Allie Beazell is director of developer marketing @ Census. She loves getting to connect with data practitioners about their day-to-day work, helping technical folks communicate their expertise through writing, and bringing people together to learn from each other. Don't be shy, say hi. 👋
This is part 3.1 of our series on evaluating reverse ETL tools. You can read part one here and part two here, or grab a copy of the companion guide, The ultimate guide for evaluating reverse ETL tools.
In this installment, you’ll learn four of the key things you should consider as you evaluate your reverse ETL vendor, including:
- Data connector quality ⭐
- Sync robustness 💪
- Observability 👓
- Security & regulatory compliance 🔒
We’re assuming (since you’re reading this) that you’ve come across material on reverse ETL. You’ve heard about how it can drive personalization and growth at a completely new scale while unblocking data teams to do their best work and unlocking operational analytics.
Good news: It’s all true. Challenging news: There’s a lot to consider when you’re choosing your reverse ETL vendor. That’s why we designed this guide to help data and business leaders understand the key things to look for when partnering with a reverse ETL vendor to add to their modern data stack.
Each reverse ETL on the market is unique, and not all are created equal. Before we get too deep into the considerations to keep in mind, you should know that there are generally two lanes of reverse ETL software on the market today:
- Bundled reverse ETL tools from companies that offer reverse ETL functionality as part of their larger suite of services. This may sound like a good deal (especially if you’re already using the overall suite), but as we talked about up top, reverse ETL isn’t just another pipeline or feature set. Providers that bundle reverse ETL into their main offering often don’t include the same or as many data integrations as dedicated platforms (Example: Domo offers some reverse ETL functionality, but only out to Salesforce, Gainsight, and Netsuite.). In the same way that you wouldn’t want your friendly, neighborhood general practitioner doing your emergency heart surgery, you want precise expertise when it comes to the tools that operate on the heart of your business (your data).
- Dedicated reverse ETL tools (that’s us 😄). There’s an increasing number of companies singularly focused on building the best reverse ETL tool they can to fuel a new era of business intelligence and automation. As the pioneer of the category and operational analytics, we believe that this route toward specialized tools is the correct choice for most companies out there.
Regardless of which lane of reverse ETL tools you decide to go with, you can use the following consideration areas to make sure you choose the option that best fits your use case. These considerations break down into four initial categories:
- Data connector quality
- Sync robustness
- Security & regulatory compliance
Let’s dive in and find the best reverse ETL tool for you.
1. Data connector quality ⭐
Reverse ETL tools are only as good as the quality of their connectors. These tools move data from your data sources and SQL models and push it into your business apps like Salesforce, Marketo, Braze, Hubspot, and Netsuite via its native APIs.
We break down data connector considerations into two categories:
- Connector breadth (features that span individual connectors)
- Connector depth (the features dedicated to the specific connectors you need)
Here’s what you should keep an eye out for within each connector lane.
Connector breadth: High-level data connector considerations
When looking at connector breadth for your potential reverse ETL vendor, there are four things to consider:
- Make sure your vendor offers the high-quality connectors you need today. It should integrate with the business tools (both sources and destinations) your organization needs today. Keep in mind that no tool offers all integration with every app there. Prioritize vendors with a reputation for high-quality connectors, and beware those platforms that boast high-velocity, shoot-from-the-hip integrations.
- Make sure extensibility is a priority for your vendor. Extensibility ensures that if you need a custom integration for a specific need, your vendor can support it well. This generally comes from a Webhook or custom API connector which can handle the last mile of integration with native APIs so you can quickly hook up a tool and enjoy sync robustness all the while.
- Consider specialized connectors vs add-on connectors. Existing ETL or DataOps vendors may offer limited reverse ETL capabilities, but this often means connectors aren’t given the full attention and time they need to stay up to date (which means more troubleshooting for your team).
- Evaluate whether or not you can (or should) use an open-source vs managed reverse ETL tool. While open source tooling offers extra flexibility, the current OSS reverse ETL connectors only support a fraction of the integrations of managed tools. If you’re concerned about getting locked into a proprietary tool, compare connector extensibility over whether or not the tool is open source.
At the end of the day, it doesn’t matter if a vendor offers a large number of half-baked integrations that you either A) don’t use or B) can’t rely on. Checking for these features when evaluating the best reverse ETL tools for your use case so you can trust your tooling as much as you trust your data.
Connector depth: Individual connector considerations
While it’s important to make sure your vendor’s reverse ETL software offers a wide range of connector options for your current and future use cases, you also need to do your due diligence to make sure they offer the depth you need to deliver real-time data where--and how--you need it.
Just as there were four tenants of connector breadth, there are four tenants of connector depth to measure your reverse ETL vendor options against:
- Check to see that the vendor offers support for all the objects you need for each key connector. Unlike the ETL tools you may have used in the past, there are structural nuances to consider when syncing data back into business systems since each business app has its own data structure. This means no standards exist between connectors and multiple objects have different rules between them.
- Determine whether or not the tool helps you simplify these complex object syncs. You shouldn’t have to spend hours architecting mapping to make sure syncs go through. Instead, your vendor automatically handles multi-object syncs and alerts you when an issue arises so you can quickly resolve it when needed.
- Check to see if your reverse ETL vendor supports event and tabular data. The ability to tap into behavioral data sources empowers a whole new generation of personalization for your outreach efforts, and, as such, the event data category is growing rapidly. Make sure you’re set up for success with a tool that can help you unlock insights from this growing data trend.
Phew! 🧠 That’s a lot of things to consider just for your connectors, but they’re all important. Don’t worry, we’ve included a checklist that consolidates all these consideration areas at the end to make it easier to check as you compare vendors.
2. Sync robustness 💪
Syncs define the rules and manage the workflow around how your data is sent from your single source of truth (AKA your data warehouse) or another source to your destination tools. You likely have multiple syncs pulling from the same data connector for different SQL models and use cases. This means sync reliability is, arguably, the most critical part of your reverse ETL tool.
When choosing a vendor, you should consider the following six sync characteristics across all syncs, regardless of connectors:
- Sync reliability or Just Works™. This sounds simple, but the advantage of sync reliability can’t be overstated. You need to know that no matter what happens (rate limits, network issues, etc.) your data will eventually sync, without duplication or loss (or, as Fivetran calls it, idempotence). Your reverse ETL should seamlessly retry syncs if they fail, intelligently handle changes to data models, and alert you to issues, unlike most point-to-point systems.
- Data validation. Your business apps are only as useful as your data is trustworthy. Your reverse ETL vendor should offer automatic deduping capabilities as a line of defense against bad data quality. This is particularly important as your business teams become more self-service and sync data to Google Ads or Facebook Ads where dupes cost you cold, hard cash.
- Incremental syncs. Your tool should only sync the data that’s changed, rather than all existing data in a “full sync”. Incremental syncs, combined with batch APIs, ensure you steer clear of API rate limits and don’t rack up high API bills from your apps.
- Automated sync scheduling. We fundamentally believe that no one should spend their time stressing over sync schedules. You should just be able to set your syncs, know they work, and get on with your data-driven life. As such, make sure your vendor supports sync scheduling automatically on a schedule of your choice, from weekly to continuous syncs.
- Triggering sync scheduling. In addition to an automated sync scheduling feature, your vendor should support trigger sync via your orchestration layer (Airflow), your models (dbt), and programmatically via API.
- Sync speed. Nothing is more annoying than waiting for data you know is in your cloud data warehouse, but not yet available in your tools. Make sure that your vendor offers reasonable sync speeds based on your largest volumes of data that satisfy your freshness demands, particularly for large volumes of data or data that fuels real-time analytics.
Now that you know what you’re looking for when it comes to sync robustness, let’s dive into the next integral part of any good reverse ETL platform: Observability features.
3. Observability 👓
Try as we (and any other company might) there are just some times that the data Just (doesn’t) Work™. In the event of this, your reverse ETL vendor should alert you quickly, help you understand what’s gone wrong and how to fix it.
To confidently help you fix data issues, this means the best reverse ETL tools must offer the following five observability features.
- Alerting. Your reverse ETL tool shouldn’t just limit to alerts for failures, but include them for invalid and rejected records and general error messages to help you take proactive action to fix issues before they cause major breaks. These alerts should be available quickly via email, as well as Slack, when something goes wrong.
- Integration with your favorite monitoring tools. Your reverse ETL vendor should meet your stack where it is, and this includes easily integrating with your favorite monitoring tools (e.g. Opsgenie, PagerDuty, Datadog, etc.). This could be via a native integration, or through other levers such as webhooks and ensuring logs can be parsed by those systems.
- Detailed logging. You should be able to easily reference and understand what records failed and why, as well as which were successfully synced so you can feel confident your data has synced correctly or quickly troubleshoot any issues you’re alerted to.
- Usage audit logs. You should be able to track all usage of your reverse ETL tool from the user level, and understand exactly who’s made any changes to models or sync configurations and when. This includes any changes made by vendor support teams.
- Sync rollbacks. Your reverse ETL tool should allow you to easily rollback syncs to an early data state when syncs fail to make remediation of errors quick and painless.
It may seem like the work your data engineers do is magic, but the reverse ETL layer of your pipelines shouldn’t be a black box of mystery. Instead, you should be able to see what the tool is doing, which data it’s accessing, and where that data goes. No white rabbits or magic wands are required. 🪄
Speaking of transparency and peace of mind, let’s talk security and regulatory compliance.
4. Security & regulatory compliance 🔒
Your customers (and regulators) expect your company to act as a good steward of the data you collect. As such, your reverse ETL should be a security and compliance asset, not another liability. When it comes to moving the sensitive personally identifiable information (PII) of your customers between systems within your company, this really isn’t a time to cut corners.
When evaluating a reverse ETL vendor there are four key considerations to keep in mind:
- Does this tool let you own your own data? Any good reverse ETL tool should seek to minimize both the number of places and the amount of time your customer data is stored. This means:
- Your vendor should minimize data retention so they don’t store your data for a moment longer than needed to sync it to the destination.
- Your vendor should ensure data stays in your environment. In order to only perform incremental syncs, it’s common for reverse ETL tools to store a snapshot of your data, so each sync can create a ‘diff’ off between data states. Your tool should seek to minimize the number of systems it stores that data in, and, ideally, store snapshots in the same place as the source. Additionally, your vendor should limit the number of systems your data moves through.
- Does your vendor meet progressive regulatory standards? As a baseline, your vendors should be SOC 2 Type 1 and a best-option will also sport SOC 2 Type 2 regulatory compliance to reflect deeper security maturity and investment. Beyond this, it’s important to make sure your vendor meets industry-specific standards that apply to your organization, such as the Health Insurance Portability and Accountability Act (HIPAA) or the Payment Card Industry Data Security Standard (PCI-DSS).
- Does your vendor encrypt data at rest and in transit? This should be covered by SOC 2 compliance, but double-check your vendor offers data encryption both when information is at rest and in transit.
- Does your vendor leverage the best security available for API connectors? Check to make sure your vendor requests as little OAuth scopes as possible for SaaS applications (subject to SaaS provider support, of course). This also means your vendor should take an active role in staying up to date with the emerging security capabilities of business tool APIs (e.g. mirroring the support for hashed data currently offered by Google Ads and Facebook Ads).
It’s not enough that your data be correct. It needs to be secure and safe, too. The best reverse ETL tools check all the above security boxes to ensure this trust today and into the future.
Ready to learn more? Check our final part of the series on evaluating reverse ETL vendors: 3 (more) key considerations for finding the best reverse ETL tool where we'll dive into requirements for ease of use, community and vendor support, and pricing. Or, if you’re ready to start shopping around, grab The ultimate guide to evaluating reverse ETL tools.