Oneflare Pty Ltd v Chernih

This appeal is concerned with an agreement for the provision of search engine optimisation (SEO) services to improve a website’s ranking in search engine results, particularly on Google. In organic search results (those excluding paid advertising), such rankings depend, amongst other things, upon the number and quality of links to the target website from other websites. Accordingly, “link building”, a form of SEO, is directed to improving those characteristics of the target website. It may take several forms, which range from encouraging so-called “natural” links (the product of an authentic editorial decision to create a link to the target website) to generating “artificial” or “unnatural” links (including on unrelated websites created or used solely for the purpose of manipulating the search algorithm). As expressed in its Webmaster Guidelines, Google opposes practices at the latter end of the spectrum. To discourage them, it uses algorithmic and human means to identify and demote websites that benefit from artificial links. In the circumstances of this case, Penguin 2.0 was the most recent update to Google’s search algorithm directed to that objective.

In May 2013, the appellant (Oneflare) retained the respondents (initially Mr Chernih and later his company, Linkbuild) to provide SEO services in relation to its website. The services ultimately provided by Mr Chernih included the creation of artificial links. As Mr Chernih freely accepted in cross-examination, the strategies adopted to boost the website’s ranking were only “good so long as you don’t get caught out”. Within a month of the introduction of Penguin 2.0 on 22 May 2013, the traffic to Oneflare’s website had diminished substantially as a result of an “algorithmic” demotion. The drop was from in excess of 22,500 visits per week to about 17,500 visits per week. On 26 July 2013, Google imposed a “manual” penalty on the website, following which its traffic dropped almost to zero. In early August 2013, the arrangement with Linkbuild was terminated. Oneflare engaged in further link building and, on 21 November 2013, a second manual penalty was imposed.

In November 2014, Oneflare brought proceedings against Mr Chernih and Linkbuild for damages for breach of contract, negligence and misleading or deceptive conduct. The last of those claims was abandoned. Oneflare’s case at trial was that Mr Chernih had developed and recommended a SEO strategy which was “incompetent" because, in the language of its opening written submissions, “the whole approach to off-page SEO - and link building in particular- was in such flagrant breach of the Google Webmaster Guidelines that it exposed [Oneflare] to the real risk (which materialised) of being punished by Google”. Those guidelines warned against engaging in “tricks intended to improve search engine rankings”.

There were three strands to Mr Chernih’s response. The first was that the directors and principal officers of Oneflare, Mr Dong and Mr Lim, were technically experienced and sophisticated users of the web and well aware of the nature of SEO services and how they interacted with search engine algorithms; and also well aware that there was a risk that in the face of “manipulative” SEO practices Google could take action that would negatively impact on the website’s search ranking results. The second was that he was retained by Oneflare to pursue and execute an “aggressive” SEO strategy which described activities that included the creation of “unnatural” or artificial links. The third was that Oneflare understood and implemented that strategy, including, of its own initiative and without Mr Chernih’s knowledge or advice, by the creation of “unnatural” links.

The primary judge (McDougall J) dismissed Oneflare’s claims: Oneflare Pty Ltd v Chernih [2016] NSWSC 1271. His Honour held that Mr Chernih had been retained to devise and implement SEO strategies that were “aggressive”; that such strategies included building “unnatural” links; that from the outset at least Mr Dong understood that the pursuit of such strategies carried the risk of adverse action by Google; and that Mr Chernih implemented and assisted Oneflare to implement those strategies. In doing so, the primary judge rejected as dishonest much of the evidence of Mr Dong and Mr Lim concerning their knowledge of link building and Google’s algorithm, and their role in the implementation of Oneflare’s strategy.

The grounds of appeal are considered in three parts. The first (ground 15) is concerned with the primary judge’s dishonesty findings. Oneflare submits that Mr Dong and Mr Lim were not afforded procedural fairness in relation to findings with respect to “fundamental parts of Oneflare’s case”: Judgment [16]. If that argument is upheld, the judgment must be set aside, and a new trial ordered. The second (grounds 1 to 9, 10, 11 and 12) challenges his Honour’s finding that Oneflare’s instruction that it wanted to pursue an “aggressive” SEO strategy involved, as was understood by Oneflare, that the link building to be undertaken was contrary to Google’s guidelines and carried the risk that Google might penalise the website. It is said that by reason of this error the primary judge’s findings as to the scope and terms of Mr Chernih’s retainer and as to breach also involved error. The remaining issue (ground 14) arises if Oneflare is successful on its challenges to those findings and conclusions. It is said that the primary judge’s application of a 30% discount to the “low estimates” of its loss involved error.

Before turning to these issues, it is necessary to refer in more detail to the background facts, and then to the findings and conclusions of the primary judge.

Oneflare operates an internet business directed to putting people who are looking for a service provider (such as a removalist or plumber) in contact with such providers. Once the prospective customer is on the Oneflare website, they are able to post details of the services they require. Service providers then provide quotations for those services. The online customer may then choose from the quotations provided. Oneflare’s revenue is derived from service providers, who pay a fee for access to the website and the opportunity to submit quotes for services.

The Google search engine generates and ranks search results by reference to its assessment of the characteristics and qualities of websites it has “indexed” and their relevance to the search enquiry. In determining a website’s ranking, Google’s algorithm takes account of the website’s “on-page” and “off-page” characteristics, where the latter includes the quantity and quality of links to the target website from other website. That component of the algorithm was explained in the following passage from Mr Weyher’s expert report, extracted at Judgment [42]:

The Google Webmaster Guidelines prohibit SEO practices which seek to manipulate the Google algorithm. At Judgment [44], the primary judge extracted Mr Weyher’s uncontroversial description of some of those activities and Google’s attitude to them:

At one end of the spectrum of link building is the generation of a link which is the product of what Mr Weyher described as an “authentic editorial decision made by a real human being”. Such a link is capable of being created intentionally for the purposes of SEO by the retaining of an original content creator to write a blog entry, press release or other article containing a link to the relevant website. That article is then posted to a reputable third party website which, exercising its independent editorial discretion, accepts or rejects the post. Link building of that kind would not, as I understand the evidence, be contrary to the Guidelines.

At the other end of the spectrum is the creation or exploitation of a website populated by hundreds or thousands of links to a target website. The content of those webpages may be very poor, consisting of nonsensical phrases or gibberish, or make sense but have no relevance whatsoever to the purpose or business of the target website. At the same time, the “anchor words” – which constitute the clickable link to the target website – may themselves be “unnatural” in the sense that they do not bear any relation to the text in which they appear. Such a website, serving no purpose other than to provide off-page links to another or other websites, is described as a “spammy site”, and the links as “spammy links”.

There is then a range of link building activity which falls in between. Some of that activity more closely replicates the features of a natural link. For example, artificial links may be created by purchasing an “aged” website, which has an established usage history in relation to a particular subject, and populating that website with a small number of off-page links, each created using original content and anchor words that relate to the text in which they appear.

Mr Chernih, whose evidence the primary judge accepted, considered the aim of SEO and the purpose of engaging a person with expertise in SEO techniques to be to apply “the right mix of SEO techniques to boost the value or rank of a website as high as possible without being penalised by a search engine for doing so” [emphasis added]. He also described that purpose as being to “artificially boost the value or rank of a website in the Google search algorithm. It is by artificial means because the increase in rank or value is being driven by actions taken by the operators of the website itself, rather than by user traffic”. That description could extend to actions “driven” or initiated by the operators of a website (such as the creation of original content for the consideration of an independent website), which would not involve breaches of the Google guidelines or at least any risk of being penalised. However, the strategies pursued by Oneflare and Mr Chernih contemplated and involved the creation of unnatural links contrary to the Google guidelines. One such strategy was the population of aged website domains, acquired for that purpose, with links to the Oneflare website.

Oneflare’s business was originally conducted under the name PickAQuote. In about January 2012, the Oneflare website became operative. Prior to that, Oneflare (then known as PickAQuote) had created a PickAQuote website and retained a freelance SEO provider to undertake link building to that website. In early May 2012, Mr Chernih was provided with a report which showed that, between May and July 2011, in excess of 7,700 links had been created to the PickAQuote website. Mr Chernih described those links as “spammy”.

In about March 2012, Mr Dong and Mr Lim learned of an American website known as Thumbtack which had a business model similar to Oneflare. They also became aware that Thumbtack had improved the traffic to its website very substantially by using SEO strategies, including link building. In their initial meeting with Mr Chernih on 27 April 2012, Mr Lim described Thumbtack as having “a very aggressive SEO policy”, and Mr Chernih was informed that Oneflare wanted to copy that business. At that meeting, and at a second meeting on 3 May 2012, it was said that the traffic to the Oneflare website was currently around 100 views per day, and that its traffic goal was 100,000 views per month within 10 to 12 months.

The primary judge’s acceptance of Mr Chernih’s evidence as to what was said at this first meeting is not challenged on appeal. Following reference to Thumbtack, the conversation (the “Me” being Mr Chernih) continued as follows:

On 4 May 2012, Mr Lim confirmed by email to Mr Chernih that Oneflare was looking to implement an “aggressive SEO growth strategy”. By an email sent on 17 May 2012, he also advised that Oneflare had decided to proceed with the SEO initiative with a budget of $1000 per month for Mr Chernih’s services for the first three months. Mr Chernih provided those services until late November 2012, from which time Linkbuild provided the services until August 2013.

The primary judge’s reasoning addresses the following four issues agreed by the parties and recorded at Judgment [5]. As his Honour also noted, that formulation concealed “a multitude of sub-issues”:

With respect to the first of these issues, the primary judge found:

His Honour then addressed whether, in the exercise of reasonable care and skill, Mr Chernih was required to warn Oneflare of the risks associated with an “aggressive” link building strategy:

The primary judge then considered the conduct of the parties in the period up to August 2013. That conduct was relevant to questions of breach of contract and duty of care. It was also relied on by Mr Chernih and Linkbuild as confirming the knowledge and understanding of Mr Dong and Mr Lim of link building and of the risks associated with it at the time the contract was made. Given the contract was informal and not evidenced in writing, the conduct was also relevant to whether the initial oral agreement had been varied or Oneflare’s instruction as to its strategy had changed. His Honour concluded at Judgment [151] that there was “nothing in the conduct of the parties following 31 May 2012 that would in any way add to the scope or terms of the retainer, as I have set out at [141], [142] and [147] above, or qualify, restrict or negate any of those terms”.

In particular, his Honour concluded that Oneflare’s basic strategy remained the same (Judgment [202]):

In relation to breach, Oneflare relied on the events after May 2012 as showing the extent to which Mr Chernih had formulated and implemented the link building SEO strategy pursued by Oneflare: Judgment [152]. For Mr Chernih, it was submitted that the dealings of the parties showed “that it was Oneflare, rather than Mr Chernih, that decided the SEO strategy to be followed” and that on a number of occasions Mr Dong and Mr Lim implemented “SEO strategies without seeking comment or advice from Mr Chernih”: Judgment [153].

Addressing these submissions, the primary judge made the following general findings. First, his Honour held that there were numerous occasions when Oneflare and Mr Chernih collaborated in the implementation of Oneflare’s SEO strategy. To the extent that strategy included link building which carried the risk of penalty from Google, it reflected what Oneflare wanted: Judgment [171]. Secondly, the primary judge found that Oneflare consistently undertook SEO activities, either through Mr Dong and Mr Lim or through other employees, without any reference to Mr Chernih: Judgment [175]. Thirdly, his Honour accepted that, from time to time, Mr Chernih did recommend software programs for Oneflare to use as part of its link-building strategy: Judgment [178]. And finally, that on the “relatively few” occasions that Oneflare referred something to Mr Chernih for advice, his advice was “rarely precise and unequivocal”: Judgment [179].

Addressing the second issue formulated by the parties, the primary judge held that the essential elements of the strategy which was pursued and implemented from May 2012 was that laid down by Messrs Dong and Lim on 27 April 2012: Judgment [218]. Mr Chernih then assisted in the execution of that strategy, both by providing assistance to Oneflare’s employees, and by carrying out aspects of the SEO work himself: Judgment [219].

With respect to the third issue, the primary judge held that no question of breach arose because Mr Chernih did not formulate the strategy pursued by Oneflare. He was retained to devise and implement an “aggressive” SEO strategy that included link building (Judgment [222]):

The fourth issue concerned the quantification of any damage Oneflare had suffered, assuming it succeeded on liability. The damage claimed was the marginal profit Oneflare lost because of the decline in both traffic and “sales” resulting from the algorithmic demotion in May 2013 and the manual penalty in July 2013. The proposed measure of that damage was the difference between the sales revenue actually achieved in the period 22 May 2013 to 17 March 2015 and the sales revenue that would have been achieved in the same period had those penalties not been imposed, less any additional direct costs which would have been incurred in achieving that increased revenue. Mr Cavanagh’s “low estimates” of the marginal profit lost as a result of those events were $464,516 due to the algorithmic demotion and $497,713 due to the first manual penalty. Oneflare contended that those estimates should be accepted as reflecting its loss: Judgment [269].

The primary judge discounted the sum of those estimates by 30% to account for two matters which he specifically identified – “the endogeneity problem and the apparent difference in the growth trendlines” – as well as for other contingencies and uncertainties, including the possibilities of increased competition and of Mr Dong and Mr Lim having engaged in alternative “aggressive” SEO strategies: Judgment [271]–[274].

In arriving at his conclusions on liability, the primary judge made significant adverse findings as to the credibility of Mr Dong and Mr Lim:

His Honour also made specific findings as to aspects of their evidence in support of his conclusion that Mr Chernih’s account of the meeting on 27 April 2012 should be accepted: Judgment [94]. Although Mr Dong was also at this meeting, he gave no evidence about it in his affidavit in chief. Mr Lim’s evidence as to what was said included that in response to a reference by Mr Chernih to his contract work in SEO having a “specific focus on link building” Mr Dong asked “What’s link building?” Mr Chernih denied that was said and maintained that during the meeting Mr Dong proffered “We know a bit about link building, we have had some SEO work done before when we were PickAQuote”. In his first affidavit in reply, Mr Dong denied saying these words and added that Oneflare’s earlier exposure to SEO (as PickAQuote) did not include “any kind of link-building or aged-domain purchasing work”.

The primary judge found that Mr Dong did not say “What’s link building?” His Honour’s second and third reasons for doing so were:

The fourth reason given by the primary judge for rejecting this evidence of Mr Dong included reference to an email sent by Mr Lim to Mr Chernih shortly after the meeting on 27 April 2012. In the course of dealing with that email, his Honour made the following finding about an aspect of Mr Lim’s evidence:

In Kuhl v Zurich Financial Services Australia Limited (2011) 243 CLR 361; [2011] HCA 11, the majority (Heydon, Crennan and Bell JJ) said at [67]:

Kuhl was a case in which the trial judge made a finding that a plaintiff claiming damages for a work injury had deliberately supressed material evidence adverse to his case concerning the circumstances in which his injury had occurred. The fact of that suppression had not been suggested to him in cross-examination and the defendant’s counsel had not submitted that finding should be made. If that submission had been made, as the plurality observed at [70], “a breach of the rule in Browne v Dunn would have taken place”. In such a case, the plaintiff’s “remedies might have included a refusal by the judge to accept or entertain the submission, and a recall of the plaintiff to the witness box to deal with the allegation”: Kuhl at [71].

That is not this case. Mr Chernih’s final written submissions concerning the credibility of Mr Dong’s evidence included:

With respect to Mr Lim, it was accepted that he was “generally more truthful and less obstructive than [Mr] Dong, and also less of an advocate for Oneflare’s cause than [Mr] Dong”. However, it was submitted that on occasion Mr Lim “appeared to be unwilling to answer questions or give truthful evidence that he believed would harm Oneflare’s case”. The most obvious example was said to be “his repetition of the mantra” that everything he or Mr Dong did in relation to SEO was done “under Nick’s guidance” or “under the advice of Nick” or “in consultation with Mr Chernih”. Where there was a conflict between Mr Lim’s evidence and Mr Chernih’s evidence, it was submitted that the Court should prefer the latter.

Oneflare’s written and oral responses to these submissions did not suggest that the credit issues raised had not been apparent or unexpected, or that the primary judge should not entertain them, or that the witnesses should be given a further opportunity to deal with them: cf Kuhl at [71]. On the contrary, the responses took issue with the submissions by ultimately arguing that there was no reason not to accept either witness as a witness of truth. At the same time, it was submitted that the Court should reject Mr Chernih as a truthful witness.

Oneflare’s complaint is that there was a breach of the rule in Browne v Dunn (1893) 6 R 67. As the following extract from the judgment of Lord Herschell LC makes clear, that rule applies where it is intended to suggest that a witness is “not speaking the truth on a particular point”. His Lordship said at 70:

Glass JA observed in Seymour v Australian Broadcasting Commission (1977) 19 NSWLR 219 at 224 (Reynolds JA agreeing) that the rule “is a rule of professional practice as well as a rule of conduct which is essential to fair play at the trial and fair dealing with witnesses”. Once it has been made clear that a witness’s evidence to a particular effect on a given subject should not be accepted as truthful, the rule does not ordinarily require that it then be put to the witness in relation to each answer to that effect on that subject that the answer is untrue. The rule is a practical one designed to ensure procedural fairness from the perspective of the witness and the party calling the witness.

Lord Herschell’s statement of the rule at 71 accepts that it is to be applied:

There are many ways in which a party or other witness might sufficiently be put on notice that their evidence on a particular subject or to a particular effect is challenged as untruthful, and as to the basis of that challenge. Depending on the circumstances, those ways include by pleadings, affidavits or statements of evidence exchanged before the hearing, and opening statements made or exchanged at the commencement of the hearing, as well as by cross-examination.

Oneflare submits that Mr Dong and Mr Lim were not afforded procedural fairness in relation to the general and more specific findings extracted at [30], [31] and [32] above. The subject matter of the general findings was their knowledge of and involvement in Oneflare’s deployment of link building that violated Google’s guidelines and was liable to be penalised. The findings made were that each deliberately sought to downplay that knowledge (Judgment [12]); that in relation to the deployment of those strategies each dishonestly maintained that all of those activities were undertaken under Mr Chernih’s instruction or advice (Judgment [13]-[14]); and that each also dishonestly maintained that in relation to that deployment he relied on Mr Chernih for instruction or advice (Judgment [15]).

Each of the more specific credit findings at Judgment [98]-[108] concerned the knowledge of Mr Dong or Mr Lim as to the kind of link building which it was proposed be undertaken with the assistance of Mr Chernih. Those findings were directed to: Mr Dong’s knowledge of link building gained from SEO services provided to PickAQuote (Judgment [98]); Mr Dong’s denial that he had said anything about link building at the first meeting and his denial that those earlier services related to link building (Judgment [100]); Mr Dong and Mr Lim’s evidence that there was reference to the Google updates Penguin and Panda at that first meeting (Judgment [102]); and Mr Lim’s evidence as to his understanding of what an “aggressive” link building strategy involved: (Judgment [108]).

The affidavit evidence exchanged before the hearing, the parties’ opening statements and the cross-examination of each of Mr Dong and Mr Lim made plain that the truthfulness, not merely the reliability, of their evidence on several matters was challenged. Those matters included what was said at the first meeting, their knowledge of link building and the implementation of Oneflare’s link building strategy. Mr Chernih’s written opening asserted:

It was also apparent by that time that the challenge to their evidence in relation to those subjects would be made by reference to Mr Dong’s earlier experience in website development, the creation and operation of the PickAQuote website, the contents of the first meeting, and Mr Chernih’s detailed evidence concerning the email and other communication he had with Mr Dong and Mr Lim in the period to August 2013.

By the time Mr Chernih’s first affidavit was served, it was plain that there were fundamental differences between his evidence and that of Mr Dong and Mr Lim that could not be explained as due to failures of recollection or of emphasis. For example, in relation to the first meeting, Mr Lim recalled Mr Dong asking “What’s link building?” Mr Lim also recalled Mr Dong asking questions about Google updates like Penguin and Panda which he maintained he had seen on Mr Chernih’s LinkedIn page. Mr Chernih denied that these exchanges occurred. He maintained that following that meeting he believed that both Mr Lim and Mr Dong had prior experience in SEO techniques and a high level of knowledge of those techniques. In his first affidavit in reply, Mr Dong described his understanding of SEO as “very limited” and denied that he or PickAQuote had any previous experience with link building.

Several of Mr Chernih’s email communications with Mr Dong and Mr Lim dealt with activities which they apparently initiated and undertook without his prior knowledge or advice. Many assumed or suggested a fairly sophisticated understanding on their part of what various SEO strategies involved. And there was an absence of communications in which Mr Dong or Mr Lim asked questions which might suggest a lack of any real knowledge of what link building involved and of Google’s attitude to it. In his second affidavit in reply, Mr Dong maintained, in relation to Oneflare’s “purchase of aged domain SEO strategy”, that at all times he relied on Mr Chernih’s “oversight, direction and management of that strategy”. That was not in accord with Mr Chernih’s evidence which was that Mr Dong and Mr Lim “generally did not consult me prior to purchasing domains”.

Similarly, there were substantial points of difference between the evidence of Mr Lim and Mr Chernih. They disagreed as to what was said at the first meeting concerning link-building and the Google updates. Mr Lim also gave evidence of the following conversation with Mr Chernih on 6 August 2013, at the time Linkbuild’s retainer was effectively terminated:

The significance of this evidence was obvious. If it was accepted, Mr Chernih had made admissions that almost certainly were fatal to his defence of Oneflare’s claim. Mr Chernih denied that there was any such conversation. That he would have so spoken was wholly inconsistent with his affidavit evidence. The question for the primary judge was who was telling the truth. Mr Lim’s evidence could not be dismissed as merely mistaken or based on imperfect recollection.

Oneflare’s opening submissions to the primary judge confirmed that there were fundamental issues as to the knowledge of Mr Dong and Mr Lim of link building and as to how Oneflare’s strategy was implemented. Oneflare’s counsel described there being a “hot dispute” as to whether “what the plaintiff did in terms of SEO activity was the product of advice and recommendation of the first defendant, or whether the plaintiff acted unilaterally. The plaintiff’s case is that it acted at all times in accordance with the advice and guidance of the first defendant.” It was also stated, in relation to that dispute, that Oneflare’s case was that it “was not aware that the activities it was undertaking at the advice and recommendation of the first defendant [were] in serial breach of the Google webmaster guidelines until it discovered that for itself following the algorithmic penalty of May 2013”.

The cross-examination of Mr Dong challenged his evidence that at the outset he did not know anything about link building and the risks that it involved; and that he relied on Mr Chernih and did not undertake any SEO activity without his advice or instruction. Those challenges as to the truthfulness of his affidavit evidence were made by reference to the subjects referred to in [46] above. The propositions put as reflecting the true position included: that Mr Dong did not ask questions at the first meeting about Google because he and Mr Lim “were well aware of the type of SEO techniques that Google rewards or punishes”; that while Mr Chernih was retained Mr Dong and Mr Lim were pursuing their own SEO strategies without any instruction or pre-approval from Mr Chernih; that his evidence that he learnt of the Google Webmaster Guidelines in June 2013 was false; that he was well aware of the existence of those guidelines at all times Mr Chernih was retained by Oneflare; and that his evidence that he did not proceed with any purchases of aged domains without first consulting, or receiving a direction from, Mr Chernih was not correct.

At the conclusion of Mr Dong’s cross-examination, the primary judge confirmed, in response to a question from Mr Chernih’s counsel, that there was no need to take the witness through each conversation with Mr Lim or Mr Chernih that was disputed. It was expressly noted that “the rule in Browne v Dunn is amply satisfied when competing affidavits are put on”. That short exchange assumed, as was otherwise obvious, that the credibility of much of the evidence of Mr Dong and Mr Lim, as well as that of Mr Chernih, was in issue.

In relation to the findings at Judgment [98]-[102], Mr Dong was cross-examined on each of the subjects on which his evidence was challenged. Whilst those challenges were directed to particular matters, their relevance ultimately lay in establishing his knowledge and experience of link building. With respect to each, he was given an opportunity to explain what he maintained the position was. It was not necessary in that context for the cross-examiner to put to him that his evidence was deliberately evasive or false. It was plain for the reasons I have already given that his evidence on that general subject was challenged as untruthful.

In cross-examination, Mr Lim was also challenged as to his understanding of the link building to be undertaken and the extent to which he and Mr Dong devised and implemented the strategies pursued. Again, those challenges as to the truthfulness of his evidence were made by reference to the subjects referred to in [46] above, insofar as they related to him. In the course of his cross-examination, the following propositions were put as reflecting the true position: that from time to time he and Mr Dong devised strategies without first seeking advice from Mr Chernih; that he had previously engaged persons to undertake link building for PickAQuote; that he believed that to increase traffic volumes to 100,000 visits per month required “an aggressive SEO policy”; that he was making up evidence because he thought it would assist his case; that he understood in June 2012 that “spam link building” could attract an algorithm penalty from Google; that he was aware from the outset of the possibility that particular SEO work could result in Google imposing a penalty; and that he had made up evidence as to his receiving advice from Mr Chernih about particular webpage content containing links.

With respect to the specific finding at Judgment [108] Mr Lim denied that an “aggressive” SEO strategy included “aggressively link building”. In context that question could only be understood as referring to link building which carried a risk of being penalised by Google. That sufficiently gave Mr Lim the opportunity to respond to the challenge to his evidence that he did not have that understanding.

For these reasons, ground 15 is not made out. The conclusion that there was no denial of procedural fairness is wholly consistent with the way in which counsel for each party conducted the trial, and with the primary judge’s observation at Judgment [11]:

Finally, I note that nowhere in Oneflare’s written or oral submissions in support of this ground was it suggested that there was any particular matter which either of Mr Dong or Mr Lim was not given an opportunity to explain or lead evidence about that was relevant to the issues on which their credibility was challenged.

Grounds 1 to 9 challenge his Honour’s finding that Oneflare’s instruction that it wanted to pursue an “aggressive” strategy entailed, to the knowledge of Oneflare and understanding of Mr Chernih, that the proposed link building would breach Google’s guidelines and thereby carry the risk of penalty. That error is said to affect the reasoning supporting his Honour’s overall conclusions as to the scope of the retainer and the breach of Mr Chernih’s duty of care. Grounds 10, 11, 12 separately challenge specific findings made at Judgment [185], [205] and [212].

Oneflare’s argument proceeds as follows: SEO includes “off-page” optimisation, which essentially is link building that may or may not be in accordance with Google’s guidelines. Google seeks to detect and not take account of or penalise link building which violates those guidelines. Oneflare instructed Mr Chernih to pursue an “aggressive” SEO strategy, which included link building similar to that employed by Thumbtack. The primary judge construed “aggressive” in this context as meaning link building which was not compliant with Google’s guidelines. His Honour thereby “conflated” aggressive link building and non-compliant link building and overlooked that an “aggressive” link building strategy could comply with Google’s guidelines. The evidence did not support his Honour’s conclusion that such an instruction described link building not in accord with Google’s guidelines. Accordingly, the primary judge should have found that Oneflare’s instruction as to “aggressive” link building required Mr Chernih to engage in link building that did not breach Google’s guidelines.

This “conflation” error is said to have induced a series of wrong findings, each of which is challenged: that Oneflare’s instruction was to pursue a SEO strategy that necessarily involved non-compliant link building (grounds 1(a), 5); that Oneflare understood instruction as requiring the pursuit of a non-compliant link building strategy (grounds 1(b), 2); that it was reasonable for Mr Chernih to understand that Oneflare understood what “aggressive” link building involved (ground 2); that Mr Chernih sufficiently warned Oneflare about the risks of such link building (grounds 1(c), 3, 4, 6); that there was no basis for imposing on Mr Chernih an obligation to warn of those risks where he rightly believed that Oneflare understood them (ground 7); and that there was no breach by Mr Chernih or Linkbuild of any obligation to use reasonable care and skill (ground 8). Finally, ground 9 formulates the favourable findings that it is contended the primary judge should have made, having accepted Oneflare’s underlying argument.

Oneflare’s argument takes issue with the primary judge’s conclusion – which involved construing the language of the parties – that its instruction to pursue an “aggressive” SEO strategy proposed link building “for its own sake” and accordingly involved the risk that Google might demote or penalise the website: Judgment [116]. His Honour’s reference to link building “for its own sake” describes the creation of artificial or unnatural links intended to “manipulate” Google’s search results in the sense described in Mr Weyher’s evidence extracted at [10] above. However, it did not follow, as his Honour’s other findings make plain, that such link building necessarily carried a high risk of detection and penalty by Google.

The primary judge recognised that, within the range of activities that may constitute such link building, the risk of detection and action by Google increased depending on the number and lack of “quality” of the links created. His Honour described those more risky activities as involving “over indulgence in (or over enthusiastic pursuit of) link building”: Judgment [118(4)], [146]. He also noted in this context that the success of any link building strategy that includes the building of artificial or unnatural links depends on the links not being detected as such by Google: Judgment [40]. These findings and observations show that, contrary to Oneflare’s submission, in concluding that its instruction was directed to the building of artificial or unnatural links, the primary judge did not overlook that link building could comply with Google’s guidelines. At Judgment [41], his Honour noted that such strategies are referred to as “white hat” strategies. He expressly recognised the distinction between those strategies and what Oneflare proposed by use of the words “for its own sake”.

The primary judge did not err in concluding that Oneflare’s reference to an “aggressive” strategy was one involving the creation of “unnatural” links or use of link schemes. Link building is the creation of links to a website, and the adjective “aggressive” is apt to denote both intensified quantity or frequency and the taking of calculated risks. The creation or acquisition of “natural” links to a website is not “aggressive” in either of those senses.

The context provided by the parties’ exchanges at their first meeting show that these were the senses in which “aggressive” was reasonably to be understood. Notwithstanding that the contract was informal and partly oral, the objective theory of contract requires, once a finding is made as to what was said, that the parties’ language be construed as reasonable persons in their position would have understood it: see Crown Melbourne Ltd v Cosmopolitan Hotel (Vic) Pty Ltd [2016] HCA 26; (2016) 90 ALJR 770 at [22] (French CJ, Kiefel, Bell JJ), [196] (Nettle J), [242] (Gordon J). In that first meeting, Mr Chernih described link building as “building links to artificially inflate your link profile in Google’s eyes” [emphasis added]. That was not a description of a strategy involving the creation of “natural” links which in no sense “artificially inflate” a website’s link profile. Mr Dong replied that Oneflare knew “a bit about link building”. That response would be understood as directed to link building of the kind described by Mr Chernih. And it was most likely that was what Mr Dong in fact intended because the link building to which he was referring was the creation of “spammy”, and accordingly “unnatural”, links: see [12] and [15] above.

Mr Chernih also explained that link building which focussed on the quantity and not quality of the links created was “more likely to get you penalised than help your ratings”. There was no point in his answering Mr Lim’s question in that way if the parties were talking about the creation of links which involved no such risk. In that context, the contrast implicit in Mr Chernih’s observation was between such an outcome and one in which the link building improved the website’s rankings without being penalised. All of this would have been apparent to someone who had a good working knowledge of link building techniques and the risks they presented.

That directs attention to the primary judge’s finding that at the time of these exchanges Mr Dong, and likely also Mr Lim, understood what was involved in link building “for its own sake”: Judgment [118]. That finding is challenged. That challenge must also be rejected. Mr Chernih explained in the exchanges to which reference has been made that he regarded an “aggressive” strategy as involving building links to “artificially” inflate the website’s link profile, an activity which carried the risk of being penalised by Google if detected. That risk of detection depended on the perceived “quality” of the links created. As those links were to “artificially” inflate the website’s link profile, their “quality” was to be measured by the extent to which they replicated or had the appearance of “natural” links. Mr Dong and Mr Lim did not ask for any further explanation or suggest that what they knew about link building did not extend to an understanding of what Google sought to penalise.

Almost immediately after Mr Chernih was retained, it was readily apparent from the email exchanges between the parties that Mr Dong and Mr Lim did appreciate that the link building being undertaken involved the creation of links of the kind Google sought to detect and penalise. It is sufficient to give four examples.

The first is an email exchange on 4 June 2012. Mr Lim had asked Mr Chernih to send him blogs written for posting onto third party websites. On receipt of those blogs by Mr Lim the following exchange occurred:

Mr Dong did not enquire what Mr Chernih meant by “suspicious”. The word was plainly understood as a reference to Google’s means of detecting “unnatural” links.

The second was an email exchange between Mr Lim and Mr Chernih on 8 June 2012, after Google notified Oneflare of a technical error in displaying its website and began to lower its ranking:

When cross-examined about this exchange Mr Lim accepted that by this time he was aware that link building carried a risk of penalty.

The third email exchange occurred on 23 and 24 July 2012. Mr Dong sent Mr Chernih an email that listed 5 off-page links to the Oneflare website on redirect pages of websites operated by the US Government. He asked:

The fourth occurred in early August 2012. On 8 August, Oneflare had received a Google Alert notifying it that there was a reference to “OneFlare Saves the Day” in a blog posted to a website address which included “cameroonchurches.org”. It was then the subject of the following exchange between Mr Lim, Mr Chernih and Mr Dong:

The comments “it’s not good”, “yeh it’s pretty shit”, “You should get away with it” and “easily on Google’s radar” make plain the understanding of each of the participants that the link building undertaken by Mr Dong was of the kind sought to be penalised by Google. There was no suggestion in any of these email exchanges of surprise on the part of Oneflare that the links being created had that character. Nor was it said that the creation of those links was contrary to Oneflare’s instruction or expectation. All of this was wholly consistent with his Honour’s findings that from the outset at least Mr Dong, and most likely also Mr Lim, were aware of those risks and were prepared to accept them (Judgment [116]), and that it was reasonable for Mr Chernih to understand, as he maintained he did, that they had “a good understanding of SEO techniques” (Judgment [118(3)]).

These conclusions require that the challenge made by grounds 1 to 9 be rejected. It was accepted before the primary judge and not argued otherwise in this Court that, if Mr Chernih was retained to implement an “aggressive” strategy and reasonably believed that Oneflare was aware of the risks involved, there could be no breach of the retainer in implementing such a strategy or breach of duty in not warning of the risks inherent in it: Judgment [222].

Ground 10 challenges the finding at Judgment [185]. That finding related to the email exchange of 8 June 2012 (see [71] above) concerning the acquisition of aged domains. The ground misstates the effect of the finding which was not that Mr Chernih did not recommend the buying of aged domains, but that the advice given by Mr Chernih on that occasion could not be interpreted as an unqualified endorsement of a strategy of purchasing aged domains and populating them with backlinks to Oneflare’s website. For this reason, ground 10 is without substance and rejected.

As to ground 11, in cross-examination Mr Chernih agreed “Your SEO strategy” was to “break Google’s guidelines at all times”. At Judgment [205], the primary judge described that evidence as a concession by Mr Chernih concerning “Oneflare’s SEO strategy”. This ground challenges that “finding” – and its reference to the strategy being Oneflare’s as distinct from Mr Chernih’s – as inconsistent with what Mr Chernih said. This ground also misstates the effect of his Honour’s finding. As is apparent from Judgment [208], among other places, the strategy to which his Honour is referring at Judgment [205] is that which “Oneflare retained Mr Chernih to devise and implement”. So understood, the finding was justified and in accord with his Honour’s findings at Judgment [218] and [219], as to which see [26]. Ground 11 is rejected.

Ground 12 challenges as a “finding” the primary judge’s statement at Judgment [212] that “Oneflare did not consult Mr Chernih to find out what could be done”. The primary judge continued “It retained Mr Chernih to devise and implement an SEO strategy that had as its centrepiece the creation of backlinks”. Oneflare’s written submissions acknowledge that this challenge depends upon its establishing that the second statement is not correct. For the reasons given above, that argument has not been made out and ground 12 must be dismissed.

Finally, none of these challenged findings, even if shown to have involved some error, is said or shown to require a different outcome on the question of liability to that which follows if grounds 1 to 9 are not made out.

In the result, these and the other grounds of appeal going to the question of liability are rejected. It follows that the appeal should be dismissed. There remains the now moot issue as to the quantification of Oneflare’s damages.

Mr Cavanagh’s methodology for calculating Oneflare’s loss required estimation of its counterfactual revenue between 22 May 2013 and 17 March 2015 but for the imposition of the algorithmic and manual penalties. There was then to be subtracted from that revenue the actual revenue earned during that period, as well as the additional direct costs it would likely have incurred to achieve the additional sales revenue. No challenge is made in the appeal as to that being a correct measure of Oneflare’s loss as a result of the alleged breaches of contract and of duty. The narrow issue between the parties is as to the discount applied to Mr Cavanagh’s estimate of loss using that methodology.

The first and third elements in Mr Cavanagh’s calculation required estimates to be made as to a hypothetical past situation if the penalties were not imposed (presumably because Mr Chernih would not have employed non-compliant SEO strategies in the period before May 2013). Mr Cavanagh estimated the first by applying assumed sales growth rates to maintainable sales revenue at the start of the loss period. The sales growth rates were estimated by taking the actual growth rates achieved by Oneflare in the six month period from March to August 2014 (when it was considered the likely effect of the penalties had dissipated). The maintainable level of sales was estimated by adopting the sales Oneflare was achieving immediately before 22 May 2013. That level of sales followed a period during which Oneflare had engaged in SEO strategies that were contrary to Google’s guidelines.

Before the primary judge, Mr Chernih submitted that the sales growth rates used to make the revenue estimate also included the beneficial influence of non-compliant SEO strategies employed during the earlier period. That beneficial influence was referred to by the parties and the primary judge as “the endogeneity problem”. In response to Mr Chernih’s submission that some allowance should be made for that problem (which was said to result in higher growth rates and thus to increase the estimated past revenue), Oneflare suggested that an allowance could be made for it by discounting Mr Cavanagh’s overall estimate of the lost marginal profit for the relevant period: Judgment [228]-[230]. The primary judge accepted that such a discount should be made: Judgment [257]-[258].

Mr Chernih submitted that the level of maintainable sales to which Mr Cavanagh applied those growth rates was overstated for the same reason. The primary judge accepted that was so: Judgment [259]-[260]. However, a further consideration revealed in his view a number of uncertainties that had to be taken into account when considering each of those estimates – the starting sales figure and the rates of growth. First, he observed that, even if only compliant SEO strategies had been employed before May 2013, the later growth rates were consistent with the sales figures in May 2013 being about the same as they were in fact: Judgment [261]. His Honour then noted that the revenue growth rates in the later period (from March 2014) appeared to be slightly higher than those during the earlier period from December 2012 to May 2013 and that, looking at the overall period from December 2012, the rate of growth may have “flattened out somewhat”: Judgment [262]-[263]. At that point, his Honour acknowledged that there “may be other possible outcomes, or other possible explanations, as to how these matters might have played out” and that, as a result, there was a level of uncertainty which must be factored into any discount.

In assessing the appropriate discount, his Honour took into account “the endogeneity problem and the apparent difference in the growth trend lines”. He also sought to take into account other factors that might have affected Oneflare’s estimated revenue in the past hypothetical period from May 2013. These included the possibility of disruption of Oneflare’s business (presumably conducted before May 2013 using compliant SEO strategies) from competition, as well as the possibility that Oneflare would have independently sought to employ some non-compliant strategies: Judgment [271], [273]. His Honour’s earlier discussion showed that these possibilities could have affected Oneflare’s revenue levels as at May 2013, and sales growth rates experienced after that time. Taking account of these matters, and accepting that the exercise was “necessarily impressionistic”, the primary judge concluded that an “appropriate discount rate” was 30%: Judgment [274].

Oneflare submits that in making this assessment the primary judge took into account irrelevant matters and failed to apply the principles referred to Armory v Delamirie (1722) 1 Stra 505; 93 ER 664, as explained in McCartney v Orica Investments Pty Ltd [2011] NSWCA 337 at [148]-[161] (Giles JA, Macfarlan and Young JJA relevantly agreeing). It is said those principles required, as distinct from permitted, the primary judge to resolve the uncertainties to which he referred against Mr Chernih “whose wrong has thrust the innocent party into a difficult task of proving a past hypothetical”.

It is sufficient, in order to deal with this second argument, to refer to the following observations of Giles JA in McCartney at [156]-[158]:

As the last observation makes clear, the principles identified do not require inferences to be drawn favourably to a plaintiff in all cases, or even in all cases where there is uncertainty which may be expressed as a range of possible outcomes. Accepting that to be the position, it is not shown that the primary judge erred in not taking that approach in this case.

Oneflare identified five matters which it submitted were taken into account, although not relevant. The first was the endogeneity problem which is said not to have existed based on the finding at Judgment [261]. That finding is that the rates of sales revenue growth in the periods December 2012 to May 2013 and March to August 2014 were “similar”. The primary judge did not, however, find that they were the same, and at Judgment [262] his Honour noted that the rate of growth in the later period was “slightly higher”. That difference was sufficient to justify a conclusion that the problem remained present.

The second and fourth matters relate to the differences in the growth trend lines for those two periods. It is accepted that there were uncertainties as to the reliability of each as an indicator of what the position was likely to have been in the past hypothetical. That is sufficient to justify the primary judge’s accounting for that uncertainty.

The third and fifth matters are the possibilities of adverse competition and independent use of non-compliant strategies. It is submitted that, as the loss being estimated was for a past period, what Oneflare might or might not have done was irrelevant. That is not so where, as here, that loss is being estimated by reference to a past hypothetical. As Brennan and Dawson JJ uncontroversially observed in Malec v JC Hutton Pty Ltd (1990) 169 CLR 638 at 639, “[h]ypothetical situations of the past are analogous to future possibilities: in one case the Court must form an estimate of the likelihood that the hypothetical situation would have occurred, in the other the Court must form an estimate of the likelihood that the possibility will occur. Both are to be distinguished from events which are alleged to have actually occurred in the past.”

The hypothetical events with which the sales revenue estimate is concerned are those that might have happened in the periods before and after May 2013, assuming that Mr Chernih in the performance of his retainer had made Oneflare aware of the SEO strategies available and the risks they entailed. The uncertainties included that Oneflare might have faced different competition in response to the different SEO strategies it adopted and that it may have continued to pursue non-compliant link building. The happening of those possibilities could have affected the level of Oneflare’s sales as at May 2013 and the rate of growth of its sales thereafter. In my view the primary judge is not shown to have erred in taking these matters into account. Ground 14 should be rejected.

In the result, the appeal should be dismissed with costs.

GLEESON JA: I agree with Meagher JA.

LEEMING JA: I agree with Meagher JA.