re talking about things that operate cup comprehensively across the foundation, maybe things that are built on the Microsoft stack or the Amazon stack, or I dunno, I shouldn’t put words in your mouth here.
Tisson Mathew (21:54): Yeah, that’s correct. There’s more of a horizontal thing. So you can go that for point solutions, right? So absolutely. So when it comes to customer data management or something more foundational, so rather than issuing RFPs and you know, so those are good. It’s extremely old school way of checking boxes. So we have no idea. Those checking boxes really works or not. So certain things buying certain things, RFP processes makes a lot of sense, but a lot of other things, it doesn’t, so this is one of them. So because there is only a select few choices you have and you have to really align those choices and test it. So the proof of concept really reveals are we even thinking the problem, right? What we did the problem statement is, right? Did the team come together to execute even at a proof of concept level? It’s like your training camp.
Tisson Mathew (22:38): So if you can’t really execute very well and show some outcomes and that, okay, this is going to work out. So that is the first step. So from a timeline standpoint, that is in 90 days, typical timeline. So once you’ve crossed that path, you can go into a case implementation. So every software platform you could get your hands on has a lot of features probably, you’ll end up using 20% of it. I mean, that’s the reality. So what are the 20% thing that you really need and can bother with features that are not really fully baked into our adopting as an organization? So that’s where the cost overrun comes into it. Like for example, if you throw an implementation effort on a feature that has very little business value and you spend all of your money trying to do email personalization, and, but you didn’t prove a concept, it pretty well.
Tisson Mathew (23:26): And your return is very, very minimal. Now another thing is customer journey mapping. Great. But if you have you tested it, your customer base is more broad market-based. And then individual journeys, really, it is not a driving factor. So you really do understand the signal from noise, and then we direct your resources and your presentation to right direction. So putting the effort into the foundation, elements of customer data management, make sure data is clean, and it is organized. It is just together. So we have the ability to connect customer data. Even if there is no common identifier between them like there is no foreign key or anything between the two data sets. We have the ability to connect them together, using machine learning and a few techniques. But we want to use those primarily to clean up and organize data first and then make sure that is there.
Tisson Mathew (24:14): And it is reliable before you start to hear anything else. So I think the cost-effectiveness a part comes into choosing the cost-effective technology that is flexible during the implementation redirecting to the right pain points and business value. Those are the two areas that we can come in. One is in the product cost alone and secondary implementation cost. But you probably have seen where the cost of a product is sometimes free like opensource, but when you want to implement it, it just examined the costs. So you ended up with the same place, right? Like, okay, I got the product for free, but I spent all this money to implement it, but hardly anybody use it.
Allison Hartsoe (24:47): Just some of this strikes me as you’re speaking from experience because you’ve been down the path many times before. For people who haven’t been down the path, how do they figure out these unknown unknowns? And let me give you an example where I have seen companies that have a ton of product and supply chain type data that has one mention of the customer in it. And they think they have a customer platform, or they can just mash the data together and get a customer platform. How do you get ahead of that?
Tisson Mathew (25:18): Yeah. So if you have no customer data at all, and we are very, you know, factions or elements of customer data and trying to figure out what your customer data should be easier to put effort into. I mean, so giving them the Amazon example, Amazon put effort into collecting customer data, and they put infrastructure to do that. So if you don’t do it and you don’t have enough, unfortunately, it has to start on a collection process or the index process first. So rather than relying on very fragmented unreliable data and making decisions, that’s when you probably humans are better. And so they have a far bar experience based on where the data is not sufficient enough to make those decisions. So then as you grow and you can remove the human 11 and rely more and more on the data.
Allison Hartsoe (26:02): Yeah. That makes sense. Awesome. Tisson, if people want to reach you, how can they get in touch?
Tisson Mathew (26:06): Yeah. Great. Couple of things. One is our website has a ton of information. We actually contributed to open source community a lot. It’s Skypointcloud.com, and that is our website. So that’s the best place to go, to learn more about our product. We have all the links of our social media and on our website Twitter and LinkedIn, both of them we use quite extensively.
Allison Hartsoe (26:26): And I will say that you’ve been very generous with explaining a lot about what goes into the tech stack. And I think the resources that you have on the site are definitely worth checking out.
Tisson Mathew (26:37): Thank you so much. Yeah, absolutely. We update it every single week with at least one blog article, something contributed to the open-source community and get hub on others.
Allison Hartsoe (26:45): Yeah. That’s all about giving back. Right? I heard you said that earlier. It makes sense.
Tisson Mathew (26:49): Yeah, absolutely.
Allison Hartsoe (26:51): So, we will link to that. And as always, everything that we discuss the links to the blog and the links to sky point cloud will be at ambition data.com/podcast. Tisson, thank you for joining us today. It’s been a very good discussion about cost-effectiveness and just the need to help businesses get ahead of that problem versus like what I call like the vendor Frankenstein nightmare. Everybody’s suffering through.
Tisson Mathew (27:15): Thank you, Allison. Appreciate it.
Allison Hartsoe (27:17): Remember, when you use your data effectively, you can build customer equity. It’s not magic. It’s just a very specific journey that you can follow to get results. See you next time on the customer equity accelerator.