Comparison

Cash flow underwriting vs. traditional credit scores

Traditional credit scores summarize reported credit history. Cash flow underwriting assesses current financial behavior from customer-permissioned account data. Carrington Labs can combine both with application and internal performance data in one lender-specific model.

Four approaches

Traditional FICO and bureau scores

Provide broad-market credit risk scores based primarily on reported credit-file information.

They are standardized inputs that lenders can use within existing policies and workflows.

UltraFICO

UltraFICO combines FICO credit information with customer-permissioned cash flow data from Plaid and uses the FICO score range. FICO made the updated product generally available in May 2026.

Standardized cash flow scores

Standardized cash flow scores convert transaction data into a consistent risk measure. Carrington Labs offers Cashflow Score as this lighter-weight option.

Lender-specific Credit Risk Models

Carrington Labs can build a model around the lender's own product, population and observed outcomes, using bureau, transaction, application and internal data in an appropriate combination.

At a glance

Dimension

Traditional score

Standardized cash flow score

Lender-specific model

Primary data

Reported credit history

Account and transaction data

Any relevant combination

Development population

Broad market

Broad or consortium population

Lender's own portfolio

Current financial behavior

Limited

Stronger visibility

Can be incorporated

Product specificity

Limited

Limited

High

Implementation speed

Fast

Usually fast

Requires development

Calibration to lender

Policy overlays

Score testing

Built, calibrated and retrained for lender

Offer optimization

No

No

Can connect to Credit Offer Engine

The core difference

The advantage of Carrington Labs is not cash flow data alone. It is the ability to test, engineer and combine cash flow, bureau, application and internal performance data against the lender's own repayment outcomes.

For lenders with sufficient performance data, the strongest approach is often a lender-specific model that tests and weights each signal against the product's actual repayment outcomes.

Why Carrington Labs goes beyond bureau-only underwriting

Bureau data remains useful, but it is only one source of risk information.

Bureau data captures reported credit history. Cash flow data adds information about current liquidity, income patterns, obligations, volatility and resilience.

Carrington Labs can combine bureau data with cash flow, application, product and internal repayment data, then test which engineered features add predictive value for the lender's own portfolio.

Underwriting with bureau scores today?

You do not have to pick a side. A combined model can be benchmarked directly against your current bureau-only approach on your own outcomes — keep whichever wins.

Carrington Labs vs. a new cash flow FICO score

UltraFICO combines FICO credit information with Plaid cash flow data in a standardized score.

Carrington Labs differs by offering:

A standardized transaction-only Cashflow Score
for fast deployment.
Custom models
built around lender-specific outcomes and commercial objectives.
Flexibility
to use the lender's chosen bureau and data sources.
Client-specific
custom-model data boundaries.
Advanced feature engineering
across transaction, bureau, application and internal performance data.
Included retraining
as portfolio data develops; the ability to engineer additional features when new data supports them; and a Credit Offer Engine for amount, term and price.

Carrington Labs is built for lenders that want portfolio-specific feature engineering and calibration, the ability to combine cash flow with traditional data, included retraining, client-specific custom-model data boundaries and a broader offer-optimization strategy.

Why a custom model may add value even when a score is strong

A broad-market score may provide weaker separation in the decision areas that matter most to an individual lender, including:

Within a lender's cutoff band
For a new product
For a specialist borrower population
Across different loan amounts and terms
Among repeat customers with internal performance history.

A custom model can be tested specifically within those decision areas.

More comparisons

See what bureau-only underwriting misses on your portfolio

Compare bureau-only, cash flow and combined approaches against your own portfolio outcomes before changing policy.