BUYER GUIDE
Anomalo Pricing: Anomalo Cost, Anomalo Alternatives, and the Only Published Rate
Anomalo has never published a pricing page. We checked the live site, then checked more than 4,000 archived Anomalo URLs going back to the start of the domain, and there has never been one. There is exactly one published Anomalo rate anywhere on the internet, it sits on the AWS Marketplace listing, and the buyer notes on that listing admit the unit it is denominated in is undefined. This page shows you the rate, what it does and does not tell you, and what teams report paying.
14-day trial, no credit card, read-only connection
Alerted #data-eng 0.8s ago.
Downstream impact · consumers at risk
What does Anomalo cost in 2026?
Anomalo publishes no price on its own website. The one public rate is on its AWS Marketplace listing, checked September 10, 2026, which shows a single contract dimension named units at 1.00 US dollar per unit for a 12 month term, with 24 and 36 month terms also offered and no tiers or add-ons. Because the listing never says what one unit covers, that rate cannot be turned into a quote. Vendr, which aggregates real contracts, reports a median Anomalo buyer paying 115,000 dollars a year, in a range from 58,000 to 164,688 dollars.
Last updated September 2026
Side by side
Anomalo pricing compared
Swipe to see all columns →
| Vendor | Published figure | Source and date | What the meter counts |
|---|---|---|---|
| Anomalo | 1.00 USD per unit, 12 months | AWS Marketplace, Sept 2026 | Units, not defined on the listing |
| Anomalo | 115,000 USD a year, median | Vendr, Sept 2026 | Negotiated contract, meter not stated |
| Monte Carlo | 50,000 USD per 12 months | AWS Marketplace, Sept 2026 | Monitors, sold as credits |
| Sifflet | 48,000 USD per 12 months | AWS Marketplace, Sept 2026 | Assets, in tiers of 500 and 1,000 |
| Bigeye | From 45,000 USD a year | AWS Marketplace, Sept 2026 | Monitored tables |
| Elementary | From 120,000 USD per 12 months | AWS Marketplace, Sept 2026 | Units sized on seats and tables |
| Collibra | 170,000 USD per 12 months | AWS Marketplace, Sept 2026 | No unit published at all |
| Acceldata | 10,000 plus 5,000 per TB-month | AWS Marketplace, Aug 2026 | Average terabytes processed |
| IBM Databand | 450 USD a month, withdrawn | IBM archived page, Oct 2025 | Resource Units, data transferred |
| Soda | 750 USD a month, Team | soda.io, Sept 2026 | Soda Processing Units |
| Dataobservability | 99, 299 and 799 USD a month | Our pricing page, Sept 2026 | Flat monthly tier, no usage meter |
Positioning and pricing models are summarized in good faith from each vendor's public pages, September 2026. Verify current terms with the vendor.
What you get
What the published Anomalo numbers actually mean
The one published Anomalo rate is 1.00 dollar per unit per year, and it is unusable on its own
Open the Anomalo Data Quality listing on AWS Marketplace and scroll to Pricing. You get three contract lengths, 12, 24 and 36 months, and a table with one row in it. The dimension is called units. The cost per 12 months is 1.00 US dollar. That is the whole published price. AWS states on the same page that the listing uses a single contract-based pricing dimension billed in units, that you buy a quantity of units for a fixed contract term rather than paying per hour or by direct usage, and that because pricing is a single unit-based dimension there are no separate tiers or add-on options to choose from. A rate of one dollar per unit is a clean number, and it is also the least informative price structure in this category, because the entire cost of the contract lives in the quantity rather than the rate. Two buyers on identical rates can be a hundred thousand dollars apart and neither could tell from the listing.
The buyer notes on the AWS listing say the unit is not defined
This is the part worth quoting, because it is unusual to see a marketplace say it out loud. Under Top-of-mind questions for buyers, in answer to what one billing unit represents, the listing reads: the marketplace listing bills in units under a contract, but does not define what a single unit maps to in concrete terms. It then adds that Anomalo monitors data quality across tables and datasets, so a unit may relate to monitoring scope, and tells you to contact the vendor to confirm exactly how units are counted for your environment. May relate to is doing a lot of work in that sentence. Every other vendor in the comparison table above at least names its unit, even where it will not price it: Monte Carlo says monitors, Sifflet says assets, Bigeye says monitored tables, Acceldata says terabytes processed. Anomalo publishes the rate and withholds the unit, which is the mirror image of Collibra, which publishes a total and withholds the unit. Both leave you unable to compare, from opposite directions.
Anomalo has never had a pricing page, and that is verifiable rather than an impression
Fetch anomalo.com/pricing today and you get a WordPress 404 whose title is literally Page not found - Anomalo. That alone would only tell you the page is gone, so we went further and pulled the Wayback Machine index for the whole domain: more than 4,000 distinct archived Anomalo URLs, spanning the life of the site, and not one of them is a /pricing, /plans or /cost path. Compare that with IBM, which published Databand prices for about eighteen months before withdrawing them, and the archive shows exactly when. Anomalo did not withdraw a price. It has never had one. That is a deliberate and entirely legitimate go-to-market choice, and it is also a fact you should factor into your evaluation timeline, because it means there is no historical figure to anchor a negotiation on and no version of this product you could ever have bought without a sales cycle.
What teams actually pay, according to the only aggregator with contract data
Vendr is a software buying platform that sees real signed contracts rather than scraped marketing pages, which makes it the most useful third-party source here. Its Anomalo page reports a median buyer paying 115,000 US dollars a year, with observed prices running from 58,000 to 164,688 dollars. Read that carefully. Vendr does not publish how many contracts sit behind the median, and a median without a sample size is a weaker claim than it looks, so treat it as an order of magnitude rather than a benchmark. What it is good for is calibration. If you are a fifty person company being quoted 40,000 dollars, you are below anything Vendr has seen and should ask what is being excluded. If you are being quoted 300,000, you are well above the top of the observed range and should ask what is being included. The 2.8x spread between the low and high figures is itself the finding: this is a negotiated product, and the scope of the negotiation is wide.
One dollar per unit and a 115,000 dollar median imply a very large unit count
Put the two published numbers next to each other and you can do one piece of arithmetic, as long as you label it as an inference across two independent sources rather than a fact either source states. If the AWS Marketplace rate of 1.00 dollar per unit per 12 months is the rate a typical buyer pays, and the Vendr median of 115,000 dollars a year is a typical contract, then a typical Anomalo contract is on the order of 115,000 units a year. That number is only useful for one purpose, which is to tell you what question to ask in the first sales call. Nobody buys 115,000 of something without knowing what it is. Ask for the unit definition in writing, ask what consumes a unit, ask whether the count is per table, per check, per run, per row scanned or per column profiled, and ask what happens when you exceed the committed quantity mid-term. If the answer is that units are sized by the vendor from your environment, you have a volume meter with a friendly rate, and your bill will move for reasons your team did not choose.
Two marketplaces, one number between them
Anomalo has put real effort into cloud marketplace distribution. There is the AWS Marketplace listing described above, and in September 2024 Anomalo announced a fully containerized Snowflake Native App on Snowflake Marketplace, running inside Snowpark Container Services, together with Snowflake Premier Partner status and eligibility for the Marketplace Capacity Drawdown program. That last detail matters to a buyer: if your company already has a Snowflake capacity commitment, the purchase can draw it down instead of opening a new budget line. On the AWS side, US buyers can also pay through the PNC Vendor Finance line of credit program, which the listing notes is not available in Nevada, North Carolina, North Dakota, Tennessee or Vermont. What two marketplaces do not give you is two data points. Nothing in the Snowflake app announcement published a price, and the AWS listing shows a rate for a unit it never defines. Distribution has been solved. Price transparency has not.
How it works
From connected to caught
Ask for the unit definition before you ask for a quote
Reverse the usual order. Most evaluations open with scope and get a number at the end, which means you receive a figure you cannot decompose. Instead, in the first call, ask what one unit is, what actions consume units, and how many units the vendor estimates for your environment. You now have a rate you already know, a definition, and a quantity, and the quote becomes something you can check rather than something you have to accept. If the unit definition is not forthcoming in writing, that is a legitimate finding to record in your evaluation notes alongside the feature comparison.
Price the same scope with a vendor that publishes a number
The fastest way to know whether a six figure quote is reasonable is to have a working, priced comparison on the same tables. Connect a platform whose price is on its website, read only, to the warehouse you were going to put in scope, run it for two weeks against the same tables, and write down what it caught. You end the evaluation with a real baseline and a defensible cost per table rather than a vendor estimate. Dataobservability publishes 99, 299 and 799 dollars a month with a 14 day trial and no card required, which is enough to run this comparison without a purchase order.
Convert every quote on your shortlist to a rate times a quantity
Enterprise observability quotes are hard to compare because each vendor counts a different thing, so annual totals are not denominated in the same currency. Build a three column sheet: vendor, unit, rate. Monte Carlo goes in as monitors sold through credits with no published per-credit rate. Sifflet goes in as assets in bands of 500 and 1,000. Bigeye goes in as monitored tables. Anomalo goes in as units at 1.00 dollar with the definition you obtained in step one. Where a cell stays empty, that empty cell is a decision input, not a gap in your research.
Decide the meter you can live with, then negotiate inside it
Coverage meters and volume meters fail differently and you should choose on purpose. A coverage meter, counting tables or monitors, makes your bill a direct function of a decision your team makes deliberately, which is predictable but taxes thoroughness: the neglected corner of the warehouse where incidents actually start is the part you are paying extra to instrument. A volume meter charges for how much data moves, which makes adding checks cheap but makes your bill move when a backfill runs or the business grows. A flat tier does neither and simply caps the surprise. Pick the failure mode you can defend in a budget meeting, then negotiate the quantity.
Why a vendor with no published price still sells through cloud marketplaces
It looks contradictory and it is not. Cloud marketplace listings solve a procurement problem rather than a pricing problem. When a US enterprise has a committed spend agreement with AWS, Azure or Google, software bought through that cloud marketplace draws down the commitment, which turns a new line item into money the company has already promised to spend. That is a powerful reason for a finance team to say yes, and it has nothing to do with whether the vendor publishes a rate. What a marketplace listing does require is that something appear in the pricing table, and the common solution among sales-led vendors is a private offer, where the public listing carries a placeholder and the real terms arrive through a negotiated offer visible only to the buyer. A single dimension called units at 1.00 dollar is exactly that shape: a valid public listing that transmits no information. Read it as a procurement rail, not as a price. The practical consequence for a buyer is that marketplace purchase is worth asking about early, because it can materially change how the spend is approved internally even when it does not change the number.
What Anomalo is actually selling in 2026, and why the price scales the way it does
It is worth understanding the product to understand the meter. Anomalo started as unsupervised machine learning applied to wide tables: point it at a table, and it profiles every column and learns what normal looks like without anyone writing rules. That is still the engine and it is genuinely good at what a rules-first tool cannot do, which is find the problem in the column nobody thought to check. By 2026 the product presents as a broader autonomous data platform organized around named agents covering table observability, data quality defined in natural language, insights, documentation and lineage, extending past tables into documents and unstructured data, and deployable inside your own VPC. That breadth explains the six figure median. You are not buying a freshness checker, you are buying a platform, and platform contracts are scoped by how much of the platform you turn on. It also explains why the unit stays undefined: a meter that has to cover column profiling on a 300 column table, document quality, and lineage generation is not naturally expressed in tables or monitors. The honest read is that Anomalo is priced like a platform because it is one, and that a team whose actual requirement is freshness, volume, schema and distribution monitoring on four warehouses is buying a lot of surface area it will not use.
When Anomalo is the right purchase despite the opacity
We compete with Anomalo and we are still going to say plainly where it wins, because a comparison page that never concedes anything is not useful to anyone making a decision. Three cases point clearly at Anomalo. First, very wide tables with no known failure modes. If you have a 300 column table and no idea which columns break, unsupervised profiling of every column will find things a rules-based system never will, because nobody was ever going to write 300 rules. Second, deployment inside your own cloud account. A bank, an insurer or a health system that cannot let table contents leave its VPC has a short vendor list and Anomalo is on it, while most SaaS-only tools are not. Third, unstructured data. If your quality problem is documents feeding a retrieval pipeline rather than rows feeding a dashboard, that is a coverage area most warehouse monitoring tools, ours included, do not touch. If one of those three is your primary requirement, the sales cycle and the opaque unit are a cost of doing business and the demo is worth booking. What none of the three describe is the much more common case: a lean data team that needs freshness, volume and schema monitoring on Snowflake or BigQuery running this quarter, with a number it can put in a budget line before anyone has a call.
How to read a category where the published numbers share no unit
Take a step back from Anomalo and look at the table above, because the pattern is the point. Of the vendors with any public figure at all, the figures are not comparable: 50,000 dollars of Monte Carlo credits, 48,000 dollars of Sifflet assets, 45,000 dollars of Bigeye tables and 170,000 dollars of Collibra with no unit at all are four numbers that cannot be divided by anything to reach a common denominator. Monte Carlo and Sifflet sit 4 percent apart on paper and could easily be twice as far apart in practice for the same warehouse, because one counts monitors and the other counts assets and a team can have very different quantities of each. IBM briefly did the opposite, publishing both a price and a quantity on the same page, and then withdrew both. Anomalo publishes the rate without the quantity definition. None of this is misconduct, it is what happens when a category prices by negotiation, but it does mean the comparison work has to be done by the buyer and it cannot be done from public pages. The only structural escape is a vendor whose price does not depend on a meter you have to reverse engineer, which is why a flat monthly tier is worth considering even when the feature list is narrower.
Questions buyers ask
Anomalo pricing FAQ
What is the only published Anomalo price?
A single contract dimension named units, priced at 1.00 US dollar per unit for a 12 month term, on the Anomalo Data Quality listing on AWS Marketplace, checked September 10, 2026. The listing also offers 24 and 36 month terms and states there are no tiers or add-on options. Anomalo publishes no price on its own website.
Why does anomalo.com have no pricing page?
Because Anomalo sells through a sales-led motion and never built one. anomalo.com/pricing returns a 404 titled Page not found, and an index of more than 4,000 archived Anomalo URLs in the Wayback Machine contains no /pricing, /plans or /cost path at any point in the history of the site. Unlike IBM Databand, which published prices and later removed them, there is no earlier figure to recover.
What does one Anomalo unit cover?
Nobody publishes an answer. The AWS Marketplace listing states in its own buyer notes that it bills in units under a contract but does not define what a single unit maps to in concrete terms, and suggests only that a unit may relate to monitoring scope. Ask for the definition in writing before accepting a quote, including what actions consume a unit and what happens when you exceed the committed quantity.
What do companies actually pay for Anomalo a year?
Vendr, which aggregates signed software contracts, reports a median Anomalo buyer paying 115,000 US dollars a year with observed prices between 58,000 and 164,688 dollars. Vendr does not disclose the sample size behind that median, so treat it as an order of magnitude for calibration rather than a benchmark. The 2.8x spread reflects how much scope varies between contracts.
Is Anomalo priced per table?
Several review aggregators say so, and Anomalo does monitor at table level, but no first-party source confirms per-table pricing. The AWS Marketplace listing calls the unit units and explicitly declines to map it to tables, saying only that it may relate to monitoring scope. If per-table pricing matters to your model, get it stated in the contract rather than inferred from a review site.
Can I buy Anomalo through AWS or Snowflake Marketplace?
Yes. Anomalo sells through an AWS Marketplace listing with 12, 24 and 36 month contracts, and since September 2024 through a Snowflake Native App on Snowflake Marketplace that is eligible for the Marketplace Capacity Drawdown program, so an existing Snowflake commitment can fund it. The AWS listing also offers PNC Vendor Finance line of credit payment for US buyers outside Nevada, North Carolina, North Dakota, Tennessee and Vermont.
More of the platform
Catch broken data before your stakeholders do
Connect your warehouse and get anomalo pricing live from one read-only connection. Transparent pricing, no credit card.