Fivetran's MAR formula is simple to state and hard to estimate blind. Here's the math, worked through with Fivetran's own published numbers, and what it looks like against a flat price.
No credit card. 14 days. Cancel in one click.
Enter rough numbers for what you'd connect. Nothing here is sent anywhere; it runs in your browser.
Counting rule: a row counts once per calendar month no matter how many times it changes that month, so this adds distinct-changed rows to distinct-new rows rather than multiplying by number of changes. Dollar figures are illustrative, built from Fivetran's own published example rates, not a quote.
Quick answer: Fivetran's Monthly Active Rows are distinct primary keys touched by an insert, update, or delete in a calendar month, counted once per row. Estimate yours as total rows times your rough monthly change rate, then apply a per-MAR rate (Fivetran publishes examples from about $0.0005 to $0.0007 per MAR depending on the connector) plus a $5 base charge per connection. QueryFlow Pipelines skips the calculation entirely at $29.99/month flat, regardless of row volume.
Fivetran's documentation defines a Monthly Active Row plainly: "the number of distinct rows synced from the source system to your destination system in a given calendar month," determined by tracking distinct primary keys, counted separately per account, destination, connection, and table. A row counts once it's inserted, updated, or deleted, and Fivetran counts each distinct row exactly once per month, regardless of how many times it changes again inside that same month. There's no keyword tool or published multiplier for "how many MAR does my table generate"; the only accurate answer comes from knowing how many of your table's distinct rows actually change in a typical month, which Fivetran's own usage dashboard shows once you're connected.
Say you're evaluating a connector for a customers table with 2,000,000 total rows, where your application logic suggests roughly 8% of rows (new signups plus profile edits) see some kind of change in a typical month.
-- Illustrative estimate only; Fivetran's dashboard is the source of truth once connected total_rows = 2,000,000 pct_changing = 0.08 estimated_mar = total_rows * pct_changing -- 160,000 MAR -- Against Fivetran's published Standard-tier example rates (pricing page): -- Google Analytics 4: 21,993 MAR -> $15.88/mo (~$0.00072/MAR) -- Google Ads: 88,240 MAR -> $48.67/mo (~$0.00055/MAR) -- using the higher-volume Google Ads rate as a rough proxy: estimated_monthly = 5.00 + (estimated_mar * 0.00055) -- ~$93/month, plus the $5 base charge
That $93/month figure is an illustration built from Fivetran's own published example rates on different connectors, not a quote for your specific connector, which will have its own rate. The point of the exercise isn't the exact number; it's the shape of the math: a static, rarely-changing table generates almost no MAR no matter how large it is, and a small table with heavy write activity can generate more MAR than a much bigger, quieter one.
This kind of back-of-envelope math is only as good as your guess at the change percentage, and that guess is often wrong in both directions. Tables with frequent no-op updates (an updated_at timestamp touched by an unrelated process, or a full-row re-save that doesn't actually change any values) can inflate a naive estimate, since Fivetran counts the row as active on any update touch, not just a meaningful one. Conversely, tables where most of the "activity" is actually new inserts rather than updates to existing rows will have a MAR count close to net-new row growth, which is easier to estimate from a growth rate than a change rate. Before making a purchasing decision, Fivetran's own trial and usage dashboard, not an outside calculator, gives you the real number for your actual data.
QueryFlow doesn't ask any of these questions, because Pipelines is $29.99/month or $199.99/year regardless of row volume. That's not automatically the better deal: a team with genuinely low, predictable data volume might land well under $29.99/month on Fivetran's own free or entry-level usage, and a team that needs true managed streaming CDC with vendor support gets real value for MAR pricing that a flat license doesn't replace. Where a flat price is a clear win is a team whose row counts are large, volatile, or hard to predict, since the QueryFlow number never moves regardless of what a table does next month, and there's no dashboard to check before finding out what the bill will be.
| Monthly change volume | Illustrative Fivetran cost* | QueryFlow Pipelines |
|---|---|---|
| ~22,000 MAR (Fivetran's own GA4 example) | $15.88/mo | $29.99/mo flat |
| ~88,000 MAR (Fivetran's own Google Ads example) | $48.67/mo | $29.99/mo flat |
| ~160,000 MAR (worked estimate above) | ~$93/mo (estimated) | $29.99/mo flat |
| ~850,000 MAR (Fivetran's own Marketo example) | $424.54/mo | $29.99/mo flat |
*Figures marked with an asterisk are Fivetran's own published per-connector examples from its pricing page, or an illustrative estimate built from those published rates, not a quote for a specific connector you might run. Actual Fivetran pricing depends on your connector, tier, and contract terms; check Fivetran's own pricing page and usage dashboard for a figure you can rely on.
The math changes shape when you're evaluating several smaller sources rather than one big table. Say a team is considering five separate Fivetran connectors (a marketing tool, a support tool, and three internal databases), each generating roughly 15,000 to 40,000 MAR a month individually, similar in scale to Fivetran's own published Google Analytics 4 and Facebook Ads examples.
-- Illustrative only, using Fivetran's own published per-connector examples as reference points connector_1: ~22,000 MAR -> ~$16/mo (Fivetran's own GA4 example) connector_2: ~34,000 MAR -> ~$22/mo (Fivetran's own Facebook Ads example) connector_3: ~15,000 MAR -> base charge dominates at low volume connector_4: ~28,000 MAR -> similar band to connector_1 connector_5: ~18,000 MAR -> similar band to connector_3 -- five separate $5 base charges plus five separate usage charges, -- versus one flat QueryFlow Pipelines subscription covering all five syncs
Even at modest individual usage, five separate connections each carry their own $5 base charge on top of their own per-MAR usage, a detail worth including in any estimate since it's easy to model connectors individually and forget the base charge multiplies with connector count, not with total row volume.
This kind of calculator is weakest exactly where the stakes are highest: a table you don't yet have real change-rate data for. A newly-launched product feature can generate far more write activity in its first month than steady-state, which means an estimate built from "normal" traffic can understate MAR during a launch period specifically. The safer approach for a genuinely new source is a short paid trial period on Fivetran itself (or QueryFlow's own 14-day free trial, which has no metering to worry about either way) to observe real usage before committing to a long-term number either way.
Fivetran's published annual discount, starting at 5% and reaching up to 22% according to its pricing page, only pays off if your MAR usage is stable enough that a year-long commitment doesn't lock in a rate against volume that later shrinks. QueryFlow's own annual pricing, $99/year for Studio against $9.99/month, or $199.99/year for Pipelines against $29.99/month, works out to roughly a 17% discount for committing annually, a comparable shape of incentive without any usage variable attached to it.
Every dollar figure in the worked examples above is deliberately labeled illustrative rather than exact, because Fivetran's actual per-MAR rate varies by connector and by where your volume sits within its usage bands, information only Fivetran's own quote or usage dashboard can give you precisely. Treat this page as a way to understand the shape of the pricing model, not as a substitute for getting an actual number from Fivetran directly before making a purchasing decision either way.
Fivetran's own MAR definition is exact (distinct primary keys changed per calendar month), but converting that into a dollar figure ahead of time requires knowing your actual change rate, which only Fivetran's usage dashboard shows accurately once connected.
No. MAR counts distinct rows, not bytes or column count, per Fivetran's stated definition.
Fivetran counts MAR per connection and per destination, so yes, the same source table synced twice through two separate connections generates MAR usage on each one independently.
Fivetran's free plan includes up to 500,000 MAR for connections at no charge, and its usage dashboard is available once you've connected a source, which is the most direct way to see a real number before committing to a paid tier.
Because Pipelines has one price, $29.99/month or $199.99/year, with no per-row usage component at all; there's nothing to estimate.
14-day free trial, no card. $29.99/month, whatever the row count.
No credit card. 14 days. Cancel in one click.