Alert!
The hidden cost
of poor
data quality
What 150 AML data resellers told us about the gap between the data they buy and the data they ship.
6 quality dimensions
Every figure sourced
The cost of bad data, in four numbers
Every figure is attributed to a named source and listed in full in the appendix. Figures marked as AML Watcher data are our own, and labelled as such.
is the average yearly cost of poor data quality to an organisation
Gartner, 2021 (general benchmark)
of rule-based screening alerts are false positives, many driven by dirty data
PwC (industry benchmark)
is laundered each year; coverage gaps are where much of it slips through
UNODC
fewer false positives when clean data feeds true-risk scoring
AML Watcher data
What’s inside
Six parts on the data-quality gap
Built to be read start to finish, or opened at the part that matches where your pipeline is leaking today.
Where Data Decays
How sanctions, PEP and adverse media records go stale, duplicate or drop out between source and shipment.
The Reseller’s Dilemma
You are accountable to your clients for data you did not create and often cannot see inside.
Six Dimensions of Quality
Accuracy, completeness, freshness, consistency, coverage and structure, defined and made measurable.
What Poor Quality Costs
Rapid onboarding with pre-trained analysts and ready-to-deploy AI models
Benchmarks & Frameworks
A data-quality scorecard and a five-stage maturity model to place your pipeline honestly.
Fixing the Pipeline
Sourcing closer to origin, validating before shipment, and a 90-day plan to close the gap.
What’s inside
Six dimensions of AML data quality
The report scores data on six dimensions. A record can be accurate and still be worthless if it shipped a week late.
Accuracy
Names, dates and identifiers match the true entity, without transcription or transliteration errors.
Completeness
Aliases, RCAs, dates of birth and secondary identifiers are present, not just a primary name.
Freshness
Listings and delistings reach the client fast enough to matter, not on the next batch.
Consistency
The same entity is defined the same way across every list it appears on, with no clashing formats.
Coverage
The regimes, watchlists and jurisdictions the client is exposed to are all actually represented.
Structure
Fields are clean, typed and machine-ready, so screening engines can use them without guesswork.
Who feels it
Five people who inherit poor data quality
The gap is not abstract. For each person along the chain, it shows up as a specific, recurring cost.
Data Reseller / Distributor
Wears the margin hit when a client churns over a match that failed upstream.
Head of Data Ops
Owns the pipeline from source to ship, and every cleaning step bolted on to make it usable.
Quality / Compliance Lead
Signs off on what goes out the door, often without visibility into how it was sourced.
Customer Success
Fields every false-positive complaint and delisting query the source data caused.
The Client’s MLRO
Inherits whatever gaps shipped, and is personally accountable for what the program misses.
Key statistics
Data-quality statistics for 2026
A blend of survey findings, named third-party research and AML Watcher’s own coverage data. Free to cite with attribution.
average annual cost of poor data quality per organisation
Gartner, 2021
of rule-based screening alerts are false positives
PwC (industry benchmark)
laundered globally each year, 2–5% of world GDP
UNODC
PEP profiles with RCAs and known aliases
AML Watcher data
fewer false positives when clean data feeds true-risk scoring
AML Watcher data
official watchlists unified into one definition
AML Watcher data
adverse media sources across 415+ categories
AML Watcher data
interval at which sanctions data is refreshed
AML Watcher data
languages for name and media matching
AML Watcher data
Dashed cards are survey placeholders, to be populated from the finalised dataset. Where research firms disagree, the report publishes the range rather than the number that flatters the argument, and figures that could not be traced to a primary source were left out.
How it was made
A report about data quality has to hold itself to that standard
A blend of survey findings, named third-party research and AML Watcher’s own coverage data. Free to cite with attribution.
150 verified resellers
Every respondent was a confirmed AML data reseller or distributor. Sample size is stated on every survey figure.
Named sources
Every third-party claim is attributed to a named source, cited in the text and listed in full at the end.
Ranges, not cherry-picks
Where estimates disagree, on the scale of laundering for example, we publish the range, not one number.
Company data labelled
Figures describing our own database are marked as AML Watcher data, not dressed up as neutral research.
Empower your AML compliance!
Publications
Publications are not catching-up but knowing what’s ahead! So take the leap into the future!
Country Guides
From the Andes Mountain Ranges of Argentina to the Palm Beaches of Australia – We cover every inch.
Compliance Co-pilot
We offer a rich collection of free AML compliance AI models, tailored to empower everyone.
Beyond the Name: Name Screening Challenges in the MENA Region
Investigation of AML Cases: Manual vs. Automation (AMLW x SumSub)