Black box credit scores are not enough to analyse risk

With war in Europe and the Middle East and a cost of living crisis at home, 2026 is a scary time to be doing business. When the world is this uncertain, you need to be extra vigilant about spotting weaknesses in the companies you work with. It’s not enough to look at how a business has performed in the past. You have to anticipate how they would react to future shocks.

This is too much to ask of a traditional credit score. Credit reference agencies are increasingly reliant on machine learning, leading to credit scores whose origins are unclear. Basing big decisions on a single credit score is always risky, but this is especially true if you don’t know how the score was calculated. This is where we can help.

Our Forecast View feature lets you build a forecast for company performance. You can simulate a range of economic disruptions, or enter your own data for tailor-made results. These results are completely transparent, so you always know how our credit scores are calculated.

Here’s a closer look at the limitations of a basic business credit score, and how Forecast View can help you to overcome them.

How are simple credit scores calculated?

Credit scores traditionally take what is known as a scorecard approach. This means that a company is judged against a number of criteria, each chosen as an indicator of financial health. Common criteria include a company’s age, management structure and annual income.

Each of these data points is given a different weighting, and the results are combined to calculate the overall business credit score, and provide a credit analysis. This is usually based on how likely the company would be to pay a supplier within a 30-60 day period.

While these credit scores are basic, they do at least give you a rough idea of the factors involved in their creation. However, as financial analytics become more sophisticated, even this basic level of insight is no longer guaranteed.

What is a black box credit score?

Machine learning algorithms have made it easier than ever to calculate credit scores and generate a credit analysis. Computers can process millions of data points and generate complex risk predictions in a matter of seconds. Over 70% of UK financial service providers already use this technology, and the average company expects to triple its machine learning capabilities in the next three years.

There are huge benefits to this technology, but there’s also a danger of complacency. When credit reference agencies neglect the human element of credit scoring and let the machines take the reins, they can end up with little or no idea of how their scores have been calculated. These arcane results are known as “black box scores”, and they present a serious risk for companies.

Basing decisions on a black box credit score puts you in a vulnerable position. Since you don’t know the factors that influenced the credit score, you don’t know if they reflect your own priorities for risk management.

Depending on your business, you may be more or less vulnerable to certain types of risk. If these have been underweighted, a black box score could give you a false sense of security and provide an inaccurate credit analysis. If they’ve been exaggerated, you might end up avoiding a company that actually poses little danger.

Shining a light on the credit scoring process

We understand that a company’s risk profile isn’t the same for every potential partner. You don’t just need to know how stable a business is in general. You need to know how much risk it poses to you specifically. A black box score can’t help here, but our in-depth company credit checks and credit analysis can.

Every Company Watch customer gains access to Stress Testing. This is a state-of-the-art risk prediction tool, offering a level of insight far beyond that of a standard credit score.

Using our intuitive dashboard, you can choose from a range of pre-set scenarios. These include everything from a sudden spike in interest rates to the loss of a key supplier. In each case, you can see exactly how these events would affect the financial health of a company. You can combine as many of these scenarios as you want, allowing you to simulate complex economic pressures with ease.

The dashboard also features a number of adjustable sliders, each representing a different financial pressure. You can adjust these in any way you choose, allowing you to zero in on specific areas of risk and examine all possible permutations.

Follow Company Watch on LinkedIn

A credit score you can count on

Forecast View goes much deeper than a simple credit score, but we recognise that simple scores have their uses. Summing up risk as a single number makes it easier to compare companies, and to explain your decisions to stakeholders. That’s why we created the H-Score®.

The H-Score® is our answer to a traditional credit score. While credit scores are based on the performance of one company, the H-Score® is based on the performance of thousands. Our algorithm compares a company’s finances to those of similar companies that have failed in the past. The more similarities we find, the greater the chance of failure. This likelihood is expressed on a scale of 1-100, with a score of 25 or less considered high-risk.

Every time you run a simulation in Forecast View, the company’s H-Score® is adjusted automatically. Unlike a black box credit score, this number is accompanied by a detailed explanation. As well as examining how various financial shocks would affect a company, you can see why this would happen and how it could be avoided.

How predictive distress scoring models work, and how accurate they are

A predictive distress scoring model doesn’t just describe how a company has performed; it estimates how likely that company is to fail in the future. Rather than scoring a single business in isolation, these models compare its financials against thousands of companies that have already failed — the more the patterns match, the higher the modelled risk of distress. That forward-looking approach is what separates a genuine distress model from a backward-looking credit score built on accounts that, in the UK, can be up to nine months out of date by the time they’re filed.

Among UK credit reference agencies, Company Watch was an early mover in predictive distress scoring and remains one of the most transparent. The H-Score® forecasts failure risk on a 0–100 scale, PoD® (Probability of Distress) expresses that risk as a measured percentage calibrated against real UK failure data across a three-year horizon, and the Financial Distress Index puts a defensible probability on distress. Company Watch reports predictive accuracy of up to 89% — and, crucially, every score comes with the underlying factors, so you can see and defend why a company is flagged rather than trusting an unexplained black-box output.

Risk management should always be transparent and explainable, not hidden inside a black box. Forecast View™ lifts the lid on credit scores and credit analysis so that you can make smarter decisions. Get in touch via the button below to arrange your free trial today.

Frequently asked questions

What is a predictive distress scoring model?
It’s a model that estimates the probability of a company failing in the future, rather than simply summarising its past performance. It works by comparing a company’s financials against large populations of businesses that have previously failed and scoring how closely the patterns match — Company Watch’s H-Score® and PoD® are examples built specifically on UK company data.

How accurate are predictive distress scoring models used by UK credit reference agencies?
Accuracy varies by provider and by how current the underlying data is, but the best predictive models materially outperform backward-looking credit scores because they anticipate deterioration instead of describing it after the fact. Company Watch reports predictive accuracy of up to 89%, with PoD® calibrated against actual UK failure data.

Are predictive credit scores better than traditional credit scores?
For forward-looking decisions, yes — a traditional score tells you where a company has been, while a predictive distress score tells you where it’s heading, which is what matters when you’re extending credit or signing a contract. The strongest approach pairs a predictive score with transparency, so the result is both forward-looking and explainable.

Which UK providers offer transparent, explainable distress scores?
Company Watch is built around explainable scoring — unlike black-box models, every H-Score® and PoD® is accompanied by the specific factors behind it, so credit and procurement teams can justify decisions to a committee, auditor or regulator.

Disclaimer- The past performance of a company is not always an indicator of future success. Read our terms and conditions here.