Triaster Ltd v Dun & Bradstreet Ltd

Triaster Ltd (“the Claimant”) is a company providing business process solutions. Specifically, it licenses its software and sells related support services to companies, public authorities and other organisations. Dun & Bradstreet Ltd (“the Defendant”) is a company operating in the business information sector. It sells data analytics services, credit analyses, credit reports, sales and marketing data packages, and other services to its clients. The focus of these proceedings for defamation is a credit report published by the Defendant to Specsavers in late September 2017. Owing to what the Claimant says are untruths in that report, it lost a potentially valuable contract with Specsavers, which had sought a credit check on the Claimant as entities in this sort of commercial situation often do.

The credit report in issue allocated the Claimant a “D&B rating” of D3. “D” is a reference to the Claimant’s “financial strength” based on a tangible net worth of £86,930. Viewed in context, it is clear that this figure was derived from the Claimant’s latest accounts publicly filed on 31^st March 2016. The “3” is a risk indicator which is said to “represent a greater than average risk of business failure”.

The D3 rating was, I should add, part of the Defendant’s initial “risk evaluation” as well as forming an important constituent of the “D&B risk assessment”. Another aspect of the risk assessment is that the “D&B maximum credit” figure was £15,000. This was the “recommended credit exposure at any given time” and was calculated “using the D&B failure score, business size and primary industry sector”.

Peeling away the jargon, the “D&B failure score” was – see page 2 of the credit report – 37 out of 100. By way of further explanation, and here I am looking at page 2 and nothing else, “D&B analytics have shown that businesses with this score have a moderate probability of failure. 22% of UK businesses are classified in a higher risk category”. Further, page 2 gave an “incidence of failure” metric of 1.79%. As the document explains, “the incidence of failure above shows the percentage of businesses with this failure score that have been subject to an insolvency event or gone out of business with outstanding debt”. In other words, a failure score of 37 out of 100 means that 1.79% of businesses with that score become insolvent or cease trading.

The report also provides a pictorial representation of the D&B failure score which it is not easy to replicate verbally. On the spectrum from “worst” to “best”, the former is in red and the latter is in green with yellow in the middle. As one might expect, because we are dealing with continuous and not discrete variables, the colours merge into one another, and the area between red and yellow appears as amber. On close examination, the Claimant’s score was in the amber zone, although I confess that I expressed a different view during the hearing. It may also be seen that a D&B failure score of 1-10, with a risk indicator of 4, does not precisely correspond with the red zone.

Page 2 of the credit report goes on to provide:

“The Scores and Ratings included in this report are designed as a tool to help credit professionals make their own credit related decisions, and should be used as part of a balanced and complete assessment relying on the knowledge and expertise of the reader, and where appropriate on other information sources.

The Score and Rating models are developed using statistical analysis in order to generate a prediction of future events. D&B monitors the performance of thousands of businesses for at least 12 months in order to identify characteristics common to specific business events. These characteristics are weighted by significance to form rules within our models that identify other businesses with similar characteristics and provide a Score and Rating.

Please note: (1) like all forward looking predictions, our Scores and Ratings are not a statement of what will happen, but an indication of what is more likely to happen based on previous experience; and (2) use of the term “insolvency event” in this report means (i) if a business has convened a meeting of its creditors, made a voluntary arrangement or proposal for any other composition scheme or arrangement with (or assignment for the benefit of) its creditors, (ii) if a business is unable to pay its debts, (iii) if a trustee receiver, administrative receiver or similar officer is appointed in respect of all or any material part of the business or its assets; (iv) if a meeting is convened for the purpose of considering a resolution, or other steps are taken for the winding up off the business (otherwise than for the purpose of an amalgamation or reconstruction) or for the making of an administration order or other appointment of an administrator in respect of the business, or any such order or appointment is made or effective resolution is passed to wind up the business.

Whilst D&B uses extensive procedures to maintain the quality of the information we hold, we cannot guarantee that it is always accurate, complete or up to date, and this may affect the Scores and Ratings we publish.”

The credit report gives further information about the Claimant which it is unnecessary for me to address specifically.

It is common ground that the credit report must be read in conjunction with the

The guide makes clear that this issue is one of risk, a concept which commercial individuals well understand; and the Defendant’s role is to provide some insight into that risk, assessed on an objective and consistent basis.

It is apparent from the guide that the “D&B failure score” is the probability that an organisation will obtain legal relief from its creditors or cease operations in the next 12 months. To that extent the incidence of failure statistic of 1.79% given in the credit report itself must be qualified. This score is a relative and not an absolute measure of risk. If the failure score is in the range of 11-50 (risk indicator 3) it is “higher than average”. As the guide also makes clear, the probability of failure (i.e. ceasing operations or obtaining legal relief) is associated with the failure score but is not synonymous with it.

In addition to these verbal explanations, the guide also provides what I have previously described as a pictorial representation of the D&B failure score by way of illustration. In order to make up for deficiencies in verbal description, the Appendix to this judgment contains the D&B failure scores as set out illustratively in the report (in relation to the Claimant), and then as set out in the guide. It may be seen that the spread of colour is slightly different in the guide, although I do not think that anything can turn on this.

Two other passages in the guide are relevant:

“Behind each Failure Score is an associated probability of failure, which rises rapidly at the low end of the Failure Score range.

The probability of failure allows our customers to set cut-offs for decisions based on their own credit policy and attitude to risk. It can be used to show the expected level of ‘bad’ applications/accounts for each Failure Score and therefore allow our customers to balance the opportunity of increased sales against the risk of bad debt.

…

The D&B Rating provides an indication of creditworthiness. The rating is made up of two parts:

-

Financial strength – Based on Tangible Net Worth from the latest financial accounts.

-

Risk Indicator – derived from the D&B Failure Score.”

According to paragraph 7 of the amended Particulars of Claim, the financial and ordinary meaning of the credit report was that “the Claimant was a serious credit risk and its creditworthiness was poor” and “its financial and trading position was such that there was a real risk it would fail owing creditors money within the next 12 months”.

Paragraph 8 of the amended Particulars of Claim pleads an innuendo meaning on the basis of various “glowing” self-appraisals set out on the Defendant’s website. These proclaim the quality and accuracy of the Defendant’s analytical services. However, it

The four preliminary issues which arises for my determination are as follows:

The legal principles governing items 1 and 2 above are relatively straightforward, and I may take these directly from the Claimant’s Skeleton Argument.

The natural and ordinary meaning is the meaning which the statement would convey to the hypothetical ordinary reasonable reader, applying the “single meaning rule”. The meaning captures both the subject matter of the imputation and the degree of seriousness of any defamatory allegation.

The principles to be applied to the determination of meaning were summarised by Nicklin J in Koutsogiannis v The Random House Group Limited [2019] EWHC 48 (QB) at para 12:

“(i)

The governing principle is reasonableness.

(ii)

The intention of the publisher is irrelevant.

(iii)

The hypothetical reasonable reader is not naïve but he is not unduly suspicious. He can read between the lines. He can read in an implication more readily than a lawyer and may indulge in a certain amount of loose thinking but he must be treated as being a man who is not avid for scandal and someone who does not, and should not, select one bad meaning where other nondefamatory meanings are available. A reader who always adopts a bad meaning where a less serious or non-defamatory meaning is available is not reasonable: s/he is avid for scandal. But always to adopt the less derogatory meaning would also be unreasonable: it would be naïve.

(iv)

Over-elaborate analysis should be avoided and the court should certainly not take a too literal approach to the task.

(v)

Consequently, a judge providing written reasons for conclusions on meaning should not fall into the trap of conducting too detailed an analysis of the various passages relied on by the respective parties.

(vi)

Any meaning that emerges as the produce of some strained, or forced, or utterly unreasonable interpretation should be rejected.

(vii)

It follows that it is not enough to say that by some person or another the words might be understood in a defamatory sense.

(viii)

The publication must be read as a whole, and any 'bane and antidote' taken together. Sometimes, the context will clothe the words in a more serious defamatory meaning (for example the classic "rogues' gallery" case). In other cases, the context will weaken (even extinguish altogether) the defamatory meaning that the words would bear if they were read in isolation (e.g. bane and antidote cases).

(ix)

In order to determine the natural and ordinary meaning of the statement of which the claimant complains, it is necessary to take into account the context in which it appeared and the mode of publication.

(x)

No evidence, beyond publication complained of, is admissible in determining the natural and ordinary meaning.

(xi)

The hypothetical reader is taken to be representative of those who would read the publication in question. The court can take judicial notice of facts which are common knowledge, but should beware of reliance on impressionistic assessments of the characteristics of a publication's readership.

(xii)

Judges should have regard to the impression the article has made upon them themselves in considering what impact it would have made on the hypothetical reasonable reader.

(xiii)

In determining the single meaning, the court is free to choose the correct meaning; it is not bound by the meanings advanced by the parties (save that it cannot find a meaning that is more injurious than the claimant's pleaded meaning).”

As for the common law position, the relevant principles were recently summarised by Warby J in Allen v Times Newspapers Limited [2019] EWHC 1235 (QB) at para 19:

“(1)

At common law, a statement is defamatory of the claimant if, but only if, (a) it imputes conduct which would tend to lower the claimant in the estimation of right-thinking people generally, and (b) the imputation crosses the common law threshold of seriousness, which is that it "[substantially] affects in an adverse manner the attitude of other people towards him or has a tendency so to do": Thornton v Telegraph Media Group Ltd

[2010] EWHC 1414 (QB) [2011] 1 WLR 1985 [96] (Tugendhat

J).

(2)

Although the word 'affects' in this formulation might suggest otherwise, it is not necessary to establish that the attitude of any individual person towards the claimant has in fact been adversely affected to a substantial extent, or at all. It is only necessary to prove that the meaning conveyed by the words has a tendency to cause such a consequence.": Lachaux v Independent Print Ltd

[2015] EWHC 2242 (QB) [2016] QB 402 [15(5)].”

I do not read para 16 of Lord Bingham’s speech in Jameel v Wall Street Journal Europe Sprl [2007] 1 AC 359 as importing a different test. A business may be entitled to sue in respect of “defamatory matters which can be seen as having a tendency to damage it in the way of its business”, but the test is not merely the latter part of this proposition after the relative pronoun. This section in Lord Bingham’s speech was addressing an anterior question of which entities could sue.

It is defamatory to impute present or future insolvency: see Gatley on Libel & Slander, 12^th Edn., para 2.30 and Whittington v Gladwin [1825] 5 B&C 180. The imputation in that Georgian case was not merely that there a risk of insolvency but that it would happen. In my opinion, an imputation of probable insolvency would also be defamatory. Where the line falls may give rise to difficult questions depending on the facts, but as will be seen the instant case does not generate these problems.

As for the difference between fact and opinion, the applicable principles are summarised in Koutsogiannis at para 16:

“(i)

The statement must be recognisable as comment, as distinct from an imputation of fact.

(ii)

Opinion is something which is or can reasonably be inferred to be a deduction, inference, conclusion, criticism, remark, observation, etc.

(iii)

The ultimate question is how the word would strike the ordinary reasonable reader. The subject matter and context of the words may be an important indicator of whether they are fact or opinion.

(iv)

Some statements which are, by their nature and appearance opinion, are nevertheless treated as statements of fact where, for instance, the opinion implies that a claimant has done something but does not indicate what that something is, i.e. the statement is a bare comment.”

The dangers of drawing too rigorous a distinction between the question of whether the statements complained of are fact or comment, and the logically subsequent question of when matters properly classified as opinion may be regarded as defamatory were highlighted in British Chiropractic Association v Singh [2010] EWCA Civ 350; [2011] 1 WLR 142 at para 32. In Zarb-Cousin v Association of British Bookmakers [2018] EWHC 2240 (QB) at para 38 Nicklin J determined the issue whether the statement complained of was fact or opinion before finding the single meaning. In my judgment, the particular facts of the instant case require a holistic or global approach rather than

Ms Wilson submitted that the credit report was “highly defamatory” and provided a “red flag” as to the Claimant’s creditworthiness. A statement about a business’s future viability could be a forecast but it was also saying something about the “here and now”. The Defendant was in reality making a statement as to the present creditworthiness of her client. The Defendant was also making statements as to the quality and reliability of its output which pointed away from the assertions in the credit report being in the nature of opinion. The overall impression of the report viewed as a whole was that the Claimant was at the margins of the red zone and that it was a poor credit risk. These matters were also borne out by the recommended credit exposure being as low as £15,000.

Ms Jolliffe submitted that the ratings are all in the nature of being assessments, and are therefore matters of opinion. Nowhere does the report or the guide state that a D&B failure score of 37 out of 100 is “poor”. On any view, this would be a matter of opinion. It is not in actuality borne out by the report, read in conjunction with the guide, because the epithet “poor” could only be properly allocated to those with a risk indicator of 4.

In my judgment, the natural and ordinary meaning of the words complained of, or the innuendo meaning, it matters not, is clear and the Defendant’s submissions about this, with some necessary modifications, are correct.

The Claimant is seeking to place its own interpretation or gloss on what the credit report says. It does not state, or imply, that the Claimant’s credit rating was “poor”. The contention that this is what the report says, or means, amounts – impermissibly in my view – to the Claimant’s own opinion about or comment upon what the report says; or, put another way, is not generated by the report as as a matter of permissible linguistic and legal inference. Furthermore, if, contrary to the above, the report was somehow saying that the Claimant’s creditworthiness was “poor”, that would be an expression of opinion rather than one of fact.

In the context of the D3 rating and the D&B failure score, the following points emerge.

First, the Claimant’s financial strength is £86,930. This is a bald piece of data derived from the financial accounts and in that respect cannot be gainsaid. In any case, in my view it tells one virtually nothing about the Claimant’s financial health because it is simply a measure of the company’s tangible net worth.

Secondly, the risk indicator of 3 indicates a greater than average risk of business failure. This is because 22% of businesses are classified as being in a higher risk category. 78% of businesses are therefore in a lower risk category and the risk of business failure is, as a matter of simple arithmetic, greater than average in relation to this Claimant. I would not characterise “greater than average” as being a statement of opinion. It is a statistical conclusion, or inference, derived from all the data in the Defendant’s possession relating as it does to the countless companies it has analysed. Because the conclusion is one impelled by the raw data using standard and straightforward statistical methodologies, I would hold that this is an imputation of fact: more precisely, of secondary fact.

Thirdly, the incidence of business failure in the next 12 months is 1.79%. I interpret this as saying that in relation to businesses with a failure score of 37 out of 100, and only 22% of businesses have a higher risk of failure, 179 out of 10,000 of these will in fact fail in the next 12 months. The Defendant’s algorithm has no doubt yielded these figures on the basis of the information and data its computer was provided. Again, these are facts and not opinions.

Fourthly, the Defendant’s explanation of this composite arithmetical conclusion is that businesses with a failure score of 37 out of 100 have “a moderate probability of failure”. The language of “moderate probability” is clearly a matter of opinion, because it is subjective and judgmental, depending on a number of factors including appetite for risk; but it is certainly one reasonable interpretation of the raw data. It is not, however, a straightforward statistical extrapolation, and therein lies the distinction.

If someone told me that my risk of a stroke or heart attack over the next year was 1.79%, I would probably call that “low” or “moderate” depending on my state of mind at the time. An epidemiologist would be able to inform me exactly where I was on the relevant graph: that would be a matter of fact, as would be any unadorned description of where I find myself on that spectrum. The point I am making is that “low” and “moderate” are subjective and/or judgmental ascriptions of risk which are inherently matters of opinion.

Fifthly, neither the language used in the report nor the pictorial representation I have appended lends any real support to the proposition that the Claimant’s credit rating would or should be understood as “poor”. The Claimant is not in the worst risk category nor is it clearly in the red zone.

For all these reasons, I cannot accept the Claimant’s argument that the credit report was saying, implying or suggesting that the company was in financial difficulty at the date of the report; or, more specifically, that it was a serious credit risk, with poor creditworthiness and a real risk of failure. The Claimant relies on the “impression given” by the report as a whole, including the pictorial representation, but what is necessary here is a reasonably attentive reading of the report in conjunction with the guide. A credit rating of this sort would merit this degree of attention by the hypothetical reasonable reader operating in the commercial world. The bottom line is that the risk of business failure was 1.79% and the Claimant was not in the worst category. Properly understood, the report was not saying that there was other than a “moderate probability” of failure (in the sense explained by the guide). Any probability above that (e.g. “poor” or “serious credit risk”) would in my judgment be a matter of impermissible extrapolation.

There was some debate with Counsel as to how much knowledge should be attributed to the reasonable commercial person who has bought the Defendant’s services. Ms Wilson strongly submitted that I should take care to avoid applying any detailed statistical knowledge I might have to the interpretative exercise. I see some of the force of what she was saying, but I have no particular expertise. I might have found it slightly easier than some in understanding this report in conjunction with the guide, but I would not agree that any of the foregoing analysis is particularly complex.

For completeness, no separate point arises on the figure of £15,000 being the “D&B maximum credit” figure. That advice has to be read in the light of the report as a whole,

On the first and second preliminary issues, I would rule as follows. The statements complained of mean that with a D&B rating of D3 and a D&B failure score of 37 out of 100, there was a greater than average risk that the Claimant’s business would fail (as defined in the guide) within the next 12 months because 22% of businesses had a higher risk, this amounted to a “moderate probability” of failure, and that the likely incidence of failure within the next 12 months was 1.79%.

I would further hold that all the foregoing statements were statements of fact, save from the inferential statement or comment that this amounted to a “moderate probability”.

Finally on the first and second issues, I would not hold that the meaning of the words complained of was as set out at either para 7 or para 8.2 of the Amended Particulars of Claim. In particular, the words complained of did not mean that the Claimant’s creditworthiness was “poor”. If, contrary to the above, they did, that would be a statement of opinion.

The final preliminary issue (the third in the list) addresses the question of whether the words complained of are defamatory at common law in the light of my findings on the first, second and fourth preliminary issues. In my judgment, a 1.79% risk of business failure over the succeeding 12 months, even if that figure were untrue because there existed some methodological error which has not so far been identified, would not be defamatory at common law. The Claimant would not be substantially lowered in the estimation of right-thinking people generally if a body of such individuals were made aware of that risk in the context of the explanations given in the guide. It is defamatory at common law to impute business failure; it would also be defamatory to impute a significant (i.e. substantial) risk. It is unnecessary for me to determine where the cutoff point falls, but in my judgment the figure of 1.79% is well below it.

I would reach the same conclusion in relation to the statement that there was a moderate probability of business failure. I have explained why this must be envisaged as a statement of opinion. Read in conjunction with the further information provided in the report (sc. 22% of businesses have a higher risk; the risk of business failure was 1.79% over the following 12 months) I would hold that the statement was not defamatory. I do not understand the Claimant to submit that it is.

For similar reasons, I do not conclude that the “D&B maximum credit figure” of £15,000 is defamatory either. As I have said at §36 above, it does not materially advance the Claimant’s case.

If, contrary to my analysis, the credit report should be understood as meaning that the Claimant had a poor credit rating, I would not wish to reach any definitive conclusion as to whether such an expression of opinion should be regarded as defamatory at common law in the circumstances of this case. Had the point been essential of determination, I would have sought further submissions from Counsel directed to some of the cases contained in the bundle of authorities which were not touched on by

I invite the parties to draw up an Order which reflects my rulings.