England and Wales High Court (Chancery Division) Decisions

IN THE HIGH COURT OF JUSTICE
CHANCERY DIVISION

		Royal Courts of Justice Strand, London, WC2A 2LL
		19 December 2002

Mr Justice Pumfrey:

Introduction

1. This is my judgment on the further adjourned hearing on the form of order resulting from my judgment delivered on 20 May 2002. The hearing has been adjourned twice (once from July, and on a second occasion in August) because further allegations which went to the scope of relief granted were raised by the claimants, on both occasions shortly before the hearing, which necessitated further adjournments.
2. In my judgment, I expressed the view that the claimants had not established a case for the immediate grant of an injunction. The case revolved around the website 'totaljobs.com' and my finding was as follows:

'89. By the trial a search on 'Reed jobs' was unlikely to elicit totaljobs.com at all, and if it did all the evidence showed that it would be ranked well below the claimants' own site, reed.co.uk. I will draw the inference that the word 'Reed' was the working part of the site until late, probably in November 2001, and odd pages may have continued to contain it in a form in which it might be spidered, depending, according to Mr Roche, on the enthusiasm of the spider and the depth of the page in the site.'

3. By trial, therefore, I considered that the word 'Reed' was ceasing to find the defendants' site.
4. The claimants devoted substantial energy to showing that this was not in fact the case, and in raising further allegations of what I consider to be bad faith on the part of the defendants Mr Hobbs QC strongly urged me to view the further allegations which were made (largely, though not entirely, through Mr Roche, the claimants' expert) as put forward on a disinterested basis, so that where it was said that a particularly matter 'could' be due to deliberate act or omission on the part of the defendants, in no case was the possibility of an innocent explanation excluded. What is more, he said, the defendants had been careless in the manner in which they had sought to remove the word REED from places where it might contribute to search engine indices. No doubt Mr Hobbs was right so far as the form of the allegations went. But the allegations were serious ones.
5. The claimant no longer suggests that injunctive relief is appropriate. I consider that this was a wise concession. The order will therefore include liberty to apply for an injunction, and a declaration in accordance with my judgment in respect of infringement and passing off.
6. The claim for an injunction was maintained until a letter from Slaughter & May for the claimants to dj freeman for the defendants. The costs generated on this issue have been substantial, for reasons which I shall now summarise.
7. Shortly before the date originally fixed for the hearing of the argument on the order, 14 July 2002, the third statement of Mr Roche was served on behalf of the claimants. It is a substantial document of thirty-eight paragraphs, twenty-four of which are concerned with the allegation that the defendants continued to use REED on totaljobs.com. On 20 May or thereabouts he says that the position with the search engines changed. He found search engine results which were identical to the data held on the DMOZ directory (paragraph 90 of the main judgment) but rejected DMOZ as being responsible for them He attributed the result to the presence of the word REED on a page of totaljobs.com between 30 January 2002 and 22 May 2002, that is, during or after the trial, and after the defendants had given evidence of their removal of all indexable instances of the word REED. He attributed the presence of the word REED to inefficiency on the part of the defendants.
8. I gave permission for this evidence to be answered. I wished to continue with the discussion in relation to costs, which I did. However, the contentions in respect of the disclosure sought prior to a decision to elect (so-called Island Records disclosure referring to the decision of Lightman J in Island Records v Tring [1996] 1 WLR 1256) ultimately evolved to the point at which it was clear that Mr Howe QC was inclined to suggest that there should be no inquiry as to damages.
9. Mr Roche clarified his report at the defendants' request in another twenty-four paragraphs. The further information appears to have been divided more or less equally between material going to damages and material relevant to the defendants' willingness to infringe the registered trade marks in suit or pass-off. A further hearing was fixed to take place before me on 27 August 2002. For this hearing, Mr Bloomer produced a very substantial report. On the material available to him, he formed the view that the 'click through' rate from search results which had been obtained using a search string containing the word 'REED', which would be a principal source of damage to the claimants if proved, were very low in number indeed. I give an example: in September 2001, there were click-throughs to the defendants' site from 16 searches containing the word 'Reed'. There were 45,498 click-throughs in total from search results, and some 648,000 visitors. Mr Bloomer was inclined in his report to take the view that any diversion of visitors who used the word 'Reed' in the search engines to find the claimants' site would be negligible. He also said that there had been no page on the site that year containing the word 'REED' that had been indexed by any search engine, thus directly challenging Mr Roche. This evidence itself caused a further adjournment, since it raised the real possibility that there was no appreciable access to the site through Reed-related search terms. If this was right, then the scope of any enquiry would be substantially reduced. Mr Roche had not inspected Mr Bloomer's data, and this was the purpose of the further adjournment. Mr Bloomer had worked on only sample months. The quantity of data analysed had been very substantial, as had the expense of generating it. It became quite clear that the investigation involved in an enquiry which investigated the rate of click-throughs on Reed-related search terms was likely to be extremely expensive. A question of proportionality therefore also arose.
10. By the hearing in October, Mr Roche had produced a further report. It appears that this report adequately sets out the claimants' contentions. Mr Roche agrees in substance with Mr Bloomer's calculations relating to click through on the sample months examined.
11. I do not think that the applicable principles are in doubt. They are well summarised in paragraphs 18-139 ff of Kerly's Law of Trade Marks and Trade Names (13th Edn). I wish to consider the basic principles of granting enquiries, however, referred to in paragraph 18-141, only when I have considered the heads of damage available in this case.
12. One preliminary matter must also be borne in mind. Candidates do not pay either the claimants or the defendants, who obtain their income from the clients, from advertisers on the website and from other sources. 'Diversion of trade' in this context has the rather special connotation that the attention of the user seeking a job website by reference to the word 'REED' is not concentrated on the claimants, but drawn to the defendants, but by reason of the word 'REED'. Thus, in analysing the material which has been made available, it is assumed that any candidate who typed a 'Reed-related search term' into the search engine wanted the claimants
13. Second, I consider that it is axiomatic that damage can be caused by trade mark infringement (use of an infringing sign) or passing off even if no diversion of trade takes place Thus an article bearing an infringing sign can continue to make an untrue representation which causes loss even though there was no deception at the time of sale Damages (calculated on the basis of a royalty or as a lump sum by way of reasonable compensation for use of the infringing sign) are available in such a case.
14. It is convenient to set out Mr Roche's contentions as they are advanced in his sixth report, as they represent the highest at which the claimants can put their case (WebTrends is the name of software which is used to analyse the detailed logfiles generated automatically by the site recording the relevant statistics of its users):

    '2.3 …Using WebTrends I have been able to view the information relating to Reed Related search phrases, including the extent to which Reed related phrases generated traffic to the site. The slightly altered configuration of WebTrends as run by myself did not produce results of a different order of magnitude from the results produced by [Mr Bloomer] in August 2002. I do however disagree with Mr Bloomer's conclusion that the figures are insignificant.

    2.4 Mr Bloomer has focussed on analysing the number of people using Reed related phrases on searches to get to Totaljobs. He suggests….that the only metric worth considering is that of click through. However, in my view this is only part of the picture. As I explained in my third report…the overall "visibility" of a site is also of prime importance. In this context I have considered the extent to which Totaljobs has been visible in connection with the word "Reed" as a result of the Defendants' use of the word "Reed". This connection has been made:

    (a) In the content of the site itself, when "Reed" appeared visibly to people visiting the site: "Reed" has appeared, for example, in the Reed Business Information logo, copyright notice and in the title (which appears at the top of the visitor's monitor screen) which for version 1 of the site read: "Total Jobs - The Recruitment service from Reed Business Information" [the headline was version 1 of the site only: see the judgment in respect of the RBI and Reed Elsevier logos];

    (b) In SERPs (search engine results pages) where the search engine has included the word Reed in the listing for Totaljobs (for example as a result of taking that listing from DM0Z for from metatags or title for the site), irrespective of whether the search phrase included the word "Reed";

    (c) where the Totaljobs site is listed in SERPS in response to a search including the word "Reed"; and

    (d) where Reed related searches triggered the appearance of a banner advertisement for Totaljobs.'

15. The click through figures are very small. I have quoted that for September 2001 (16 click throughs) above. June 2002 had 22 click throughs on Reed related phrases, some of which Mr Bloomer attributes to Mr Roche's activities. In the same month the word 'jobs' alone generated 16,612 click throughs. February 2002 showed 119 click throughs on Reed related phrases out of 148,251 search engine click throughs. February 2001 showed 33 click throughs on Reed related phrases from a total of 20,428. September 2000 showed 30 out of 5,385 and April 2000 showed 43 out of 7,059.
16. Since the experts are agreed in general as to the click through results, I can come to the conclusion that the actual number of visitors who arrive at the totaljobs site having searched against a Reed term and clicked upon a link to totaljobs.com on the search engine result page is and always has been negligible.
17. Mr Roche provided a number of estimates which were not challenged. In April 2000, there were some 237,809 visits to pages of totaljobs.com which included the word 'Reed'. This means in the form of 'Reed Business Information', 'Reed Elsevier' with or without their respective logos or in the copyright line to which I refer in the judgment. It was not used in any other way visible to the user on the site. Over the period January to June 2000, he estimates that there were 1,426,854 visitors to the site, and 4,395,246 views of the pages.
18. I can now turn to Mr Roche's headings (c) and (d). I have no difficulty in saying that they also are trivial. No doubt many sites having no connection to the claimant appear in response to a search on 'Reed' containing phrases. They are not all guilty of passing off or trade mark infringement. Subject only to the question whether its appearance on the results page is in consequence of the use of the word 'Reed' on the defendants' website itself it seems to me that (c) has nothing to do with trade mark infringement or passing off. In the main judgment, I have already indicated that (d) has no causal relationship to the use of the word 'Reed' and amounts neither to infringement nor passing off.
19. Mr Roche's class (b) attempts to separate out the case in which the word 'Reed' appears on the results page against a link to totaljobs.com although the search did not include the word 'Reed'. This potentially catches all searches which find the defendants until the site was effectively cleaned, and all searches which gave the DMOZ entry until quite recently. In. this case, as in the others, Mr Roche states that a 'link' is established in the mind of the user between the word 'Reed' and totaljobs.com. He relies also on links to totaljobs.com from sites established for RBI's various magazines, a contention which I have to consider fanciful: why should a link to totaljobs.com on a webpage for Hospital Update, which carries the RBI logo without objection from the claimants, in some way 'associate' totaljobs.com with RBI?
20. In his address to me, Mr Hobbs QC referred to an 'absolutely massive exposure' of the public to the infringing use. No doubt if one takes Mr Roche's estimates at face value, a very large number of people saw the defendants' logos, but of course one cannot know whether, and to what extent, they had any impact.
21. At the same time, point (b) is not completely devoid of substance, but since we can never know what prompted a user who did not use a Reed related search phrase, saw totaljobs.com with its descriptive matter including the word Reed, and clicked on the link, was thinking, the matter must be approached in a different way.

Principles

22. The law is that an enquiry will generally be ordered at the claimant's risk as to costs unless the court is satisfied either that there is no loss or that it can assess the appropriate relief by way of damages on the material available. For this proposition, one may cite McDonalds v Burger King [1987] FSR 112, Brain v Ingledew Brown Bennison & Garrett [1997] FSR 511. An enquiry may be awarded even if the court feels strong doubt about the availability of substantial damages. There is no reason why a claimant who seeks an enquiry should be in any worse position than a claimant who does not seek a split trial. But on the same basis the court has a duty not to involve the parties to litigation in very expensive disputes about matters which are trivial. I was told that the costs of the work involved in producing the data which was analysed by Mr Roche was very substantial (the matters arising after trial gave rise to costs of nearly £300,000) and this figure horrifies me. I cannot believe that it would be right to encourage expenditure on this scale, on top of a trial the costs of which were in total not less than £2m, unless I were satisfied that there would be a proportionate recovery.
23. Fortunately, in relation to one matter at least, I have no doubts. The investigation of click-throughs on Reed related search terms should go no further. On the sample months to which I have referred, there is no evidence of any substantial rate of click-through, and the amount of investigation required is anyway disproportionate. Although it is arguable that persons who clicked through to totaljobs.com after searching of a Reed related term can reasonably be assumed to have been deceived, their numbers are so trivial that not only should they not be investigated further but the fact that the click through rate is so close to zero is a valuable indication of the advantage derived by the defendants from their course of infringement by totaljobs.com for which they should pay. I have no doubt that they should pay upon a use basis: but in the calculation of payment for use it is necessary to hypothesize the willing licensor and licensee. The general principles applicable where the claimant cannot show an actual loss are described by Lord Wilberforce in a much-quoted passage in General Tire v Firestone [1976] RPC 197. Although a patent case, the principles are equally applicable to trade mark infringement.

"One who infringes the patent of another commits a tort, the foundation of which is made clear by the terms of the grant. This, after conferring the monopoly of profit and advantage upon the patentee, concludes by declaring infringers "answerable to the patentee according to the law for damages thereby occasioned."

As in the case of any other tort (leaving aside cases where exemplary damages can be given) the object of damages is to compensate for loss or injury. The general rule at any rate in relation to "economic" torts is that the measure of damages is to be, so far as possible, that sum of money which will put the injured party in the same position as he would have been in if he had not sustained the wrong (Livingstone v. Rawyards Coal Co. (1880) 5 App.Cas. 25, per Lord Blackburn, at p. 39).

In the case of infringement of a patent, an alternative remedy at the option of the plaintiff exists by way of an account of profits made by the infringer - see Patents Act 1949, section 60. The respondents did not elect to claim an account of profits: their claim was only for damages. There are two essential principles in valuing that claim: first, that the plaintiffs have the burden of proving their loss: second, that, the defendants being wrongdoers, damages should be liberally assessed but that the object is to compensate the plaintiffs and not punish the defendants (Pneumatic Tyre Co. Ltd. v. Puncture Proof Pneumatic Tyre Co. Ltd. (1899) 16 R.P.C. 209 at page 215.)

These elemental principles have been applied in numerous cases of infringements of patents Naturally their application varies from case to case Reported authorities, many of which were cited in argument, may be useful as illustrations of judicial reasoning, but are capable of misleading if decisions on particular sets of facts and observations in judgments leading up to such decisions are later relied upon as establishing a rule of law. Nevertheless I think it useful to refer to some of the main groups of reported cases which exemplify the approaches of courts to typical situations

1. Many patents of inventions belong to manufacturers, who exploit the invention to make articles or products which they sell at a profit The benefit of the invention in such cases is realised through the sale of the article or product In these cases, if the invention is infringed, the effect of the infringement will be to divert sales from the owner of the patent to the infringer The measure of damages will then normally be the profit which would have been realised by the owner of the patent if the sales had been made by him (see United Horse-shoe and Nail Co Ltd. v. John Stewart & Co. (1888) 13 AppCas 401). An example of this is Boyd v Tootal Broadhurst Lee Co Ltd (1894) 11 R P C 175 where the plaintiff manufacturers proved that a profit of 7s per spindle would have been made, and settlements of litigation for lesser rates were discarded

2. Other patents of inventions are exploited through the granting of licences for royalty payments In these cases, if an infringer uses the invention without a licence, the measure of the damages he must pay will be the sums which he would have paid by way of royalty if, instead of acting illegally, he had acted legally The problem, which is that of the present case -the respondents not being manufacturers in the United Kingdom - is to establish the amount of such royalty. The solution to this problem is essentially and exclusively one of evidence, and as the facts capable of being adduced in evidence are necessarily individual, from case to case, the danger is obvious in referring to a particular case and transferring its conclusions to other situations.

Two classic cases under this heading are Penn v. Jack (1866) 14 L.T. 495; (1867) L.R. 5 Eq. 81 and Aktiengesellschaft fur Autogene Aluminium Schweissung v. London Aluminium Co. Ltd. (No. 2) (1923) 40 R.P.C. 107. In Penn v. Jack the patentee was shown to have approached all users of the invention and to have successfully required the vast majority to pay him a royalty of 2s 6d per horse power. The defendant was one of the few who refused and it was held that he should pay damages for infringement based on the accepted royalty rate on the basis that he might have expected to have got a licence at the same rate.

The Aluminium case contains a clear statement by Sargant J., at pp. 113-1 14:

"... what has to be ascertained is that which the infringer would have had to pay if, instead of infringing the patent, he had come to be licensed under the patent. I do not mean by that that the successful patentee can ascribe any fancy sum which he says he might have charged, but in those cases where he has dealt with his property merely by way of licence, and there have been licences at certain definite rates, there prima facie, apart from any reason to the contrary, the price or royalty which has been arrived at by means of a free bargain between the patentee and the person desiring to use the patented article has been taken as being the price or royalty that presumably would have to be paid by the infringer. In doing that, it seems to me that the court is certainly not treating the infringer unduly harshly; he should at least, in my judgment, have to pay as much as he would in all probability have had to pay had he to deal with the patentee by way of free bargain in the way in which other persons who took licences did in fact pay."

These are very useful guidelines, but the principle of them must not be misapplied. Before a "going rate" of royalty can be taken as the basis on which an infringer should be held liable, it must be shown that the circumstances in which the going rate was paid are the same as or at least comparable with those in which the patentee and the infringer are assumed to strike their bargain. To refer again to Boyd v. Tootal Broadhurst Lee Co. Ltd., 11 R.P.C. 175: when it was argued that because numerous other persons had agreed to pay at the rate of 4s per spindle the infringer should also pay at the rate (rather than at 7s per spindle, which represented the normal profit), it was relevant to show that the rate of 4s was negotiated by way of settlement of litigation in which the validity of the patent was in doubt. This was not the equivalent of that which the court had to assume: for that purpose the patent must be assumed to be valid. This line of argument is very relevant in the present case, for, as I shall show, the appellants adduced a great deal of evidence as to the royalties actually agreed by various licensees, and this was discarded, totally, by the learned judge and the Court of Appeal. They had every right to discard it if the bargains which led to these royalties being agreed were reached in circumstances differing from those which must be assumed when the court is attempting to fix a bargain as between patentee and infringer. The central question in the present case is whether this difference existed.

3. In some cases it is not possible to prove either (as in 1) that there is a normal rate of profit, or (as in 2) that there is a normal, or established, licence royalty. Yet clearly damages must be assessed. In such cases it is for the plaintiff to adduce evidence which will guide the court. This evidence may consist of the practice, as regards royalty, in the relevant trade or in analogous trades; perhaps of expert opinion expressed in publications or in the witness box; possibly of the profitability of the invention; and of any other factor on which the judge can decide the measure of loss. Since evidence of this kind is in its nature general and also probably hypothetical, it is unlikely to be of relevance, or if relevant of weight, in the face of the more concrete and direct type of evidence referred to under 2. But there is no rule of law which prevents the court, even when it has evidence of licensing practice, from taking these more general considerations into account. The ultimate process is one of judicial estimation of the available indications The true principle, which covers both cases when there have been licences and those where there have not, remains that stated by Fletcher Moulton L.J. in Meters Ltd (1911) 28 R.P.C. 157, 164-165: though so often referred to it always bears recitation.

"There is one case in which I think the manner of assessing damages in the case of sales of infringing articles has almost become a rule of law, and that is where the patentee grants permission to make the infringing article at a fixed price - in other words, where he grants licences at a certain figure. Every one of the infringing articles might then have been rendered a non-infringing article by applying for and getting that permission. The court then takes the number of infringing articles, and multiplies that by the sum that would have had to be paid in order to make the manufacture of that article lawful, and that is the measure of the damage that has been done by the infringement. The existence of such a rule shows that the courts consider that every single one of the infringements was a wrong, and that it is fair - where the facts of the case allow the court to get at the damages in that way - to allow pecuniary damages in respect of every one of them. I am inclined to think that the court might in some cases, where there did not exist a quoted figure for a licence, estimate the damages in a way closely analogous to this. It is the duty of the defendant to respect the monopoly rights of the plaintiff. The reward to a patentee for his invention is that he shall have the exclusive right to use the invention, and if you want to use it your duty is to obtain his permission. I am inclined to think that it would be right for the court to consider what would have been the price which -although no price was actually quoted - could have reasonably been charged for that permission, and estimate the damage in that way. Indeed, I think that in many cases that would be the safest and best way to arrive at a sound conclusion as to the proper figure. But I am not going to say a word which will tie down future judges and prevent them from exercising their judgment, as best they can in all the circumstances of the case, so as to arrive at that which the plaintiff has lost by reason of the defendant doing certain acts wrongfully instead of either abstaining from doing them, or getting permission to do them rightfully"

A proper application of this passage, taken in its entirety, requires the judge assessing damages to take into account any licences actually granted and the rates of royalty fixed by them, to estimate their relevance and comparability, to apply them so far he can to the bargain hypothetically to be made between the patentee and the infringer and to the extent to which they do not provide a figure on which the damage can be measured to consider any other evidence, according to its relevance and weight, upon which he can fix a rate of royalty which would have been agreed If I may anticipate, I have to find that the process carried out by the courts below does not satisfy this requirement."

24. Any case in which the claimant cannot show lost sales or loss of business must be approached on the 'user' basis: how much should the defendant have to pay for the use he has made of the sign? If there is damage to goodwill, as by selling inferior goods under the mark, a sum will be awarded for that damage The present case is purely a use case There is no evidence of complaint to the claimants about the defendants' services or anything of that sort. At the same time there are undoubtedly heads of infringement where I consider that an award is appropriate. These are as follows:

i) There should be an inquiry as to damages for the so-called visible use on the totaljobs.com site itself (judgment paragraph 119). This follows on orthodox principles. Since no-one has ever suggested that there would be evidence of actual confusion over and above the evidence to which I have referred in the judgment, I apprehend that damages will be assessed on the user basis, globally.

ii) The same goes for the isolated case of the Yahoo! use of a banner triggered by the word 'Reed' as a search term. I have not discussed this above, since it is a straightforward parallel to an inquiry in respect of appearance on the totaljobs.com site itself. The scale of diversion was small (judgment paragraph 62) but not entirely negligible.

25. To follow the same logic there should also be an inquiry in respect of the invisible use of the word 'Reed' on the site, the damages to be fixed on the basis of a reasonable payment between willing licensor and willing licensee on the footing that clickthroughs on the search results using the word 'Reed' will be vanishingly small, and that any benefit accruing to the defendant arises only from the juxtaposition of 'Reed' and the link to totaljobs.com on search engine response pages. In the judgment, I discuss the extensive advertising of the site off the internet. To attempt to disentangle the effect of this advertising on numbers visiting the site from the effect of the word 'Reed' in the phrase 'Reed Business Information' on search engine response pages would be difficult or impossible. I wish to emphasise that inherent in my finding that the number of click-throughs on search terms using the word 'Reed' is vanishingly small is a finding that the presence of the word 'Reed' in the phrase 'Reed Business Information' on search engine result pages is unlikely to have increased the number of visitors to the site. This will be the basis of the enquiry. The inquiry will be at the risk of the claimants as to costs and I shall reserve it to myself. I shall stay the inquiry pending an appeal to the Court of Appeal.
26. A minor issue arises in relation to press releases The press releases do not amount to infringement or to passing off They do not fall within the scope of the registration, and no-one looking at them fairly could conclude that they made a relevant misrepresentation to the interested public The claimants failed in their claim in respect of press releases.

Costs

27. The claimants having succeeded at the trial, they are on the face of it entitled to the costs. There was an open offer of settlement on 13 November 2001, the salient features of which are:

i) An undertaking not to use REED in connection with the provision of employment agency services, and not to pass-off their employment agency services as and for those of the claimants by the use in connection therewith of the word REED;

ii) A proviso permitting the defendants to carry on their established businesses other than employment agency services, including all activities carried out electronically, under or by reference to REED or any mark or corporate name including REED;

iii) A further proviso permitting identification of source, author or copyright owner in accordance with honest industrial practices;

iv) £250,000 or one-third of the claimants taxed costs as November; and of 21 days from 13 November; and

v) £10,001 in full and final settlement of the claim for damages or profits.

28. It seems to me that the offer was beaten. The defendants made no relevant admissions in respect of the claimants' reputation. Much evidence was given over to this topic, which should not have been necessary. Nonetheless, I had an impression that the expenditure of time and effort at trial was disproportionate to the results actually achieved. Much of what the plaintiff practically obtained had been brought about well before the trial, and the internet complaints required a great amount of elucidation (and further evidence) at trial. All this should, and could, have been worked out before the trial. There was a huge diversion into 'search engine optimisation' which only incidentally had anything to do with the issues in the action. It may be (and I am still not entirely sure) that the claimants were seeking to allege that the totaljobs corn site had been optimised to respond well on 'Reed' in search phrases. In this they failed. The one issue to which I refer in my judgment on which a wholly unnecessary amount of time was spent at trial was the allegation concerning 'robots.txt', which, it was suggested, raised an issue relating to the defendants' good faith.
29. I am always inclined to take an issues-based approach to costs Ms Lanzkron's witness statement shows broadly how the defendants consider that their costs were expended In respect of each issue on which the party successful overall fails, there are three questions should the party successful overall recover his own costs of that issue, or should there be no order for costs, or should the successful party pay the unsuccessful party's costs. The costs which cannot be allocated to any particular issue are paid by the unsuccessful party to the successful. The only issue in respect of which I consider that the claimants should not recover their costs is the robots.txt issue. It is true that they failed on minor issues (Invoices, the Harrogate show, Press releases and so on) but those were peripheral and took little time. But since the robots.txt issue was advanced to provide a support for an allegation of bad faith, which failed, I would give the defendants their costs of that issue. My impression is that to say it represented 5% of the trial would put it too low, and 15% too high I shall settle on 10%, which means that the claimants are entitled to 80% of their costs assessed on the standard basis This is somewhat more than one-half of Ms Lanzkron's estimate, but she includes matters (including the work done by Outrider) which in some degree should have been in the case anyway. I make no allowance for the matters to which I refer above upon which I felt some doubt. These are matters best dealt with on detailed assessment.

Costs of the post judgment investigations

30. As I have indicated, Mr Howe told me that the defendants' costs after judgment amounted to just short of £300,000, including the costs of obtaining and analysing the information relating to sample months. He makes a separate application for the costs of this exercise, since on this point his clients have succeeded, as they would have succeeded on an enquiry in which this would have been investigated at even greater length and expense. Some of this expenditure is attributable to Mr Roche's evidence on the 'injunction issue' where no notice was given of an investigation carried out on 12 June 2002 until shortly before the hearing in July, when it was no longer possible to replicate Mr Roches's results. Moreover no leave was obtained for Mr Roche's report. This caused a very substantial satellite dispute involving much investigation.
31. I accept these submissions. The claimants started a dispute which was without substance, and should have attempted to resolve it with the defendants before unleashing the relevant part of Mr Roche's third report shortly before the July hearing. I consider that they should pay the costs of this operation. I will order that the claimants pay the defendants their costs of dealing with the matters arising out of paragraphs 1-24 of Mr Roche's third report to be assessed if not agreed, and there will be a set-off. The remainder of the costs arising from Mr Roche's report, which relate, in effect, to the question of financial remedies, is reserved to the inquiry.