Foundations For Success: Data Platforms Evolve With Chris Jones And Saket Saurabh

So much more than a database, the modern data platform delivers a range of functionality that enables better experiences for customers, partners and corporate employees. Whether for customer data, third party data or other use cases, modern data platforms enable organizations to be more aware, responsive and accountable. How can your company take advantage?

Check out this episode of DM Radio to find out! Host @eric_kavanagh will interview Chris Jones of Amperity and Saket Saurabh of Nexla.

—

Transcript

[00:00:39] Eric: We’re going to focus on a pretty important topic in the world of data. I can tell you that data platforms evolve. We’re going to talk about what data platforms provide these days. They’ve been around for a while, but they’re getting very sophisticated in this modern world. That’s very good news for anyone in the business. We’re going to be talking to a couple of experts, Chris Jones, from a company called Amperity, and our friend, Saket Saurabh from Nexla.

First, I’ll give some background on what we mean by data platform. A lot of people know what a database is. There are so many databases out there these days. A data platform is much more than a database. It does persist data, of course, but it also persists data in a way that makes it much easier to leverage certain kinds of functionality.

A lot of times, it’s for customer experience and built around understanding who a customer is and being able to cater to them. Tracking what you’ve purchased, tracking your preferences, things of this nature, to optimize marketing as well. There are lots of other reasons you could build a data platform. One of our guests from Nexla started off as a clearing house for third-party data, but then they moved into some other areas. That often happens in the software world.

You have your kernel. You build out some solution, and you build on top of that in the following years and do interesting things. We’re going to talk about what data platforms are. Why they’re important? Why you should use them? What to look for if you are shopping out there for a data platform? First, maybe let’s get an opening statement from each of our guests. I’ve got Saket Saurabh and Chris Jones. Saket, tell us a bit about yourself and what Nexla is doing in the data platform space.

[00:02:16] Saket: Thank you, Eric. I’m Saket, Cofounder and CEO at Nexla. You’re right about where the world of data platforms is going. What Nexla does in this space is we are all about making data ready to use for the users of data. There are many types of users of data. It could be someone in marketing who’s trying to analyze the ad spending. It could be somebody in sales who is analyzing how effective the campaigns are. It could also be data scientists who are building some machine learning models, or someone in finance or accounting who’s looking at numbers, converting currencies, and all sorts of operations that people do with data.

Our function that makes our mission is to make data ready to use for each of these types of users, which means making sure that the right data is in the system they use. It could be a spreadsheet, a database, or an API. Wherever people need data, it is structured and validated. Errors have been managed and so on. The way we do that and make it possible is through automating data engineering. Data engineering is a function where people write code and do complex operations so that data can become useful for people. In the line of sophistication, we are all about making even that piece automated, intelligent, smarter, and much more current.

[00:03:32] Eric: We’re going to get an opening statement too from Chris Jones from Amperity. Chris, tell us a bit about yourself and what you folks are doing at Amperity to leverage data platforms.

[00:03:42] Chris: My name is Chris Jones. I’m the Chief Product Officer at Amperity. The company was founded a few years ago. Prior to that, I was 27 years at Microsoft, leading a bunch of different product and engineering roles. The fundamental problem that Amperity was founded to solve is helping large-scale consumer brands, access and build a unified view of the customer. When we say unified view of the customer, it is a view of the customer that can be used and serve all aspects of the business, marketing, analytics, and information technology. Fundamentally, the way to think about the uniqueness of our platform is that we applied cloud-scale technologies, including machine learning technologies, to the problem of organizing and managing first-party data and being able to take a whole set of messy data around the customer and doing two things.

Number one, find the linking key or the matching key across messy customer data. Number two, build all the normalized tables that you might need around the customer, what they’ve bought, what products they’ve used, and where they visited. Make all that accessible to the different roles in the business and make sure that data doesn’t just sit in the data platform, but gets published into the downstream systems. That’s the customer call center, the marketing automation systems, the business intelligence dashboarding systems. Think about Amperity’s role as the beating heart and engine that pumps high-quality customer data to all parts of a consumer brand business so they can serve their customers better.

[00:05:13] Eric: It’s funny. You mentioned a couple of things that struck my interest. You’re talking about getting that data back into the apps. There’s this whole thing now that people are talking about, reverse ETL. It seems to me that you are already doing that and that’s a term to describe that part of what you do. Is that about right?

[00:05:29] Chris: Yeah, it’s about right. Let’s use my analogy of there’s an engine and a heart that’s pumping this data into different kinds of places. You need to make sure it gets mapped into the schema of the downstream system. In other words, that ticketing system has a schema and it’s got to get mapped into it. It’s not being able to do reverse ETL, but it’s also being able to do it in a way that has semantic knowledge and understanding of what the attributes are, clear role-based permissions for what things can go into that system, and regularly schedule and monitor workflows. Every single day new data comes in every single minute. You’ve got to have all of those things working together to make this function.

[00:06:15] Eric: It’s important because we spend all this time and effort capturing data and trying to persist it somewhere. You mentioned the messiness of it. That’s so true for so many different reasons. You’re looking for the signal between the strings. It’s very interesting stuff because what you are trying to do is you’re trying to optimize marketing, sales, and the experience that the customer has. You’re using the data you have about them to do that.

Data Platforms: Data engineering is a function where people write code and complex operations so that data can become useful for people. It’s to validate and structure data so that there are no errors.

[00:06:45] Chris: It’s a very simple thing to ask. Walk into any consumer brand and ask the following question. How many people do you have that have spent more than $1,500 lifetime to date and aren’t in the loyalty program? It turns out that’s a hard question to ask most consumer brands. Typically, the access pattern of that question might take days or hours or weeks. That answer needs to be known instantly. It needs to be known when your person lies on the webpage. It needs to be known right in the customer call ticketing system. It’s not enough to get the data right and accurate. That’s hard. It also has to get to the right place. It can’t just sit in the data warehouse. It has to get into the employee’s hands that are serving the customer.

[00:07:27] Eric: That’s important. Maybe I’ll bring in Saket again to comment on this. You’re in a slightly different take. As I recall, you were a clearing house at first. One thing I liked about your approach was that someone would take clickstream data, for example, and publish it in Nexla. Other folks can come along and use that data for their own purposes. Obviously, there are some terms and conditions, but you would also publish in a variety of formats like parquet format or other ETL or data integration ready formats, such that someone can come in and whatever solution they’re using, they can patch it right into that particular target. Is that about right?

[00:08:06] Saket: It’s in the right context in the sense that we started with the problem that when companies work with other companies together, there is the data flow between them. That is usually one of the most complex because there are not a lot of standards around it. For example, our customers like DoorDash and Instacart get a huge amount of data from grocery stores. Which store has what products? How many units are there? What aisle is it in? For these companies, getting this data from CVS or Albertsons or different stores, different formats or different APIs can be a challenge. The automation of data engineering is about making that part simpler, automated and easy. Therefore, we were enabling that data that was going across companies and bringing them that.

We now have expanded a lot more because the concept of that data ready to use in your hands and it’s coming from all places and so on is common. Johnson & Johnson is using us when it comes to data that helps them in their pharmaceutical research area. We have LinkedIn which uses us for their marketing team. Companies like Varsity in the education space. One of the largest banks uses us for the AI team. The applications are across the board but ultimately, all of these people need the data in their hands and the tool where they can use it.

To Chris’s point, that was basically what you’re saying, it’s that having the data is not sufficient. Having it be applied to the right place is necessary. We are at the place where we get the data ready to use and users on platforms like what Chris has will connect it to something specific and get an outcome like, “My customers who are not in the loyalty program and let’s send them a mail or something to join the loyalty program.” That action can then be taken on that data.

[00:09:52] Eric: I’ll throw it back over to Chris. The blessing and curse of information systems these days is there are so many ways to get any particular thing done. You go from one company to another and there’s very little similarity in terms of an end-to-end solution that they put together. You get these legacy environments with all these different applications, interfaces and price points. You get all these new solutions coming down the pike.

It can be downright bewildering for the professional in our industry to figure out which tools to cobble together. The key in both of these cases, because you provide data platforms, is that you’re trying to deliver the data provisioning, and I’m guessing to a certain extent, enriching or cleansing services that would be expected for such a solution. That can feed into any of these target systems that someone is using. Is that about right, Chris?

[00:10:49] Chris: That’s about right. One of the observations that we made, and it’s what our two companies come at and have taken similar approaches to probably different domains. The observation we made is that 80% to 90% of what every consumer brand wants to know about the customer is the same. They all have to understand who they are. They need the linking key across personally identifiable information that they collect with consent. They need to know what products they buy, what channels they visit, and where they like to respond.

Most of the time, if you go talk to someone in analytics or data science or IT, they’re spending all their time doing the 80% of the stuff that’s common and not their time doing the 10% to 20% that’s unique. What we’ve done is we’ve automated that 80%. The example I always use is that 90% of our DNA is the same. It’s true, but it’s the small parts that make us different. When we think about a data platform, we think about two things.

Number one, it standardizes a set of things that you want to do. It standardizes and creates out of messy data these normalized tables that are 80% to 90% of what you want. Two, it empowers a trained user to program the platform so that it benefits the employees of the brand. That means that we don’t start with some fixed notion of the schema around customers because for Brooks Running, it’s very important to know what kind of shoes people use. At T-Mobile, it’s more important to know where their subscription is. Did they start as a pay-as-you-go user and then want to go and turn into more of a subscriber? What kinds of offers are relevant?

This combination of standardization of the 80% that’s common with programmability for the 20% that’s different, and flexible conductivity so you can drop it into any existing systems and get the results you want. Another observation we’ve made is typically what happens when we walk into a brand, you’re running the right campaign with the wrong data.

Most of the time, what people do is spend time replacing tools as opposed to replacing the data. If you get the data right, all your campaigns get better. All your ad campaigns are more reliable and the return on investment goes up. Your analytics get cleaner. Your people get more productive, and then you’re able to program the system to start to optimize as opposed to spending all your time just doing cleansing.

Data Platforms: Ask any consumer brand, “How many people do you have that have spent more than $1,500 lifetime and aren’t in the loyalty program? That’s a hard question that needs to be known instantly.

[00:13:24] Eric: If you get the data right, all your campaigns will be better. That is so true. It does take a bit of time and effort to get there. The key is to have a technology that allows you to leverage the power of your crowd and to put the power in the hands of the people who were either running the campaigns or designing the campaigns or in charge of customer service. The closer you get to the data with a meaningful ability to modify or change it, the better off you’re going to be.

[00:13:55] Chris: When we look at the problem, you look at an employee of a brand, someone who’s in marketing or customer support. They don’t want to send unwanted emails. No marketer gets up and says, “I’m going to have a great day today. I’m going to send a whole bunch of unwanted emails to people.” Nobody in customer support gets up and says, “I’m going to have a fantastic day today. I’m going to answer the phone and not know who the customer is, even though they’ve given me all the information I should know.”

We think about our tool as empowering and unlocking with better data. The natural tendency for employees at all these brands is to have the information to do their best work, to have the campaign be the right campaign, to have the support experience be the right experience, to have the analytics view that you give to the CFO be accurate, complete and up-to-date with the insights they expect.

[00:14:46] Eric: That’s a good point. I’ll bring Saket back to comment on that. There is this value to having some standardization and golden record, as we’ve called it for years, around customers. What we’re trying to vie for is to solve that 80%, which is what Chris is talking about, and then let the rest of the 20% be the gravy or the secret sauce or whatever it is that goes on top of that.

[00:15:12] Saket: Having a good understanding of who that customer is, is essential to the service we deliver. These days everybody expects that. Behind that is a ton of data because there are more data systems that the customer information is sitting in more and more places. Many of them are in your system, but many of them are also outside the SaaS services that you may have, so various places with data sets and bring that in together. To tie back into the general conversation with the back form and we’re asking about that reverse ETL and different things. What is happening right now is that ultimately, the way the landscape has been until now is that if I need the data from some files in a database, I need one tool, which is an ETL. If it’s a SaaS service that I want the data in a cloud warehouse I need an ELT tool.

If it goes back from a cloud warehouse to a SaaS service, I need a reverse ETL tool. I need streaming tools, iPaaS tools, and data API. There’s a fragmentation of different types of tools effectively moving data from A to B. It’s just different tiers of systems. The modernization of many of these things is also heading towards that.

That customer data could be hundreds of places. How do we not spend most of our time figuring out the in and out of it, and get most of our energy on what are we going to do with that data? Is it ready to use? What can I do with it? What campaign or customer support and so on? A big need for many companies with this explosion of data and data systems is to start to bring this together and simplify.

[00:17:34] Eric: Chris, I’ll throw this one over to you. We were dancing around it in the first segment here, but there is this huge disruption that’s already happening in the world of advertising with cookies, in particular, third-party cookies. It puts tremendous importance on what’s called first-party data. First-party data for those who don’t know means data that you have in your company about your customers. Whether it’s gathered through transactions over the years, surveys, phone calls, customer service, reps talking to people, whatever it is. All the data that you have about your customers is your first-party data.

Historically, in the last several years, companies have relied heavily on third-party cookies to know where people are going, what they’re looking at, sharing this data across different platforms, and that is apparently all going to go away. Google has said that third-party cookies won’t be supported. Apple is a big player in that scenario. It’s a very strange dynamic, but I’ll bring Chris back in to comment on it. It’s disruptive and we need to figure out what to do. I think customer data platforms are a big part of the solution. What do you think?

[00:18:35] Chris: I think it’s a big part. What I would say is our observation when we’ve talked to our customers is there are three things that are going on. One is they’re increasingly moving to directly transact with their consumer. That’s a big shift for every industry, whether it’s insurance companies or banks or retailers. That’s step one. Step two is because everybody is shifting there, the cost of acquiring a customer on an ad platform is going up. Acquisition costs are going up.

Step three is the way you used to do targeting and acquisition, which is based on third-party cookies or mobile ad IDs, is going away. What’s happening is what used to be viewed as a funnel in your customer life cycle, where you would use some data to acquire customers, put them into the funnel, and have some other systems manage them is now turning into a wheel.

You acquire customers, collect first-party data, understand what’s happening to them, find your high-value customers, and then that key becomes the seed into ad networks to find lookalikes who are likely to be high value because you want to find people who are going to be appropriate to go and acquire. This is completely changing how the brand thinks about this. Your first-party data is now important in every part of the life cycle, how you acquire a customer, how you turn them into a repeat purchaser, how they become loyal, how they become high value, and how those high-value people tell you how to go drive in the ad network to promote and find other people who are likely to be high value because you can’t afford to acquire a one and done person. You’ll lose margin on the transaction.

Data Platforms: Most people spend time replacing tools as opposed to replacing the data. If you get the data right, everything gets better. Your ad campaigns get more reliable, your ROI goes up, and your analytics get cleaner.

What this is causing is all the brands to ask the question now across all touchpoints, I have to have a single view of the customer based on first-party data. I have to change my website so that I’m collecting with consent on that first-party data. I have to use that data across my brand to continue to drive people through the wheel and the faster I can spin the wheel, the better my business goes.

[00:20:40] Eric: That’s a good point and that explains why every freaking site you go to these days has a pop-up within ten seconds of you going there. On every single site, the pop-ups went away and they all came roaring back.

[00:20:54] Chris: The thing that will start to happen is that this makes a brand need to have a different kind of promise to you as a user. In the past, they could just go collect it with a third-party cookie and you didn’t consent at all. Now that you have to have consent, the conversation matters. If your experience of that conversation is a disconnected pop-up, you’re unlikely to be willing to give the brand your information versus some other kinds of things.

What you’ll start to see is brands putting an economic value on acquiring that data and personalizing that for such experience in the same way that they would send you a thank you for shopping message after you’ve transacted. That experience is going to become more curated because it matters to the bottom line.

[00:21:41] Eric: There’s a related topic here. I’ll throw this question over to you and we’ll see what Saket says about it. There is all this concern about privacy, which I do understand. It’s important to recognize, not just from a security perspective and governance perspective. With GDPR, you’ve got the right to be forgotten, for example. I’ve built upon that with my own theme that I think you guys would probably enjoy. I call it the right to be respected. That is the principle by which organizations should operate.

If you do that, you’re already going to be taking the boxes necessary to keep an auditor happy. You’re already going to be doing that as part of your corporate DNA. The other thing I think that could be very interesting here. I’m seeing a little tiny bit of this, like from energy companies and from some of the credit card companies, etc., is sharing data with the end user about what they know about you and saying, “We’ve noticed X, Y and Z purchased these products in the past. You’re not purchasing now. We’re curious to know why.”

It’s those outreach efforts geared around sharing useful information that they have about you. To me, that’s a huge opportunity to show, “We’re paying attention to you. We’ve known what you’ve done. We appreciate you as a customer. Here’s some insight we have based upon your buying patterns in the past.” I’m seeing a little tiny bit of that. I think we’re going to see a lot more of that to open the conversation between consumers and the companies that they work with. What do you think?

[00:23:05] Chris: I completely agree. I think what we’re seeing is an evolution of what I’ll call the human connection between an individual and the employee of the brand that’s serving them. What’s changing is that social contract if you think about it. I like to say what was creepy five years ago on Amazon is now rude if you don’t do it. Remember the first time you logged in and you looked in like, “They know everything that I’ve purchased. That’s so weird.” Now, if I shop with a brand and they don’t know everything I’ve purchased, that’s so weird in a bad way.

It’s important to recognize that what’s changing is the expectations that consumers have about how they’re treated by the brand. Some of that is privacy, but some of it is tone and respect. It will vary by region. It varies by the type of customer you have. Brands are going to have to become not just data intelligent but socially aware.

I have the data. What is respectful to use about the data? What should I be sharing with the customer? What did they expect me to know at that moment? That again, if we go back to the human connection, it involves not just giving teams the right data, but also training them on how to use it in a way that meets the customer where they are.

[00:24:27] Eric: Also, speaking out loud. There’s a guy I had on the show a few months ago who was very impressive individually. He does coaching for executives and things of this nature. He has a lot of interesting thoughts that he shares. He goes, “What happens at the door of indecision? Nothing.” If you’re stuck in indecision, that means nothing is happening. Write and talk about it. Put it down on paper. Our ideas tend to crystallize as we put them either in words through our mouth or fingertips and crystallize it. Talk to your employees about these things, and get their feedback on stuff. You’re going to find or figure it out because the strange thing about these cultural mores is they do change. You’re aware of it.

If you start talking about things, everyone will hear that and go, “That’s right. That’s funny. I did think that was weird. They knew it five years ago and they don’t know it now.” That’s because things have changed and information is a big part of that life cycle. We’re advancing as a culture very rapidly with information. Look at young kids and some of the information that they know because they’re sitting there on YouTube and TikTok. There are little bits of data coming their way like brands. There was some story that went around about how these young kids could identify all these brands. How the heck can you identify those brands? They’re watching TV and videos and stuff, and it’s persisting in their brains. It’s funny. What do you think, Chris?

Data Platforms: Your first-party data is now important in every part of the customer life cycle. From how you acquire a customer to how to find other customers who are likely to be of high value.

[00:25:51] Chris: First of all, I will acknowledge that now I’m old enough to be out of touch. I’m sure I don’t everything. Your point is exactly right. The speed at which the consumer is changing is only growing. What it means is that brands have to be able to respond. Part of the value of having a fantastic data platform is not just getting it in, but having it be agile and iterative enough to respond.

As an example, TikTok wasn’t a thing, but now you have to have a presence on TikTok. Does your data platform let you do that? Does your data platform let you change the way you’re talking? Does your data platform let you consume the news signals around that? Is your data platform flexible and allowing your employees and your brand to change? Are you getting the signal quickly enough to learn?

The speed of learning is so important now. Anyone who thinks they’ve figured it out is already wrong. The world is going to change too quickly for us. Fantastic brands and companies are embracing that notion of agility. I want to go back to that means it’s more important than ever to give tools. This is the other thing that I think platforms do. They’re employee empowerment tools. They make the analysts’ jobs easier. They simplify things for the data scientists. They make the marketer’s job easier. They unlock more time for employee creativity and less time for data wrangling. That is a very clear benefit that you should get from your data platform.

[00:27:27] Eric: That’s a great quote. Anyone who thinks they’ve figured it out already is wrong. That’s funny. I’ll throw it over to Saket to comment on. There is something to be said about having enough flexibility in your system. I’ve been working with an interesting company. They have some offshore resources to do things. I’m finding out through asking questions, which parts of their platform have the agility and which absolutely do not.

There are some things you wouldn’t expect like, “What? I can’t change that at all now that I’ve started working with you. Maybe someone should have told me that when we onboard this whole program.” There are certain things in platforms that have to remain stable because they’re holding up the whole solution, but you do have to be flexible. I think Nexla has done a pretty good job of evolving as you’ve watched the use cases come along and watched people use the technology you’re evolving. What do you think about that, Saket?

[00:28:21] Saket: I think what Chris was saying also connects to the point that you have tools and tools that are meant to solve a particular problem. You may have a tool to do marketing on a certain channel like YouTube, but a platform is more of a foundation that you can, in a flexible way, use and build on top of that. That’s what we mean when you say it’s a platform.

To be able to build on top of that is that use case of marketing. You can also handle that because you’re not framed into a fixed box. There’s more flexibility around that. What we have seen in this space is that you have to straddle both those things together. While for example, we got a no-code tool for many people that they can do their things, it’s also a platform that people can build on top of.

In the technology world, when we think about platforms, we often understand that there are APIs and interfaces to them. There’s a software development kit that they can build on top of. There are more flexible ways of using the product alongside new things that I may buy or other services that I may use. To your audience, when you think about the platform part of it, a lot of it becomes not just what I can do with it today, but does it have the bones to carry me through the new things that come along with it?

[00:29:34] Eric: I like that reference. Does it have the bones? That’s the hard part of your body and it’s going to keep you going. You have to be flexible. That’s all the joints.

[00:29:42] Don’t start to break apart because you put a new workload here.

[00:29:48] Eric: We do need to have scale these days too. That’s the other key critical success factor for organizations. It’s to have some solution that can scale up, that can handle increased traffic, that is significant and that can also scale down when traffic is very low. We’re getting very good about being able to not just scale up, but scale back down to manage costs and give visibility into costs. That’s the other side of the equation as you have to be able to afford these solutions and know what that total cost of ownership is.

[00:30:22] Saket: There are two parts to scale. There’s a scale of data and people. What organizations are also struggling with is how I scale. I have people who work with data or do these things. There’s a scale of people and of data. The scale of data is not just about scaling up but also scaling down. The traditional approach was that I get peak traffic to my website on Sunday evenings and I get ten million visitors an hour. All my capacity is built to handle that. On a Tuesday evening, that capacity is going based. Now you are in a cloud environment. You’re not committed to that.

Data Platforms: Data platforms are employee empowerment tools. They make the analyst’s job easier. They simplify things for the data scientists. They unlock more time for employee creativity and less time for data wrangling.

Interestingly, what happens is that simply moving to the cloud doesn’t get you the scale down. Your application has to be smart enough to now say, “I don’t need all these servers running at this point. I can do with less capacity and be able to get rid of that.” For most companies, if they don’t know how to scale down and scale up dynamically, the cloud would end up being more expensive. They’re suddenly running more servers all the time, but in a more expensive environment, which is a cloud.

To be honest, even within the cloud, when we talk about the total cost of ownership, cloud is very interesting. You can buy a server demand. You can reserve a server for much lower costs ahead of time. There is a marketplace where you can on demand get a server at a much lower cost than anything. If your application doesn’t know how to use all these three types of infrastructure in the right way in the right combination, you’re not making the most of it. The total cost of ownership is a big part of that structure and everything.

[00:32:30] Eric: Chris, I’ll throw it over to you to take it where you’re going to go. Go ahead.

[00:32:34] Chris: We were having a nice conversation about this transformation that’s happening in consumer companies where they’re shifting from a world where they might have been transactional in nature. They are focused on selling products to new customers and moving to a world where they are retention-oriented in nature. They care about their high-value customers and they’re changing their business.

One of the things we talk about at least in the IT infrastructure, there’s the grand re-keying that’s happening. Every brand is re-keying around the customer and trying to pivot all of their views around the customer. In old-scale consumer brands, the way to get to scale was to sell through a distribution channel. It was expensive to own the last mile. You ended up being keyed in your IT systems on product or region or store or things like that.

We all walk around with our distribution channel in our pocket right now. We’re directly connected to these brands. All the brand needs to be able to re-key around the customer. One way to think about the problem that we solve for brands is we’re a re-keying solution. We’re able to connect to all their different systems and give them a unique new customer key that connects all their customer data together. That allows them to have a foundation or a data spine that lets them think about and reason about their customer as the first class thing.

[00:33:58] Eric: I liked this concept. I think you’re right that all these companies are trying to re-key as you describe it around the customer. It’s a significant shift, but it’s a very important one because once you understand your customer types then you can go out to the ad networks and search for that persona that you have honed in your data science program, for example. That’s very interesting. Saket, I’ll throw it over to you to comment on what you think about all that.

[00:34:24]: Saket: Customer data and being able to understand who the customer is and serve them well. Part of it is customers expect that nowadays because they see the quality of the results that they get. All that is super core. From a data platform perspective, we also come across this aspect where there is a whole variety of data, which is technically not customer data.

l spoke about inventory data, medical devices, education-related, markets, financial and stuff like that. From a data platform perspective, companies are thinking about how I scale forward. How do I scale with the data? How do I scale my people in my org? The data platform side of things is also moving towards a point where instead of many point tools, I need a tool to do this and that.

It’s moving more towards, “I need a platform because I have to get data from a customer data platform to my marketing system. I have to get data into my customer data platform because I have data in all these places. I need to then find the data.” It’s not just getting the data in the right place, but also being able to discover it. It’s the ability for people to collaborate. That’s where the people scaling comes from.

There are hundreds of people in our company. Everybody is using data and somewhere they’re going to collaborate together. These people know customer data well. These people know the sales data well. They need to be able to work together and feed off of the data that they each own. This concept that is coming nowadays is about data producers, data consumers, data as a product, and as an entity itself that people can work with. There’s a lot of work going on to say, “Do we build a data mesh and so on with that?”

I like that things are heading towards product-based broader solutions, capable of handling many more scenarios, but all these things are also moving towards a collaboration nature. Without that, the companies are finding that they can scale working together as teams. I’m pretty optimistic about where things are going. We are talking a lot about automating data engineering and a lot of work that goes into making these things possible. That’s where the evolution is. It’s an exciting time for the data platform.

Data Platforms: A data platform is a connective tissue that makes everything work together. It provides the guard rails for better collaboration between silos. It removes inefficiency which will drive less dysfunction.

[00:36:41] Eric: That’s a good point. Maybe I’ll throw this last question over to both of you to wrap up here. The other big thing that’s happening here is that organizational hierarchies are changing and remote work has something to do with that because we all figured out we can do a whole bunch of stuff remotely. The office is always going to play a role and we’re going back, but many roles are changing in part because you don’t bifurcate your organization the way you used to.

There used to be this business-IT divide. The cloud has softened that a little bit as the dev ops and some other things, but the point is roles are changing and you can’t so easily compartmentalize what people do anymore, nor should you, to the point that Saket was making there. We need to collaborate more and collaborate around the data, not around the technology, around the data products that help us understand our particular role and how it fits into the big picture. What do you think, Chris?

[00:37:37] Chris: First of all, I completely agree. The thing I would build, and it would be great to get what your perspective is, when we think about the customer data platforms, the customer data platform in your business has every system you have that has data about the customer and touches the customer. It’s not one thing. I think people quickly say, “I’m going to have this one thing and it’s going to do everything.” There’s a reason there’s a ticketing system. It’s designed to help the customer call center representative understand the ticket. There was a reason there’s a marketing automation system. There’s a reason that there are BI tools that understand how to do visualizations and graphs. The key is to go and embrace the fact that what we want to do is to facilitate easy access to common data in those systems and to streamline what is a very expensive process. Let’s say I’m a business user and I want an answer to a question. Can the analysts go and give me that answer by working in basically the same data platform? The result of their analytics is available to me to segment the business in that instant.

There are too many silos now in how that happens, which causes incredible inefficiency in these companies. By the way, the inefficiency drives more dysfunction because then they’re convinced they can’t rely on each other. They silo even more, which puts more barriers to collaboration, and creates more expense for the brand. Whichever way, you’re looking at this, we look at it through the lens of customer data and data about customer like products and things like that. We believe the data platform is the connective tissue that makes all of this stuff work together and provides the rails for collaboration between these groups.

[00:39:18] Eric: Those are good analogies. You’re hitting on one of my main themes right there, which is morale, friction points and frustration. Friction points are when things aren’t working exactly as they should and if that doesn’t get fixed, thus the frustration and the morale is going down. That hurts productivity. It hurts collaborative efforts. It hurts everything. As a company manager or a senior executive, you have to look for the ties that bind. These data platforms do play a very significant role and can help facilitate. If you got 5, 7 or 9 different systems where different pieces are, that’s all part of the big system that is your information landscape or your topography. You need to figure out how to either orchestrate it or consolidate it or do something to where you’re not having all of these friction points getting people upset about stuff. Right, Saket?

[00:40:17] Saket: Yeah, absolutely. The efficiency of companies is not spending too much time chasing tools, learning them and training people, and finding that two tools don’t talk to each other. It’s about having a unified system. To your point about roles, it’s very much like that. If you could be someone in human resources, you now have to be data-driven. Who is applying for a job? Where are they coming from? How many people go through which stages of an application? Is there compensation? How do we determine that?

Everything is a data-driven exercise. That data feeds into your finance and budgeting. When people start to do their work, the tools that they’re using, whether they’re writing code in GitHub. How many bugs they’re creating? How many deals they are creating? All the data is across so many places and during the whole life cycle if you think of marketing, sales or HR, everybody has to be data-driven in a company. That’s how they are operating.

It becomes important to stitch all of this data together so that they can get a good view as Chris is saying about the customer. Every function is like that. What is happening is that there is a specialization of tools. You have the data warehouse, especially the cloud data warehouse is a big normalizing entity right now. A lot of data can come in there, and so are a bunch of other data systems. Putting those together so that the friction, we don’t want to spend time doing tools, clicking buttons there. We want to spend time getting the stuff done.

One of the things I always remind people about is data is not end. Data is the means to the business end. The less friction there, the more automation, and the more intelligence, the better off you are in focusing your team’s energy there. I see a lot of convergence coming. Data has been there for years. Tools are now starting to come together into platforms.

[00:41:59] Eric: It was a fantastic show. We’ve been talking to Chris Jones from Amperity, and Saket Saurabh from Nexla. Look these guys up online.

Important Links

• Amperity
• Nexla
• DoorDash
• Instacart