- Reverse ETL vendors use four different billing units: monthly tracked users (MTUs), active rows, sync operations, and destination count. The same data volume can cost dramatically different amounts depending on which model applies to your sync pattern.
- Hightouch prices on MTUs for its AI Audiences and activation features, with a free tier capped at 1,000 MTUs; Census moved to Fivetran’s consumption-based model after its acquisition, which changes the cost math significantly for high-frequency syncs.
- RudderStack separates event volume from warehouse sync costs, making it cheaper for teams that already run event pipelines but want to add reverse ETL on top.
- Segment’s reverse ETL is priced inside Twilio Segment’s broader MTU model, so teams already paying for Segment may add reverse ETL at low marginal cost, but teams buying it standalone rarely find it competitive.
- The cheapest vendor flips based on three variables: record volume, sync frequency, and destination count. A team syncing 500K rows monthly to two destinations at daily frequency will rank vendors differently than a team syncing 50K rows to six destinations every hour.
Reverse ETL pricing varies by billing unit, not just by tier. Hightouch charges on MTUs, Census shifted to consumption-based pricing under Fivetran, RudderStack bills on event volume and warehouse syncs separately, and Segment bundles reverse ETL into its existing MTU model. For most mid-market teams, the cheapest option depends on sync frequency and destination count more than raw row volume. No single vendor wins across all use cases.
Why Reverse ETL Pricing Confuses Even Technical Buyers
Most people approaching this category assume vendors price roughly the same way, the way cloud storage does: more rows, more cost, scaled linearly. That assumption leads to bad vendor selections. Reverse ETL vendors built their pricing around different bottlenecks in their own infrastructure, and those choices compound quickly at real production volumes.
There are four billing primitives in this market. MTUs (monthly tracked users) count distinct identifiers synced across any destination in a billing period. Active rows count the records actually written during a sync, regardless of how many users are in your warehouse. Sync operations count the number of pipeline runs regardless of record count. Destination slots charge per connected output endpoint, sometimes regardless of volume. Most vendors combine two or three of these.
The practical consequence: a team running nightly syncs of 200K CRM records to Salesforce and HubSpot looks completely different to each vendor’s billing engine. Before requesting a single quote, you need to know your record count, your sync cadence, and your destination count. Those three variables determine which vendor is cheapest by a factor of two or three, not a rounding error.
For teams still figuring out whether reverse ETL is the right architecture at all, the ETL vs reverse ETL vs CDP explainer covers how these tools fit into a modern data stack before you get into pricing specifics.
How Does Hightouch Price Its Reverse ETL Product?
Hightouch’s public pricing page offers a free tier supporting up to two destinations and syncing up to 1,000 MTUs. Paid plans start at what Hightouch positions as a self-serve entry point, with pricing scaling on MTU volume. For larger deployments, Hightouch moves to annual contracts with custom pricing.
Hightouch’s MTU model means that a user appearing in a sync to Salesforce and in a separate sync to Marketo counts as one MTU, not two. That cross-destination deduplication is a meaningful design choice: teams with high destination counts are not penalized for routing the same user to multiple systems. The catch is that even users synced with no change in data still count toward your MTU ceiling if they are part of an active sync model.
Sync frequency does not appear as a separate line item on Hightouch’s self-serve plans, but at enterprise volumes the contract terms often include limits on how frequently syncs can run. Teams planning sub-hourly syncs to multiple destinations should confirm this during any enterprise negotiation. Hightouch also offers a warehouse-native AI Audiences product layered on top of the core sync infrastructure, priced separately, which can inflate the total cost of ownership if teams adopt both products.
How Does Census Price Its Reverse ETL Product?
Census joined Fivetran in late 2023 and has since moved toward Fivetran’s consumption-based pricing philosophy. That is a notable shift from Census’s earlier per-seat, per-destination model. Consumption pricing ties cost to the volume of data processed rather than to user counts or destination slots, which advantages teams with moderate data volumes and many destinations, but penalizes teams running large full-refresh syncs frequently.
Under the consumption model, a daily full-refresh sync of 1 million rows costs materially more than an incremental sync of only the changed records, even if the end result in the destination system is identical. This puts a premium on having well-structured change-data-capture logic in your warehouse before you connect Census. Teams running raw, unoptimized full-refresh syncs will see their costs escalate faster than they expect.
Census’s free plan historically supported up to 10 destinations and a limited number of syncs per month, though the post-Fivetran pricing structure is evolving. Any team currently comparing Census to alternatives should treat public pricing pages as a starting point and request a current quote directly, since the product has been in pricing transition.
How Does RudderStack Price Its Reverse ETL Product?
RudderStack’s pricing covers both its event streaming pipeline and its warehouse-native reverse ETL syncs under different counters. Event data, which flows from your application into the warehouse, is priced on event volume. Reverse ETL syncs, which pull data back out of the warehouse to destinations, are priced separately based on the number of rows synced.
This dual-meter model creates an interesting scenario for teams already running RudderStack for event collection. If you are paying for event pipeline volume anyway, adding reverse ETL destinations adds row-sync cost but does not compound your event costs. For a team already embedded in RudderStack’s infrastructure, the incremental cost of reverse ETL can be lower than bringing in a standalone vendor like Hightouch or Census, because you avoid a second vendor’s base platform fee.
RudderStack’s free tier (Starter) supports limited event volume with access to most destination connectors. Its Growth tier and above add SLA guarantees and higher throughput limits. RudderStack’s open-source option, available on GitHub, allows self-hosting with no per-row cost, though operational overhead of managing the infrastructure replaces the licensing cost.
Does Segment Charge Separately for Reverse ETL?
Twilio Segment’s public pricing bundles reverse ETL functionality inside its broader Connections product. Teams on Segment’s Team or Business tiers can connect warehouse sources as data inputs and sync to downstream destinations, but the cost is not broken out as a separate reverse ETL line. Instead, records synced from the warehouse count against the plan’s MTU or event volume allotment.
For teams already on Segment, this bundling is occasionally an advantage. If your warehouse sync volume fits within your existing MTU cap, reverse ETL is effectively included at no additional charge. For teams evaluating Segment purely as a reverse ETL tool, the economics rarely compete with dedicated vendors. Segment’s MTU pricing is structured around companies that also use it for event collection, CDP functionality, and audience management. Buying it only for warehouse-to-destination syncs means paying for significant overhead.
Teams evaluating Segment alongside dedicated reverse ETL tools will find a broader comparison useful. The Hightouch vs Segment breakdown covers how the two architectures differ before you get into line-item costs.
The AboutMartech Billing-Unit Test: Four Checks Before You Request a Quote
Before comparing vendor pricing pages, run these four checks against your own data. The answers will tell you which billing model advantages you and which vendors to prioritize.
- Distinct record count: Count the distinct identifiers (user IDs, account IDs, contact IDs) that will appear in any sync in a given month. This is your effective MTU exposure. If this number is high but your actual changes per sync are low, active-row billing may be cheaper than MTU billing.
- Change rate: What percentage of your records actually change between syncs? If 95% of your warehouse records are static month over month, a full-refresh MTU model charges you for records that never needed to move. Incremental sync support with row-level diffing is worth paying a premium for.
- Sync frequency: Daily is standard. Hourly is not. If your use case requires near-real-time sync, confirm which vendors support sub-hourly runs on self-serve plans and which require enterprise contracts. Billing often changes at that threshold.
- Destination count: Count your intended destinations now and in 18 months. Per-destination models penalize breadth; MTU models do not. If you plan to sync the same audiences to Salesforce, HubSpot, Intercom, and Google Ads, destination-slot pricing adds cost that MTU models absorb.
What Does Reverse ETL Actually Cost? A Worked Scenario Across Four Vendors
Consider a B2B SaaS marketing team with 75,000 active contacts in their data warehouse. They run daily syncs to three destinations: Salesforce (for sales routing), HubSpot (for email nurture), and LinkedIn Ads (for audience matching). About 15% of records change between syncs on any given day, meaning roughly 11,250 rows are written per sync run. They run 30 sync cycles per month.
The following table uses each vendor’s publicly available pricing information to estimate monthly cost for this hypothetical scenario. All figures for Hightouch and Census are estimated ranges based on their stated billing models, not confirmed vendor quotes, contact each vendor for actual pricing. Where pricing is not publicly disclosed, ranges are noted and the recommendation is to request a quote directly.
| Vendor | Billing Unit | Scenario Exposure | Estimated Monthly Cost | Key Cost Driver |
|---|---|---|---|---|
| Hightouch | MTUs | 75,000 MTUs, 3 destinations | Estimated paid tier range; request quote for exact figure | MTU count; destinations included |
| Census | Consumption (rows processed) | 337,500 rows/month (11,250 × 30) | Estimated range based on consumption model; request quote for current rates | Full vs incremental sync; row volume |
| RudderStack | Rows synced + event volume | 337,500 rows/month reverse ETL only | Free tier up to limit; Growth tier quote-based | Row volume; event pipeline counted separately |
| Segment | MTUs (bundled with CDP) | 75,000 MTUs against CDP allotment | Included in Team/Business tier if MTU cap fits | Overall MTU ceiling; reverse ETL not a separate SKU |
The scenario above favors Segment if this team already pays for Segment’s CDP tier, because the 75,000 MTUs may fit within an existing allotment. It favors RudderStack’s free or low-cost Growth tier if the team does not need event streaming. Census’s consumption model becomes more expensive if the team is not using incremental syncs, because 337,500 rows per month processed in full-refresh mode costs more than the same rows synced incrementally. Hightouch’s MTU model is destination-count-agnostic, which matters if this team adds a fourth or fifth destination.
How Does Sync Frequency Affect Reverse ETL Costs?
Daily syncs are the default assumption in every vendor’s self-serve pricing. Moving to hourly changes the economics for most vendors. The same 11,250 changed rows synced 720 times per month (hourly) instead of 30 times costs vastly more under consumption-based billing than under MTU billing, because MTU models charge per unique user per month regardless of how many times that user’s record is touched.
This is the single most underappreciated cost driver in reverse ETL. A team running hourly syncs to keep Salesforce current with real-time behavioral signals from their warehouse can face 24x the row-write cost compared to daily syncs, while paying the same MTU cost under Hightouch’s model. Teams with near-real-time requirements should explicitly model their sync frequency against each vendor’s billing unit before assuming that the cheapest daily-sync vendor is also the cheapest hourly-sync vendor.
RudderStack charges on row volume written, so hourly syncs multiply that cost directly unless the team uses incremental sync (writing only changed records each run). Census’s consumption model behaves similarly. Hightouch’s MTU model is more insulated from frequency changes, which makes it structurally advantaged for high-frequency use cases where the same users appear across many sync runs.
What Is Active Rows Pricing, and Which Vendors Use It?
Active rows billing counts the number of records actually written to a destination during a sync run. If your warehouse has 500,000 contacts but only 50,000 changed since the last sync, active rows billing charges you for 50,000 records, not 500,000. This model rewards teams that implement proper change-data-capture logic and run incremental syncs.
Census’s consumption model is the closest to pure active-row billing in this comparison. RudderStack’s row-sync pricing for its reverse ETL product also approximates this model. Hightouch’s MTU model does not behave this way. A user is an MTU whether their record changed or not, as long as they are in an active sync model within the billing period.
For teams with large, relatively static contact databases, active rows pricing is almost always cheaper than MTU pricing. A company with 500,000 contacts where only 5,000 change per month pays for 5,000 rows under active-row billing, versus 500,000 MTUs under a flat MTU model. The crossover point depends on your specific churn rate within the database, but for most B2B databases where contacts are stable, active rows wins below a certain change-rate threshold.
Per-Destination Pricing: Does It Still Exist?
Per-destination pricing, where vendors charge a flat fee per connected output (Salesforce costs X, HubSpot costs Y, and so on), was common in earlier reverse ETL pricing models and still surfaces in some vendor tiers. Census’s historical pricing before the Fivetran acquisition used a destination-based structure. Hightouch’s free tier limits destination count (two destinations on free) but paid tiers do not charge per-destination incrementally.
Per-destination billing is generally disadvantageous for teams building a modern activation stack that routes data to many downstream tools. If you intend to sync to six or more destinations, any vendor still using per-destination billing will price you out of comfortable range quickly. When evaluating vendor tiers, confirm whether destination count affects pricing independently of data volume.
Teams thinking through which tools belong in their activation stack can find the broader vendor comparison in the nine best reverse ETL tools for warehouse data activation, which covers the full competitive field beyond these four vendors.
Annual Commit vs Monthly Billing: What Reverse ETL Vendors Actually Offer
All four vendors offer annual contract pricing at a discount to month-to-month rates. Hightouch, Census, and RudderStack each follow the SaaS convention of approximately 15-20% savings for annual prepay, though exact figures require direct negotiation at enterprise tiers. Segment’s Twilio-parent pricing structure means annual commit terms are often negotiated as part of broader Twilio agreements, which can be opaque for teams not already inside that relationship.
The annual commit question matters most for teams whose usage is predictable. If your contact database is growing 20% month-over-month, committing to an annual MTU tier that fits today’s volume creates an overage problem six months in. Most enterprise contracts include a negotiated overage rate for this reason. Self-serve plans typically allow upgrade to the next tier during the billing period, which avoids hard overage charges but resets cost at the higher tier rate.
Teams evaluating annual commits should model their expected data volume at month 12, not month 1, and choose a tier that fits that projection. Paying for headroom you need in six months is often cheaper than triggering mid-year tier upgrades.
Which Vendor Is Cheapest for a Small Team?
For a team under 50,000 MTUs running daily syncs to two or three destinations, Hightouch’s free tier is the starting point. It supports two destinations and 1,000 MTUs with no credit card required. This covers many early-stage teams validating whether reverse ETL belongs in their stack before committing spend.
RudderStack’s open-source option is cheaper in licensing cost (zero), but requires infrastructure to run and engineering time to maintain. For teams with a data engineer who can manage a self-hosted deployment, it is the lowest total licensing cost at any volume. For teams without that resource, the operational cost exceeds the licensing savings quickly.
Census and Segment are harder to access at low cost without a conversation. Census’s post-Fivetran pricing structure is not as clearly self-serve at the entry level. Segment’s cheapest tier requires MTU volume that most small teams find expensive for a tool that functions here primarily as a connector rather than a full CDP. Small teams are better served starting with Hightouch free or RudderStack open-source, then revisiting Census and Segment when scale or specific connector requirements justify the cost.
The broader question of whether a dedicated reverse ETL tool or a warehouse-native CDP better fits your stack is covered in the comparison of warehouse-native CDPs, which includes pricing context for that adjacent category.
How Do Reverse ETL Vendors Handle Data Volume Tiers?
Each vendor structures tiers differently, but the pattern is consistent: a free or entry tier with hard caps, a self-serve growth tier priced on the primary billing unit, and an enterprise tier that is custom-priced and contract-based. The jump between self-serve and enterprise is where the pricing becomes opaque.
Hightouch’s public pricing page documents the self-serve tiers clearly, with MTU limits at each tier level. Enterprise pricing is not listed. Census’s pricing requires more direct engagement at growth-stage volumes, particularly after the Fivetran integration changed the product’s packaging. RudderStack publishes a Starter, Growth, and Enterprise tier structure with event volume and row limits documented, which gives the clearest self-serve transparency of the four.
Data volume tiers in this market typically reset monthly. Overages above a tier cap are either billed per unit (common in consumption-based models) or trigger an automatic upgrade to the next tier (common in seat-based and MTU models). Confirm which mechanism applies before you exceed a tier, because automatic tier upgrades on an annual contract can create billing surprises that take a quarter to resolve.
Reverse ETL Pricing Side-by-Side: The Quick Reference Table
| Vendor | Primary Billing Unit | Free Tier | Self-Serve Entry | Enterprise | Best For |
|---|---|---|---|---|---|
| Hightouch | MTUs | 1,000 MTUs, 2 destinations | Paid self-serve tiers by MTU band | Custom annual contract | High-frequency syncs, many destinations |
| Census | Consumption (rows processed) | Limited; contact for current terms | Consumption-based growth tier | Custom via Fivetran | Teams with incremental sync, moderate volume |
| RudderStack | Events + rows synced | Starter (event + row limits) | Growth tier, quote-based | Custom annual contract | Teams already on RudderStack event pipeline |
| Segment | MTUs (bundled) | Free (limited MTUs) | Team tier, MTU-based | Business/custom | Teams already paying for Segment CDP |
Frequently Asked Questions About Reverse ETL Pricing
What does MTU-based pricing mean for reverse ETL?
MTU stands for monthly tracked user. Under MTU billing, you pay based on the count of distinct identifiers (user IDs, contact IDs) that appear in any active sync during a billing month, regardless of how many syncs run or how many destinations receive that record. Hightouch and Segment both use this model. It advantages teams with high sync frequency or many destinations, but it charges for inactive records if they remain inside an active sync model.
Is active rows pricing cheaper than MTU pricing?
It depends on your change rate. Active rows billing charges only for records actually written during a sync run. If your database has 500,000 contacts but only 10,000 change per sync, you pay for 10,000 rows, not 500,000 MTUs. For large, relatively static databases, active rows billing is materially cheaper. For databases with high churn or frequent updates across most records, MTU billing can be more predictable since the cost ceiling is the distinct user count rather than a multiplier of changes times frequency.
How does Census pricing work after the Fivetran acquisition?
Census joined Fivetran in late 2023 and has been integrating Fivetran’s consumption-based pricing model. This means costs are tied to data volume processed rather than a fixed per-seat or per-destination fee. Teams that ran Census under its earlier pricing model should re-evaluate their contracts. The practical effect is that teams running full-refresh syncs of large datasets pay more under consumption billing than they did under the older model. Incremental sync logic becomes more important for cost control.
Does sync frequency affect how much reverse ETL costs?
Yes, and significantly under consumption and row-based pricing models. A team running hourly syncs processes 24 times more rows per day than a team running daily syncs, even if the underlying records change at the same rate. Under MTU billing, frequency does not change the cost because the same users are counted once per month regardless of how many times their data is synced. Teams with near-real-time sync requirements should prioritize MTU-based vendors or confirm their row-based vendor supports efficient incremental syncs.
Is there a free reverse ETL tool for small teams?
Hightouch’s free tier supports up to 1,000 MTUs and two destination connectors with no credit card required, making it the most accessible entry point for teams validating the use case. RudderStack’s open-source version is also free in licensing cost but requires self-hosting. RudderStack’s cloud Starter tier has limited free capacity before requiring a Growth plan. Census and Segment do not have clearly documented free tiers for reverse ETL specifically at this time.
Does Segment charge extra for reverse ETL?
Segment does not sell reverse ETL as a separate SKU. Warehouse sync functionality is bundled into its Connections product within the Team and Business tiers. Records synced from a connected warehouse source count against the plan’s MTU or event allotment. For teams already on Segment whose warehouse sync volume fits within their existing plan limits, reverse ETL adds no incremental cost. Teams buying Segment purely for warehouse-to-destination syncing will find the overall plan pricing uncompetitive against dedicated reverse ETL vendors.
How does RudderStack compare on price to Hightouch for reverse ETL?
For teams already running RudderStack for event streaming, adding reverse ETL is incremental to an existing contract, which can make the total cost lower than bringing in Hightouch as a standalone vendor. For teams with no existing RudderStack relationship, Hightouch’s free tier and self-serve MTU model are more accessible at low volume. At scale, both require enterprise contract negotiation. RudderStack’s row-based billing for reverse ETL is cheaper than Hightouch’s MTU model when your distinct user count is high but your change rate is low.
What is an annual commit in reverse ETL, and should I take one?
An annual commit is a prepaid contract for a fixed volume of usage over 12 months, typically offered at a 15-20% discount versus month-to-month rates. It makes sense when your data volume is predictable and you are confident the vendor fits your stack long-term. It becomes a liability if your usage grows faster than the committed tier, triggering mid-year upgrades or overage charges. Teams still validating their sync patterns should stay on monthly plans, even at higher cost, until their volume stabilizes.
The Billing Unit Determines the Winner
No vendor in this comparison consistently wins on price across all scenarios. Hightouch’s MTU model is the right choice when sync frequency is high and destinations are many, because those variables do not compound cost. Census’s consumption model rewards disciplined, incremental sync architectures, and penalizes teams that sync full datasets without change-data-capture logic. RudderStack makes the most sense as an incremental addition for teams already in its event pipeline, where the marginal cost of warehouse syncs is low. Segment is rarely the economical choice for reverse ETL standalone, but is often the zero-additional-cost choice for teams already on its CDP.
The vendor comparison tools worth bookmarking are the ones built around your specific variables. Running the four-check billing unit test before any vendor demo will tell you more about your likely bill than any pricing page will. Teams that go into vendor conversations knowing their MTU exposure, change rate, sync frequency, and destination count negotiate from a position of clarity, not guesswork.
For teams still mapping out which activation tools belong in their stack alongside reverse ETL, the Hightouch vs Census feature and pricing comparison and the RudderStack vs Hightouch pipeline comparison cover the next layer of evaluation once the cost model is settled.
