To get the most of your data, you need a robust analytics stack that provides valuable, effortless insights to both data scientists and business users alike.
Unfortunately, getting there isn't always half the fun. When vetting tools to make it possible, you have to consider variables like price, performance, support, and so on.
If you Google your options, you’ll come across many articles talking about the benefits of using heaps of possible data tools. It can quickly get confusing and overwhelming if you don’t narrow it down. The first step is to identify the right ETL tool to Extract data from disparate sources, Transform that data by cleaning it and making it easy to analyze, and Load it into a data warehouse.
When you research your options, you’ll notice that Fivetran and Stitch come up a lot. That's no surprise: Stitch was acquired by Talend for $60M a few years back and Fivetran recently became a “unicorn.” But are these popular ETL tools right for your business?
We did our homework sifting through the different benefits of using these tools. We compared costs, integrations, transformations, and more to give you an idea of what you’re in for with these popular data tools.
To provide a bit of background, Stitch is a cloud-based ETL tool built for developers and small businesses. As a Talend company that’s part of Talend Data Fabric, it’s equipped with a ton of integrations and rapid data syncs, but that’s all it does. If you’re looking for storage, a SQL workbench, or report generation, you’ll need to look elsewhere.
Aside from niche integrations that Stitch may not support, Stitch’s biggest drawback is the inability to do in-tool data transformations. This means that you’ll have to depend on another platform to prep your data for analysis. Another drawback is that Stitch relies heavily on JSON, which could be annoying—or an outright blocker—depending on your data and team.
Fivetran is also a cloud-based ETL tool built for data analysts and data engineers. It comes with a wide array of built-in integrations that help users push data into a data warehouse or BI tool. In this scenario, you can quickly extract data from a variety of external platforms like Adwords, Hubspot, Google Analytics, Mailchimp, and Salesforce.
Although Fivetran is reasonably user-friendly, new users report that the UI is a bit clunky. Managing transformations in Fivetran is a non-starter, but you may find a viable workaround in its prebuilt packages for dbt. To store your data you’ll have to push it into a BigQuery data warehouse (for an extra fee, of course) so you can work with your data directly or pipe it into a BI tool.
One of the key factors that impact any decision around ETL is price. Understanding how each tool structures their pricing model is essential to understanding what you’re likely to pay.
Stitch boasts a two tier pricing model: Stitch Standard (up to five users and ten data sources) and Stitch Enterprise (custom), both based on the number of raw rows you use. In the standard package, you can use five million raw rows for $100 per month or go up to 300 million raw rows for $1,250 per month (and there’s also an option of paying annually).
These raw rows include deletes, new rows, and record updates from a limited number of sources. To access advanced features like custom integrations, priority support, and robust data security protocols, expect to pay extra.
While Stitch focuses on raw rows, Fivetran pricing follows a consumption-based model concentrating on active rows. This means that whether the record is updated once or a million times, it will still count as a single “active” row. More often than not, compared to active rows, raw rows tend to have a higher count. This is attributed to the manner in which the system generates new records and updates existing ones.
There’s no longer a minimum spend requirement and pricing starts at $1/credit. However, that starter tier may not have all the features you need, so be prepared to pay more for more robust security or user permissions.
Regardless of which ETL tool you go with, actual cost will vary so keep an eye on what’s really included and what might cost extra. Another thing to consider is the need for other tooling to work with your data. Adding on a data warehouse is possible with either option, but can be costly and time consuming.
While price is certainly an important factor to consider, ETL is all about the integrations. The good news? Both tools are pretty well stocked with data connectors.
Stitch supports about 100 database integrations, including SaaS integrations, data sources, and eight data warehouses. If those aren’t enough, you can use their open-source toolkit to write scripts and move data according to the open-source Singer project’s standards.
Fivetran supports over 150 data sources from apps to social media platforms as well as a variety of databases and data warehouses where you can store your data. Along with completed connectors, their integration page lists connectors that are on the roadmap, so be sure to keep an eye on whether a connector is actually complete before buying.
As mentioned earlier, both Stitch and Fivetran are ETL tools that both do some common transformations...and that have some limitations.
As Stitch is part of the Talend family, you can use some of their tools to engage in pre-load transformations. You can also transform data using an external processing engine like MapReduce. These transformations are defined in Java, Python, and SQL.
With Stitch, you also have the option of pushing the data into a data warehouse and processing it there (aka, can opt for an ELT process if you choose). However, within the Stitch pipeline you can only transform the data based on the tool’s limited compatibilities and relevant data destinations.
With Fivetran, you can’t transform the data before loading. However, they now offer a pretty sweet copy and paste SQL feature for post-load transformations. There’s also an option of using dbt integrations—including a handful of prebuilt packages—for data transformation.
Even if you have a team of data scientists on board, support is just as important as any other variable in the ETL decision process. There’s nothing worse than submitting a ticket when a connector fails and not knowing when you’ll get a response.
Whether you’re on the Standard or Enterprise tier, you’ll get in-app chat support. However, phone support and video chat (including product and SLAs) are only available to Enterprise customers. As Stitch is open source, there’s quite a bit of documentation out there. If you come up with your own improvements, you can also add them to their library of content.
Fivetran’s Starter and Standard plans come with 24/7 global support, but if you want to get your answers within an hour, you have to pay for the Enterprise tier. There are rave reviews online about Fivetran’s customer service representatives' technical knowledge, which might be worth considering. Documentation is comprehensive and easily accessible to everyone.
One word of caution: While both Stitch and Fivetran offer different levels of support, neither provides any product training. That could make onboarding and ramp a lot trickier than it needs to be.
Stitch is great at replicating relational databases. It’s also good at managing the conversion of data types and schemas and is capable of incremental replication. However, it’s best for companies with limited consumption. It’s also a good choice if you only have a marginal need for support.
Fivetran is easily accessible to business users and is suitable for data analysts who want to do more analysis than coding. However, users report that the error messages aren’t always helpful and that the UI comes with a learning curve.
If you don’t want to deal with finding an ETL tool and setting up and managing a data warehouse to go with it, neither Stitch nor Fivetran is a great option. However, Panoply might be the answer you’ve been looking for.
Panoply makes it easy to sync, store, and seamlessly access your business data. It combines a robust data warehouse and code-free ETL tool so you can connect to your data, query it directly in the Panoply platform, and push it to your favorite BI tool. Best of all, Panoply is seriously easy to set up and demands minimal upkeep.
To explore what Panoply has to offer, start with a 14-day free trial—there are no limitations on the features you can use and no credit card required.