🦄 Making great presentations more accessible.
This project aims to enhances multilingual accessibility and discoverability while maintaining the integrity of original content. Detailed transcriptions and keyframes preserve the nuances and technical insights that make each session compelling.
Overview
📖 AWS re:Invent 2025 - EverCommerce’s Secret to AI-Ready, Multi-Brand Data (ANT101)
In this video, John Sedleniek from Fivetran and Kalpit Patel, VP of Enterprise Data and Analytics at EverCommerce, discuss EverCommerce's data modernization journey. EverCommerce, serving 700,000 customers across 40+ countries, faced challenges with siloed data from 50+ acquired companies. Patel explains how they built a "data factory" using Fivetran's Managed Data Lake on AWS S3 with Apache Iceberg, transitioning from decentralized systems to an open data platform. The implementation enables cross-brand visibility, improved targeting, and AI/ML readiness for customer churn analysis and propensity modeling. They leverage DBT for shift-left data quality checks and plan to integrate Census for reverse ETL, creating a comprehensive 360-degree data journey that empowers business users with self-service capabilities.
; This article is entirely auto-generated while preserving the original presentation content as much as possible. Please note that there may be typos or inaccuracies.
Main Part
Introduction to Fivetran and EverCommerce's Data Modernization Challenge
Welcome everybody. Thank you all for taking the time to be with us for this session. We're going to go into some detail regarding EverCommerce's journey from a data modernization effort. Let me start by introducing myself. I'm John Sedleniek with Fivetran. I'm the Vice President of Enterprise Sales. With me is Kalpit Patel, Vice President of Enterprise Data and Analytics with EverCommerce. We're excited to be here and have a really great lineup of questions to explore in detail.
For those of you not familiar with Fivetran as a company, we are a data ingestion and data movement company focused on automated, secure, and scalable data movement. We have about 1,500 employees around the world. We focus specifically on operational data stores and SaaS applications, and we also work around data tooling. The portfolio that we use is quite broad, as you can see.
Our portfolio is very broad. We have about 700 pre-built connectors that our customers leverage. Think about things like Salesforce and Zendesk, Workday, and Google Ads from key SaaS applications all the way through to large operational data stores like Oracle, SQL Server, Sybase, and DB2, and the list goes on. We have about 7,000 customers that trust Fivetran's technology and Fivetran's teams throughout the world, and we have a global footprint in North America, Europe, and Asia Pacific. You can see a listing of some of our customers in addition to EverCommerce that have entrusted their business and their data movement to Fivetran.
Let me talk quickly about how we do this, and then we'll jump into the conversation with Kalpit. As it relates to AWS, I'll try to be as specific as I can. When you look at the left-hand side of the slide, you'll see the portfolio of source systems that we move. You can see everything from operational data stores and SaaS applications to digital tooling like GitHub and so forth. The magic happens on the inside. Fivetran has automated, scalable, and secure data pipelines that we deploy for our customers to accelerate time to value and speed to value, reducing overhead and reducing cost. These are maintained by Fivetran, so these are automated pipelines with no additional management or costs related to that. There's no schema drift. All of these things are automatically accommodated and engineered as part of our platform.
We ingest and move that data into storage systems like S3, and we'll talk a little bit more about that. The outputs on the far right-hand side support everything from real-time analytics and business operations to cloud migrations and what everybody is excited about around AI and ML as well. With that, I think what we'll do is sit down and have a conversation. Is that good for you? We'll talk a little bit with Kalpit. Why don't we start with you just introducing yourself and your role at EverCommerce?
Thank you so much, John. Good morning and good afternoon, everybody. My name is Kalpit Patel. I'm the VP of Enterprise Data at a company called EverCommerce. I'm not sure if you're familiar with EverCommerce, but we are an interesting company. We are a leading service commerce platform that is digital native. We provide solutions to small business owners. We operate in about 40 plus countries with about 700,000 customers. We are very unique in providing services to our customers, and our purpose is simple: to empower and digitally transform service owners of small companies so they can be savvy. They should not be worried about what they have to deal with from a technology landscape. They just have to worry about running their businesses, and that's what we really strive for.
Everybody here knows about the uniqueness of 50+ organizations, and all of them being independently operated. What was the data infrastructure like as you approached this modernization effort? What was the original state? It was like every other company when you go through buying a portfolio of companies. You really have to deal with three different nuances: people, process, and technology. But when you look at the data infrastructure at scale, you think about silos of data, duplications of data, and really making sure the harmonization is done properly.
When I joined the organization about 18 months ago, it really struck me when I did about 60 interviews in my first three months. I went on a listening tour and discovered that we are a very data-hungry company. In order to turn that vision from a data-hungry to a data enterprise and a data-led company, you really need to provide accessibility to data with the speed of light. That's how we started our journey on what I call a data factory.
So you had these many organizations that were data-hungry, but there was really no organization or plan around how to consolidate those from a silo perspective into something more centralized. Is that fair? Yes, so within our commerce, we don't say centralization. We really say we are an Enablement Center because our goal is to enable the businesses and take them to the next level, whether in terms of technology. We are not just building a data factory; we are enabling the future of what our customers want and need from us with the speed of light.
Building the Data Factory: Strategic Implementation with Managed Data Lake on AWS S3
As the business scaled, what were some of the biggest challenges as you got involved and did a lot of interviews and understood the current state? What were some of the biggest challenges you encountered with that decentralized data? The three things that always come to data infrastructure and building an intelligence layer are: first, you have to understand the business. Everybody wants to do digital transformation, but if you don't understand the business landscape and what the company is made up of, it's going to be a very challenging task.
For us, it's not about the issues at hand; it's about the opportunities that bring us, and the opportunities are endless. Having data in multiple layouts and multiple formats because we bought a portfolio of companies, it really is about having them all de-siloed. That's number one. Second is how do you provide data accessibility. The third is how do you really put governance on top of all of this in the world of AI. Everybody lacks two major scenarios: how do you bring data together and how do you create a data foundation with quality and a governed infrastructure.
Once you have that, we call it a data factory. That's our foundation. Our goal is to build a factory. That's phase one. Second is lay the machinery in the factory. Third is build a factory and make it ready for rollout. Those are the three phases. With Fivetran, we have been working for almost about 8 to 10 years now. Last year when I joined the organization, we took a very strategic approach with Fivetran. Rather than using Fivetran in individual companies, we molded ourselves to say how can we use Fivetran to build a data factory and empower it.
We started using their product called Managed Data Lake. Our whole infrastructure and the platform that we are building is completely on an open data platform on Apache Iceberg, leveraging the technologies of Fivetran, but we are building it on the AWS S3 foundation. That's super solid and foundational in nature, and it provides the resiliency and the infrastructure that we need. With Fivetran, it was always a question about whether we should build it or buy it. As we have been using Fivetran, we really decided to say how can we provide acceleration, speed, and the momentum that the business has been hungry for since so many years.
We decided to use that approach and then we started really integrating all of these disparate sources, leveraging Fivetran and Managed Data Lake. You talked a bit about the technical implementations. Maybe talk a little bit about now with the large portfolio of companies, how does that affect cross-brand visibility?
How does that affect cross-selling and upselling by utilizing the managed data lake and the technology decisions we've made? How has that affected the business? When you build a vision for a data factory, today we are able to run the mission and vision of our commerce and do the reporting, but it's not at the level that we want in terms of empowerment of our customers.
When we enable the data factory, we are able to build once and consume many times. When you bring in data assets, govern those data assets, and provide quality on those data assets in one lens, rather than having 10, 20, or 30 different disintegrated systems, you gain significant advantages. For example, because of the data factory, we now have a very focused view on how targeting is working, what our achievement scores are, how the funnel is working, and what our lead generation looks like. We can now start putting more focus on enterprise reporting and standardized data definitions.
Everybody focuses on the technology side of the house, but for us, it was more around change management, really making sure that the business has buy-in. It's not about rolling out a data factory. It's about solving the business problems we see currently while also building a data factory that is built for the future, that is AI-enabled, and that has an AI migration lens coming from legacy systems as well. We focus on enterprise-level solutions, building once and consuming many times, and providing self-service capability to our business users.
Regarding the Amazon S3 side, what was the decision process like to choose S3 from a storage perspective? In our current landscape, we have AWS, Azure, and many other cloud providers. Because we have been working with AWS for quite a while, we wanted a rock-solid foundation layer. Everything has to be fundamentally shaped so that you are not just able to bring the data, but you are able to transform and mold it. S3 provides us that capability.
When you have about 100 different connectors running on a minute, 10-minute, or hourly scale, you really want that integration to be flawless, and S3 is probably one of the best engines out there. When you couple that with Fivetran and a managed data lake, building medallion architectures on an open architecture using Apache Iceberg, you are not just building for the future, but you are building for use cases that might come when you are looking at the infrastructure from a scale perspective.
Future Roadmap: AI-Enabled Analytics and the 360-Degree Data Journey
You answered the follow-on question I was going to ask about how Fivetran helped accelerate that centralization and the managed data lake service. Are there any other projects specific to the managed data lake service and S3 that you can describe to the audience that you have in the roadmap? From a future project perspective, our first goal was to lay out the factory. Now that we are done with that goal, we are really empowering and enabling our infrastructure to be AI and ML ready.
What that means is that today, if we have to look at customer churn or customer retention, because the data was in so many different landscapes, it was very hard to do that. Now that we are bringing the data in one place, we can run sophisticated algorithms across the portfolio of companies to see, for example, if a customer was buying products 17 and 10, we now have full visibility into how you can do upsell, cross-sell, propensity modeling, customer churn, and retention analysis.
In terms of helping us, we are also planning to leverage Fivetran's ecosystem holistically. We are a big DBT shop right now, and with the integration of Fivetran and DBT, that is going to accelerate us in our journey of automating what I call shift-left technology. Shift-left means you really want to govern the data and qualify the data as soon as you load them, rather than bringing them into your ecosystems all the way on the tail end. We are a big fan of doing quality checks at shift-left when you bring it from the source systems because bad data is dead data. If you bring in bad data, you will have to pay a lot of technical debt.
So you began to touch on a topic that I wanted to bring us back to, which is that 360-degree journey.
As some may know, Fivetran and DBT are in the process of becoming one company, which is very exciting, and more of that is unfolding as the weeks go on. We have also gone through a process of acquiring a company called Census around reverse ETL. When you think about that 360-degree journey, incorporating Fivetran, DBT, and Census, how does that play into the architecture that you have already developed up to this point?
I think with EverCommerce, as I said, we are very unique. We are a very digitally savvy company, so we do not think technology first. We think about what business transformation we want to lay out. From a future perspective, we created something called the DNA flywheel, and the DNA flywheel really understands who is the producer of the data and who is going to consume that data.
Fivetran, sensors, DBT, and a lot of other technologies are part of this. However, for us, we do not go to the business and say here are the technologies. We say we are building a factory for the future, which is the data factory, and the data factory is powered by an engine like Fivetran and DBT. This is to really make life easy for our engineers. Then we are putting AI and ML on top of it with API as a service and data monetization, which is our nirvana state going forward, to really see how we can enable the businesses at scale. It is all about digitally savvy business units driving forward innovation.
There is a lot there. You have deployed a lot, and you have been with the company for not a long time. What you have accomplished with your team in that short period of time is amazing, with partnering with the AWS team and the Fivetran team. I know you have a whole series of other vendors that have been a part of that as well. We are going to make everybody available to stop by our booth, booth 1660. Kalpit will be there as well to answer any questions as it relates to the information that we shared today regarding Fivetran and what we are doing with EverCommerce and what we are doing with some of the other companies out there.
I want to thank everybody for joining us. We appreciate it. Thank you everybody for letting us share our stories. Thank you, Fivetran, for being good success partners along with AWS. We appreciate it.
; This article is entirely auto-generated using Amazon Bedrock.
Top comments (0)