Law firms must find smarter ways of handling e-disclosure that get to the heart of a case without breaking the bank. Grania Langdon-Down reports.

Given the ubiquity of electronically stored information (ESI), which includes both the information in emails and documents as well as from databases, social media, digital audio files and mobile phones, it is not surprising that e-disclosure has developed into a multibillion-pound industry.

The software market is continuously evolving to keep pace. Analytics for early case assessment and data reduction, predictive coding (a form of artificial intelligence which involves lawyers ‘training’ a computer program to search for relevant and privileged documents) and information governance are highlighted as dominant trends by IT research company Gartner.

From its research, Gartner estimates that the market for e-disclosure tools to manage the cost of lawsuits will grow from $1.8bn in 2014 to $3.1bn by 2018.

But the importance of being on top of data does not stop at lawsuits, whether civil, criminal or increasingly regulatory, as bodies such as the Competition and Markets Authority ramp up investigations. E-disclosure tools are also being used in arbitrations, M&A due diligence, subject access requests and data breach investigations.

What is clear is that it can be damaging both in costs and reputation if you get it wrong. In 2009 Judge Simon Brown issued a stringent warning in Earles v Barclays Bank that it would be ‘gross incompetence’ not to know the rules governing e-disclosure when ‘over 90% of business documentation is electronic in form’.

Six years on, Mr Justice Edwards-Stuart, in charge of London’s Technology and Construction Court (TCC), and US magistrate Judge Peck in New York, who has led the way in encouraging the use of predictive coding, tell the Gazette that technology-assisted review (TAR) is increasingly going to be the technology of choice (see box below).

Simon Tolson, chair of the Technology and Construction Solicitors’ Association (TeCSA), says it is no longer acceptable in commercial litigation and arbitration for a lawyer to be unaware of the technology used in e-disclosure.

Earlier this year, TeCSA, along with TECBAR, the corresponding association of construction barristers, and the Society for Computers and Law, launched an updated version of their 2013 e-disclosure protocol. Developed in consultation with TCC judges, it is aimed at creating a more level playing field in terms of understanding this challenging and fast-moving area of practice.

It provides a voluntary framework for parties to conduct e-disclosure, including a ‘clawback’ agreement that any confidential or privileged document disclosed inadvertently will immediately be returned without it amounting to any waiver of privilege.

The protocol is backed by guidelines with a checklist of 138 boxes, taking practitioners through the e-disclosure process. An e-disclosure guide, written by Andrew Haslam of Allvision Computing, acts as a user’s manual.

‘The announcement that the protocol was to be enshrined in the TCC guide gave us “street cred” overnight,’ says Tolson, senior partner at construction and energy law specialists Fenwick Elliott. While the protocol is designed specifically for cases in the TCC, he says it is capable of being used in any piece of civil litigation.

Barrister Michael Wheater, of Hardwicke Chambers, is co-writing a book on e-disclosure law and practice to be published next year. He specialises in large-scale construction litigation and says one of the big fears for solicitors is the cost. But, he argues, e-disclosure does not have to be expensive. ‘Many people say there are cases where it is not economical to do it – but that suggests e-disclosure is different to ordinary disclosure,’ he says.

‘It doesn’t matter if a claim is worth £15,000 or £15m, your duty is to carry out a reasonable search for documents that might help or hurt your case or that of other parties. But that doesn’t mean bringing in an e-disclosure consultant to take a forensic image of someone’s hard drive if the case is worth £15,000.’

In Goodale v MoJ [2009] Senior Master Whitaker said there is no reason why a party that complies with its e-disclosure obligations should not be able to recover the fees of its e-disclosure service provider, subject to the usual cost recovery principles.

But Wheater warns there has been an increase in the number of contentious costs applications where parties are becoming less tolerant of deficiencies in disclosure provided by others and could seek to recoup the resulting cost of reviewing large volumes of irrelevant material upfront.

What is clear is that the need for smarter search and analysis tools is driving the use of predictive coding. It is not a ‘magic bullet’ – it only works on fully electronic documents, so is not effective with plans or drawings. But those who have used it say it is more effective than the ‘blunt instrument’ of keyword searches.

Earlier this year, the Irish High Court, in the first ruling of its kind in Europe, followed the US courts in approving its use, with the judge stating that TAR of large datasets is at least as accurate as, and probably more accurate than, a manual or linear method in identifying relevant documents.

Celina McGregor, senior associate with Herbert Smith Freehills and a member of the Junior London Solicitors Litigation Association Committee, says her firm has used predictive coding in a number of cases, including a substantial disclosure exercise in the Commercial Court. ‘If you look at the US, it has taken three years to reach the point where it is now part of procedural management decisions,’ she says.

‘We are certainly not there yet because of cost, lack of judicial endorsement and inertia, but it can lead to substantial cost savings.’

So, with such a fast-moving industry, how do you procure the best deal for your client? Electronically stored information is notorious for becoming a black hole of ever increasing volumes, says Tolson, which makes it difficult to price the work. ‘Service providers have three main components to their cost base: storage, software and their professional services, so flexibility on pricing will depend whether they own or rent the storage space or own their own software,’ he says.

‘Just as the courts and clients are looking to pass risks on to the law firms, so you could be looking to pass some of that on to your litigation support partner. Note the deliberate use of the word “partner”. You cannot get a good price by ringing round the suppliers and seeing who will give you the cheapest quote for “processing 50GB”.’

‘This will become the norm’

Two judges from opposite sides of the Atlantic give their perspective on e-disclosure

Mr Justice Edwards-Stuart has been in charge of the Technology and Construction Court since 2013; Judge Peck is a federal magistrate judge for the Southern District of New York.

Judge Peck is best known for his Da Silva Moore (2012) and Rio Tinto (2015) decisions on the use of technology-assisted review (TAR). He is keen to allay fears that predictive coding is handing too much control to the ‘black box’.

‘There are US studies that show TAR is at least as effective, if not more so, than anything else and is significantly cheaper,’ he says. ‘The volume of data nowadays is just so great that it is too cost-prohibitive to have lawyers or paralegals review every document. As I said in Rio Tinto, in all cases where the producing party wanted to use predictive coding, the court said “sure, go ahead”. But we are not at the stage where courts will force the producing party to use it merely because the requesting party wants it.’

However, he says predictive coding is clearly being used in the US and probably in the UK more than is apparent from reported cases – ‘the various vendors could not be in business based on about 20 cases in the US’.

Mr Justice Edwards-Stuart does not believe predictive coding needs judicial endorsement here to take off, though he does predict that, as parties become more comfortable with it and the software more sophisticated, it will become the norm within a couple of years.

‘We haven’t got to the stage where we have had a dispute over it which needed a reported decision,’ he says. ‘Disclosure is very much a hands-on, negotiated issue. The parties come with various positions and you talk them through it and, after 20 minutes to-ing and fro-ing, you usually get them to an agreed position.’

Views differ about the efficacy of predictive coding, he says: ‘There is a school of thought that, if you do it properly with enough iterations and a big enough population sample, it becomes more reliable than half a dozen good solicitors.

‘However, it is only as good as the people doing the initial run – garbage in, garbage out.’

A big issue for practitioners is the risk of inadvertently disclosing privileged material.

Judge Peck says: ‘With the volume of data, some privileged documents may slip through, no matter how careful you are. Under Federal Rule of Evidence 502(d), if you have an agreement with an adversary that is approved by the court, it is then binding on the world – at least the US version of the world.

‘Otherwise you have to go through the common law waiver approach, which is now codified, and show you had been careful in your screening for privilege, the disclosure was inadvertent and you did something quickly to get it back. Sometimes it is easy to prove, sometimes difficult, but it is always embarrassing.’

In the TCC, the TeCSA e-disclosure protocol has a ‘clawback’ provision for inadvertent disclosure. ‘It is something solicitors welcome and feel strongly about because it is always a trap for the unwary,’ Mr Justice Edwards-Stuart says. ‘This is not something the court can impose but something parties voluntarily sign up to because it suits both of them. They know that it might be them that makes the mistake. Since the protocol, I am not aware that there has been an application to get a privileged document back.’

He says the protocol is, for the most part, being used very successfully. Could it be applied in other courts? ‘I would have thought other courts would be daft if they didn’t use it,’ he says. ‘They may want to take a slightly different approach because this is geared towards construction disputes. But it would be tinkering, nothing fundamental, and certainly as far as the Commercial Court is concerned it could almost be taken lock, stock and barrel. The Chancery Division has a much wider spread of litigation so it might have to deal with different classes in different ways.’

But how far should courts be driving the use of technology? The irony, says Mr Justice Edwards-Stuart, is that the ‘cheapest form of disclosure on a per-document basis is to dump the lot on the other side and let them look for what they want. But that carries the risk of inadvertent disclosure. Narrowing the harvest always puts up the per-document cost, so it is a balance between the cost of the exercise and how refined you make it.’

He does not think the Jackson reforms have made much difference to e-disclosure. ‘If the computer can do it, it can do it, and the cost of using it to look at a hundred thousand documents rather than one million isn’t huge because all the grunt work is being done by the computer. It is more about whether the per-document cost for refining the search justifies the end result.’

What is difficult is to make a costs order in advance, he says, ‘particularly if you are sceptical and think the request is extravagant or strongly suspect it is justified, so you wait and see the outcome. If you do that, parties are less confident because they know the proof of the pudding will be in the eating’.

What will the process look like in five years?

‘I think the changes in technology will be less than in the last five years, except for predictive coding,’ says Mr Justice Edwards-Stuart.

‘We have held several paperless trials. As the courts move more in that direction, the more e-disclosure will become the norm because the two will be related. If you are having all documents scanned in readable form for the court, it is immediately interfacing with the e-disclosure process.’

He stresses it is important that dispute resolution lawyers ‘don’t allow the technology tail to wag the litigation dog’, reflecting: ‘About four to five years ago I had a real concern that the technologists in data management and hosting were going to hold us all to ransom, and it has taken that time to get the “dog back on its lead” – if you’ll excuse the idiom.’

For Haslam, practitioners should see e-disclosure as a business development opportunity. His guide warns that clients and law firms, desperate to keep costs down, may propose to collect the data themselves. But without the requisite expertise, there are potential dangers. His advice: ‘Get a professional to do the job: then if it does all go wrong their insurance can take the hit, not your reputation.’

Instead, practitioners should approach clients, possibly with a technology partner, and discuss how the client can become ‘litigation-ready. You supply the detailed legal and business-specific knowledge and the client is better prepared for the “evil day” of litigation’.

The challenge for law firms which can offer a range of e-disclosure services, says Nina Barakzai, immediate past chair of the Commerce & Industry Group and group head of data protection and privacy at Sky, is to provide it ‘at a cost that will not make an in-house counsel balk, or, more likely, prompt them to search for ways to do this kind of activity with the remote teams themselves’.

She says networks such as the C&I Group and In House Counsel Worldwide are good sources of recommendations to find the best service providers. ‘But,’ she notes, ‘in any market, the service is only ever as good as the instructions. It would be unrealistic to say the market is perfect. The best service comes from collaboration between all parts of the supply chain who understand each other’s positions and capabilities. There is nothing worse than working under strict time and regulatory pressures, only to find part way through that what one party expected to happen bears no relation to what the other party thought.’  

Vijay Rathour, vice president of Stroz Friedberg, a computer forensic and e-disclosure technical services provider, says it is seeing more and better instructions coming directly from general counsel who may have ‘fewer legal reasons to be experienced at this work than private practice lawyers, but more commercial reason to be so’.

Looking at procurement from a provider perspective, he says: ‘If you give us clarity on expected data volumes, breadth, complexity and so on, we’ll reciprocate with greater clarity on cost and timeframes.’

But ‘almost every matter goes off plan,’ he says. ‘Many projects go bigger than expected, scopes broaden, keywords get increasingly layered on top of each other, so establishing a dialogue between the parties is important, rather than everyone being at loggerheads.’

The ‘healthiest’ matters he has worked on are where the lawyers have allowed the experts to sit with them and work through a pragmatic protocol. ‘In this way messy, complex projects can be boiled down to: “I’ve got 5% of this mound of data that’s actually relevant to these proceedings – here’s why… if you trust me, show me yours and let’s use that as the basis for our mutual disclosure”.’

In June, Herbert Smith Freehills announced it was creating a global team of 240 lawyers, legal assistants and technologists to offer a range of alternative legal services including document review – with the option to use predictive coding.

The team combines the firm’s Belfast office, which opened in 2011 and has a large team of solicitors and law graduates undertaking intensive document review, with its Australian information logistics team and its London-based disputes data management service, which bridges the gap between the law, practice and technology.

McGregor explains: ‘Belfast provides a nearshore solution for clients and means we can retain close sight of the quality. If you separate the individual pieces of our offering, you could outsource the review to a third party more cheaply, on paper, than Belfast, but we wanted to keep the whole offering risk-wrapped by the firm.

‘It’s about the space in between – the cost of going back and forth between you and the reviewers and making sure the knowledge of the person reading the documents isn’t lost.’

It is a big investment, she acknowledges, adding that the need for technology could become a barrier for smaller firms: ‘Will it lead to more consolidation? Potentially.’

Another practical implication, particularly for large-scale e-disclosure exercises, may be increased engagement between parties before disclosure about how to deal with privileged data which is inadvertently disclosed.

Alongside the TCC e-disclosure protocol ‘clawback’ provision, parties will look to the recent Court of Appeal decision in Rawlinson and Hunter, which says the test is whether it is obvious that the privileged document was disclosed by mistake.

Practitioners flag up other potentially risky areas such as metadata – essentially ‘data about data’, such as when an email or document was written and who amended it – and track changes in Word documents, both of which can contain privileged information.

Another major concern is security. Don Randall is a senior consultant with Bivonas Law, a niche City firm specialising in high-value and complex disputes. Former deputy head of the City of London Police fraud squad, Randall was the Bank of England’s first chief information security officer.

‘Law firms probably hold more commercially sensitive information than other industries, and the SRA places the responsibility for that data firmly in your hands,’ he says. ‘There is a tendency to focus on how the data is going to be processed; on the quality of the end-product. But of equal importance is who will have access to this data? Where and how will the data be held? Will it enter another jurisdiction, such as the US? You must be satisfied that the other side has in place the requisite cybersecurity and risk management systems to protect your client’s data.’

Rathour says future developments will include emotional sentiment analysis,  which will pick out documents that show X was angry with Y when they wrote it; better capabilities to turn unstructured or ‘messy’ data into sense; and automated language translation of text and audio.

But will the focus on increasingly intelligent software lead to the doomsday scenario where lawyers’ jobs are taken over by robots?

Films play on this fear, says McGregor, but machines do not replace people: ‘I am a technology optimist. I don’t think future developments will result in fewer jobs or less work. We are a service industry and we can distinguish ourselves in the market by understanding the pressures general counsel are under and helping them be more efficient.’

Grania Langdon-Down is a freelance journalist