← Back to the Risk Index

Most credit scores are a black box. You put a company in, a number comes out, and the logic that produced it stays locked inside a bureau's algorithm. The Risk Index takes the opposite approach. This page sets out exactly what the tool does with your inputs, which data it draws on, and how it turns thousands of simulated outcomes into a single number from 1 to 99.

What the score measures

The Risk Index is a single number from 1 to 99 that summarises how exposed your firm is to trade credit losses, given the shape of your customer book and the economic conditions you are trading into. A higher number means more exposure. The score deliberately blends two views of the world:

• A typical year. What losses tend to look like when nothing unusual happens.
• A stressed year. What losses could look like in an unfavourable one-in-twenty scenario, the kind of year where the economy turns and several customers struggle at once.

A firm can look comfortable in a typical year and still carry serious exposure in a bad one. Holding both views in the same number is the point. It is what stops the score from flattering a customer book that is quietly fragile.

What you tell the tool

The score is built from four inputs. Each one maps to a specific lever in the model.

Your input	What it represents	How it enters the model
Annual revenue band	The rough size of your receivables, the money owed to you at any one time	Converted into an estimated receivables balance. This sets the scale of what is at risk. If you upload an aged debtors report instead, the exact figure is used.
Customer concentration	How much of your sales sit with your biggest customers	Drives tail risk. A book where one or two customers dominate carries a real chance that a single default becomes a catastrophic loss. A spread book absorbs the same shock far better.
Industry	The sector your customers trade in	Selects a risk profile, a baseline failure rate, how volatile that rate is, the typical payment behaviour, and how sharply the sector moves with the wider economy.
Payment terms	How long you give customers to pay	Longer terms mean money is exposed for longer and there is more time for a customer's position to deteriorate before you are paid. Measured against a 30 day baseline.

The live data the tool pulls in

Your inputs describe your book. The other half of the score is the economic weather you are trading into. The tool draws on eight public data sources. Several are fetched live in your browser at the moment you run the assessment. Others do not offer a live feed that a browser can reach, so those figures are calibrated from the most recent published release and refreshed as new data lands.

Source	What it contributes	Live or calibrated
World Bank	UK GDP growth and how volatile that growth has been	Live where reachable
Bank of England	The current base interest rate	Live where reachable
IMF	Forward growth forecast from the World Economic Outlook	Live where reachable
OECD	Composite Leading Indicator, an early read on whether the economy is expanding or contracting	Live where reachable
ECB	EUR/GBP exchange rate and recent currency volatility	Live where reachable
ONS	UK company insolvency rate and its recent trend	Calibrated from latest release
UK Insolvency Service	Company insolvency volumes and year on year change	Calibrated from latest release
Coface	UK country risk grade and sector risk ratings	Calibrated from latest release

If a live source cannot be reached, the tool does not fail or leave a gap. It falls back to the calibrated figure and carries on, so the score is always built on a complete picture.

How the economy adjusts the baseline

Before any simulation runs, the tool builds a single economic modifier. This dials the baseline risk up in a weakening economy and down in a strengthening one. Four signals feed it, each weighted by how strongly it tends to move trade credit losses:

The recent insolvency trend then nudges the baseline failure rate on top of this. The result is that the same customer book scores higher when rates are biting and growth is stalling, and lower when conditions ease. The model is anchored to the conditions in front of you, not a long run average that may no longer apply.

The simulation

This is the core of the tool. Rather than apply a fixed formula, the Risk Index runs 10,000 simulated years of your customer book and watches what happens across all of them. No single year tells you much. The shape of 10,000 of them tells you a great deal.

In each simulated year the model samples four things at random, drawn from distributions set by your inputs and the economic modifier:

• A shared economic shock. One draw that hits every customer at once. This is what models a genuine downturn, the systemic risk that your customers fail together rather than independently.
• A sector failure rate. How many customers run into trouble this year, scaled by how cyclical your industry is. A highly cyclical sector swings much harder when the shared shock is bad.
• A payment delay factor. How much your terms stretch out actual payment, driven by your terms against the 30 day baseline.
• A concentration outcome. Where one or two customers dominate the book, the model allows for the catastrophic case, a major customer defaulting and taking a large slice of your receivables with it. A spread book instead sees smaller, distributed losses.

Each simulated year produces a loss figure. After 10,000 of them, the tool has a full distribution of outcomes, from the quiet years to the genuinely bad ones.

From losses to a score

From that distribution the model reads off a few key numbers:

• The average loss across all 10,000 years. This is the typical year view.
• The 95th percentile loss, the level a bad one-in-twenty year reaches. This is the stress view, shown on your results page as the stress downside.
• The gap between the stress level and the middle of the distribution, which captures how heavy the tail is, in other words how much worse a bad year is than a normal one.

These combine into the headline number, then compress onto the 1 to 99 scale:

Weighting the stress year as heavily as the typical year is deliberate. A tool that only reported the average would tell you how things usually go, and usually is not the problem. The problem is the year that does not go to plan.

Reading the score

The 1 to 99 number lands in one of five bands. The bands govern how much attention the profile warrants and how to size limits and terms, not whether to protect the exposure.

Alongside the headline number, the results page breaks out three component readings, your receivables exposure, your concentration, and your payment terms, so you can see which part of your book is doing the most to move the score.

How industries are calibrated

Each industry carries its own risk profile, set by two main characteristics: a baseline failure rate, how likely customers in that sector are to run into trouble in a normal year, and economic sensitivity, how sharply that rate worsens when the economy turns. A sector can be steady in good times yet swing hard in bad ones, which is why both matter. The current calibration groups the sixteen sectors as follows:

Risk tier	Sectors	Profile
Higher	Construction, Hospitality, Commodities, Retail	Elevated baseline failure rates and strong sensitivity to economic conditions. Commodities in particular swings hard with global cycles.
Moderate	Energy, Logistics, Media, Wholesale and Distribution, Manufacturing, Agriculture	Mid range baseline rates with meaningful but not extreme cyclicality.
Lower	Recruitment, Finance, Tech and Software, Professional Services	Lower baseline rates and limited economic sensitivity. More resilient through a downturn.
Lowest	Healthcare, Cyber	Defensive sectors. Low baseline failure rates and very little movement with the economic cycle.

These groupings reflect UK insolvency patterns and Assured Trade's own experience of where losses cluster across sectors. They are calibrated estimates, refined over time, not fixed empirical constants. They set the starting point for the simulation, which is where your specific book and the live economic picture take over.

What the Risk Index deliberately does not do

Being open about the method means being just as open about its limits.

• It is not a credit rating or a probability of default. It is an indicative signal about exposure, designed to point to where structuring terms, sizing cover, or revisiting limits matters most.
• It does not assess any single named customer. It models the shape of your book as a whole. A view on a specific buyer is a separate exercise.
• It works from bands, not audited figures. Unless you upload an aged debtors report, the receivables scale is estimated from the band you select. The score is only ever as precise as what goes in.
• Published data carries a lag. Macro and insolvency figures reflect the most recent releases, which describe where things were, not where they will be next quarter.
• No model can see what has not happened yet. Even a Very Low score is a snapshot of visible risk. It cannot price in the customer who deteriorates between filings or the shock no one forecast.

Why protection is the baseline

This last point is the most important, and it is why the score is a starting point for a conversation rather than the end of one.

A good score means you can likely trade on competitive terms with sensible limits. A poor one means the structure needs more care, more security, or tighter limits. In both cases the protection sits underneath. That is the difference between a number that flatters you and a number that helps you trade with your eyes open.