The Center for Data Innovation spoke to Tejas Manohar, co-founder of Hightouch, a startup in San Francisco that helps data move from data warehouses to customer-facing teams within a business by using its reverse ETL platform. Manohar discussed various uses for the reverse ETL approach, and what he envisions to be the future of data integration.
Gillian Diebold: Most people are probably familiar with ETL (extract, transform, load), the process that organizations use to make data available for queries and analytics. What is reverse ETL, and what are some everyday situations where this technology can alleviate issues businesses might encounter?
Tejas Manohar: The whole concept of reverse ETL is in taking the data warehouse, which is generally seen as a destination of data, and turning it into a source for all of your business systems across sales, marketing, and other departments. Two really common examples for reverse ETL are one, for helping sales teams at B2B companies, and two, helping marketing teams at B2C companies. The common sales use case for reverse ETL is mostly at newer generations of companies where the products are digital or SaaS. Sales reps typically live out of a sales CRM like Salesforce. Salesforce typically includes information like the last time a customer was contacted, notes, etc. but in an increasingly digital world, a lot of the key information isn’t manually recorded, it’s digital. If I’m a B2B sales rep at a company like Zoom, I want to know how many users are using Zoom at the company I’m trying to sell a larger contract to or how many minutes they’re spending on Zoom per month. This information exists in the data warehouse, and Hightouch allows you to easily push it to Salesforce via reverse ETL without any coding or engineering work. The use case for marketing is the same but more automated and at scale. Instead of equipping sales reps, you’re powering the automated marketing campaigns over email, push, online ads, etc. with all the user data in your warehouse. You can use any data points from your warehouse to inform who you’re targeting or what they’re seeing.
Diebold: What are the benefits of allowing more people within a company to access and use data?
Manohar: In this day and age data is one of the big differentiators between companies. Most large companies see their ability to utilize data as the core way they can differentiate against their competitors given that so much of production is getting commoditized across software or even manufacturing. The biggest problem in analytics is not actually crunching numbers or consolidating data into one place anymore. Data infrastructure has advanced quite a bit, and a lot of the challenges there have been solved by innovation in cloud data warehouses. The biggest problem is really analytics enablement, or enabling people to actually use that data effectively. Typical analytics tools are not very approachable to the masses. There’s a whole movement in the analytics and business intelligence (BI) space to create so-called “self-service” BI tools that anyone can use to poke around and explore data themselves. Personally, I don’t think self-service BI is the right solution. To really make the data useful to the masses, it needs to be in the tools they’re used to using, whether that’s Salesforce for a salesperson, Braze for a marketer, or NetSuite for your finance team. That’s where reverse ETL comes in.
Diebold: How does Hightouch help businesses better understand their customers?
Manohar: What Hightouch actually helps with is the process of helping a business use customer data in all of its operations. So, if you imagine that you are a salesperson at a company with a thousand potential customers that you can reach out to, you want to figure out which ones have the highest value—which ones actually have an opportunity that you can help them out with. Hightouch would help you get the right information about what customers are doing in your application or on your marketing website, and sync that up to your sales CRM so that you can help define how you’re structuring your day as a salesperson. On the marketing side, even consumers don’t really enjoy getting emails that aren’t personalized and targeted to them. Relevance is one of the most important things when you think about marketing. If you actually want the product or the offer that an advertisement is targeting you with, you’d likely appreciate the marketing campaign. The biggest step to making marketing relevant is allowing marketers to drill into the data themselves and use it to build personalized, targeted campaigns. Hightouch allows marketers to do that with all the data that companies already gather in one place for analytics.
Diebold: What do you see as the future of the data integration space?
Manohar: If you look back 10 years or so, most data integration solutions were point-to-point solutions and totally proprietary solutions. I think the future of integration is a lot more composable, so people want to use the software for different tasks. All the software should be interoperable with each other through standards like the cloud data warehouse. What we’re seeing in the data integration stage is that everyone is centralizing all the data in a cloud data warehouse for data analytics purposes. The next generation of data integration platforms like Hightouch will be built on top of standards like the cloud data warehouse rather than solving every problem from scratch, like data ingestion, data transformation, and syncing data to other tools all in one platform. I also think the future of the data integration space is declarative, which means you tell the software what you want to happen, for example, “I want this field to show up in Salesforce this way.” And then the software actually figures out how to make that happen, versus in legacy data integration software, where you have to tell the software how to do every little step to get anything done—for example, what APIs to call in Salesforce, how many times to retry an operation and how much time to wait in between, how many records should be included in each API call, etc. The world is moving towards an approach that is much more declarative, where you just tell the computer what you want.
Diebold: How do you see your product evolving in the future?
Manohar: At Hightouch, we see the base layer of our product, which we call reverse ETL, as the foundation of what we’re doing for the future. We want to create a really good platform to help you move data from the data warehouse to different functional systems around a company. The platform should be really robust in that it should not just move the data but it should be efficient, provide a lot of visibility into what’s happening under the hood, be configurable and customizable to support the needs of the most complex enterprises, and be able to alert customers when things go wrong.
Past that, we believe that the future is to build vertical interfaces on top of the Hightouch products to make the process of reverse ETL accessible to more populations within a company. An example is a product that we released a couple of months ago called Audiences, which allows not just data engineers and data analysts to perform reverse ETL, but also allows marketers to move data from the warehouse to the systems they choose. The process in which they move the data is called building an audience. An audience is a list or a cohort of users that match certain criteria. Once you build an audience in Hightouch on top of your data warehouse, instead of the typical process of emailing CSVs around and uploading them to various systems, marketers can sync these audiences to the marketing and ad tools of their choices through the same process of reverse ETL without needing to be technical and have knowledge of SQL. Every day, we’re learning more about how our customers are using reverse ETL and a deep understanding of those use cases will feed into our future vertical product development at Hightouch. But, one thing we are confident in is that we should never become the marketing tool or sales tool itself. That is a very different problem. Hightouch is and always will be a data platform, but we want to make it accessible to everyone at a company that has a warehouse, not just the data analysts.