Fivetran pricing in 2026: how MAR billing works and what drives the bill
Fivetran bills each connection by monthly active rows. On Standard, the first million MAR costs $500 a month at list price, and the rate falls per connection.
Fivetran charges each connection separately for its monthly active rows (MAR). On the Standard plan, Fivetran’s Service Consumption Table (V.11, effective November 1, 2026) lists a $5 base charge plus $495 per million MAR for a connection’s first million rows, which comes to $500 a month at 1 million MAR. The rate then falls to $170 per million, then $58, but only within that one connection. Rows from different connections are never pooled.
MAR (monthly active rows): the number of distinct primary keys a connection inserts, updates or deletes in your destination during one calendar month. A row that syncs 30 times in a month counts once (Fivetran docs, read November 3, 2026).
Every figure below comes from Fivetran’s official pages, read on November 3, 2026. The per-MAR rates live in the legal Service Consumption Table; the pricing page only gives the formula, Total monthly usage x spend rate = connection monthly cost.
What exactly does Fivetran bill for?
Three meters, added together each calendar month:
- Connection MAR. Rows moved from a source (Shopify, Google Ads, a Postgres database) into your warehouse.
- Activation MAR. Reverse ETL, rows pushed from the warehouse into tools like Salesforce or HubSpot. Since February 1, 2026 this is also usage-based and has its own price curve (2026 pricing updates).
- Transformation model runs. The first 5,000 successful runs a month are free; after that the list price is $0.01, $0.007 and then $0.002 per run as volume grows (pricing page).
The rest of this piece is about connection MAR, the meter every pipeline pays.
How is a connection’s monthly cost calculated?
The Service Consumption Table gives the formula in plain words: Base Spend per Month + (actual MAR − MAR Threshold) × Incremental Spend per Million MAR. These are the list prices for connection MAR, as published in V.11:
| Connection MAR in the month | Standard: per million above threshold | Standard: base at threshold | Enterprise: per million | Business Critical: per million |
|---|---|---|---|---|
| 1 to 1M | $495.00 | $5.00 | $661.67 | $1,061.67 |
| 1M to 4M | $170.00 | $500.00 | $226.67 | $362.67 |
| 4M to 16M | $58.00 | $1,010.00 | $77.33 | $123.73 |
| 16M to 64M | $19.70 | $1,706.00 | $26.27 | $42.03 |
| 64M to 250M | $6.72 | $2,651.60 | $8.96 | $14.34 |
| 250M to 1,000M | $2.29 | $3,901.52 | $3.05 | $4.89 |
| Above 1,000M | $1.18 | $5,619.02 | $1.57 | $2.52 |
Source: Fivetran Service Consumption Table, V.11, effective November 1, 2026, read November 3, 2026. List prices before any contract discount.
The formula checks out against Fivetran’s own worked example. The pricing page shows a Standard-plan company with 1 to 200 employees whose Facebook Ads connection moved a median of 34,479 MAR for $22.06, and whose Google Ads connection moved 88,240 MAR for $48.67. Plugging those volumes into the first row of the table gives $22.07 and $48.68. The four-connector example on that page (Facebook Ads, Google Analytics 4, Marketo, Google Ads) totals $511.15 a month.
Read the table per connection. Fivetran’s optimization guide says “Your usage and costs are determined at the connection level,” and the consumption table adds that “you cannot combine MAR across multiple Connections, Activations, or Accounts to get a more favorable consumption rate.”
What moves the bill up or down?
Five things, all documented by Fivetran.
1. How MAR is spread across connections. Because each connection rides its own curve, the same total volume costs more when it is split. A synthetic example with round numbers, on Standard list prices: two connections with 2 million MAR each cost $670 apiece, $1,340 in total. One connection with 4 million MAR costs $1,010.
2. How many small connections you run. Since January 1, 2026, every standard connection with 1 to 1 million MAR in a month carries the $5 base charge (2026 pricing updates). Twelve connections at 5,000 MAR each cost about $90 a month on Standard, of which $60 is base charge. Fivetran’s own advice is to pause “connections with a low usage that don’t bring any analytical value” (optimization guide).
3. How often your source rows change. Sync frequency is irrelevant, per the documentation: “Your connections’ sync frequency makes no difference to your monthly active rows because MAR is based on how many unique rows are updated, not on how many updates occur.” A Shopify order that gets paid, fulfilled and partially refunded in the same month counts once. A customer table where a marketing tool touches every profile daily counts every profile.
4. Deletes and history mode. From January 1, 2026, “inserts, updates, and deletes will count toward paid MAR,” and in tables with history mode on, each tracked change counts too (2026 pricing updates). Deletes reach your warehouse as a soft delete, with _fivetran_deleted set to TRUE (usage-based pricing).
5. Duplicated pipelines. Two connections that read the same source into a staging and a production destination are counted separately, “even if the connections sync the same data from the same source (with the same primary keys)” (usage-based pricing).
Ad platforms add one more effect. The usage-based pricing documentation counts “reporting connector rollback syncs that occur during a conversion window” as incremental MAR, and says rows a rollback sync fetches from the previous month count as new rows in the current month. Facebook Ads, Google Ads, Google Analytics 4 and Amazon Ads are on its list of ad reporting connectors.
What does Fivetran not charge for?
The usage-based pricing documentation lists free MAR, including:
- Initial syncs. “Initial syncs don’t incur a cost for the historical data they sync.” Backfilling three years of Shopify orders is free; the changes after it are not.
- Most re-syncs, such as historical data pulled because you added a column.
- The first 14 days of every new connection.
- Fivetran system tables (
FIVETRAN_QUERY,FIVETRAN_API_CALLand others) and the Fivetran Platform connector, which loads your own usage data into the warehouse. - The Free plan, up to 500,000 connection MAR, 3,500 activation MAR and 5,000 model runs a month (billing and plans).
How do the plans and contracts differ?
| Monthly pay-as-you-go | Annual contract | |
|---|---|---|
| Plans available | Free, Standard, Enterprise, Business Critical | Standard, Enterprise, Business Critical |
| Minimum | $5 per connection with 1 to 1M MAR | $12,000 |
| Price | List price, billed monthly in arrears | 5% built-in discount, rising with contract size |
| Where to buy | Fivetran, AWS, Azure or Google Cloud marketplace | Order form with Fivetran |
Source: Fivetran billing and plans, read November 3, 2026.
The pricing page advertises savings of “up to 22%” on an annual contract. For Standard it publishes the discount bands: 5% up to $37,900 of annual list price, 5% to 22.6% between $37,900 and $145,340, and 22.6% or more above that.
All plans use the same meter and differ in features and rate. Enterprise adds 1-minute syncs, enterprise database connectors and hybrid deployment; Business Critical adds customer-managed encryption keys, PCI DSS Level 1 certification and private networking. The same 1 million MAR costs $500 on Standard, $666.67 on Enterprise and $1,066.67 on Business Critical at list price.
What would a Shopify brand’s bill look like?
A synthetic example with round numbers, not a quote. The MAR volumes below are invented for illustration; the prices come from the Standard row of the table above.
| Connection | Assumed MAR / month | Standard list price |
|---|---|---|
| Klaviyo | 3,000,000 | $840.00 |
| Shopify | 400,000 | $203.00 |
| Amazon Selling Partner | 250,000 | $128.75 |
| Recharge | 120,000 | $64.40 |
| Google Ads | 90,000 | $49.55 |
| Facebook Ads | 40,000 | $24.80 |
| Google Analytics 4 | 25,000 | $17.37 |
| TikTok Ads | 15,000 | $12.42 |
| Total | 3,940,000 | $1,340.29 |
In this example one high-churn source drives most of the bill. Klaviyo is 76% of the rows and 63% of the cost. Pooled into one curve, the same 3.94 million MAR would cost about $1,000, and the consumption table does not allow pooling.
To see your real numbers, Fivetran shows MAR per connection and per table under Account Settings > Billing & Usage, updated daily and calculated in UTC (optimization guide).
How can you lower a Fivetran bill?
The documented levers, from the optimization guide:
- Block schemas or tables you do not analyze. Not every connector supports blocking; check the connector’s page.
- Block frequently changing columns in tables without a primary key, where Fivetran builds a synthetic key from the columns.
- Pause low-value connections that only pay the $5 base.
- Consolidate where your architecture allows it, since volume discounts apply only inside one connection.
- Switch history mode tables to soft delete or live mode where you do not need every version of a row, an option named in the 2026 pricing updates.
Changing sync frequency will not help, by Fivetran’s own statement.
When is Fivetran the better buy?
When you have a data engineer, a warehouse you already pay for (BigQuery, Snowflake, Databricks) and dbt. Since June 1, 2026 Fivetran and dbt Labs are one company (merger release), so extraction and modeling come under one vendor, and the catalog lists Shopify, Amazon Selling Partner, Klaviyo, Facebook Ads and Google Ads (Fivetran connectors). It also fits a small stack: under 500,000 MAR a month across connections, the Free plan covers it.
The full comparison with Airbyte and managed options is in Fivetran alternatives for ecommerce data, and the scored list of nine tools is in best ETL tools for marketing data.
When is a managed platform the better fit?
When the cost you cannot predict is people. Fivetran moves data into your destination; modeling net revenue, monitoring loads and maintaining the warehouse remain your team’s work. If nobody on the team does that today, a per-row price is the smaller part of the cost, which is the argument in build or buy a data lake.
Precisian is one such platform: ingestion, a data lake isolated per client on our infrastructure, a semantic layer with versioned metric definitions, and access through MCP, API or your own BI. It bills a monthly subscription with no setup fee and no per-seat charge. The infrastructure is ours, the data is yours: you can access it at any time and, if you leave, you take all of it. It is the wrong choice if you already run a data team on dbt, if you want a ready-made dashboard (we don’t sell one), or if your company operates only in the US: Precisian serves companies in Brazil.
What does this article not cover?
Negotiated prices, Enterprise License Agreements (Fivetran lists them as “Fixed price per year with no limits on consumption” on the pricing page), the full activation price curve, contracts signed before March 1, 2025, which are billed at the account level, and the cost of the warehouse that receives the data. Fivetran says it may update the consumption table, and changes reach existing customers at renewal; check the version number at the top before you budget.
If your Fivetran estimate is mostly a few high-churn connections plus the people who model what lands in the warehouse, compare it against a subscription that includes both. What Precisian’s monthly fee covers is on the pricing page. No contract commitment, and no promise of an answer in 30 minutes either.