Google Confuses Search & eDiscovery; May Cost Billions

Community Topic(s):

Keywords: ediscovery, google, search, Oracle, email

Current Rating:
(0 ratings)

 

Google is a company run by engineers.  Steven Levy’s great book “In the Plex” (http://books.google.com/books/about/In_the_Plex.html?id=V1u1f8sv3k8C) documents how engineers are the rock stars of one of the world’s most successful companies.  But Google’s culture assumes that everything is best addressed as a technical problem and the company is just now learning that sometimes technology isn’t perfect.

The real irony is that Google is likely to learn from a multi-billion dollar mistake related to, of all things, search.

Oracle is in a lawsuit with Google over license fees associated with Java.  Oracle acquired Java when it purchased Sun and most companies licensed Java.  Google did not and it forms a critical part of Google’s popular Android operating system.  Google gives Android away for free as a way to drive mobile phone users to their ecosystem and extremely profitable searches and advertising.  Given the success of Android, a loss to Oracle is expected to result in billions of dollars in damages.

Computerworld’s James Niccolai wrote a great article going over the particulars (http://www.computerworlduk.com/in-depth/it-business/3314548/how-google-was-tripped-up-by-a-bad-search-in-oracle-lawsuit/) but the bottom line is that when Google responded to Oracle’s discovery request, they used search technologies and produced an email that was particularly incriminating.  (Google argues that the email is out of context and was part of a legal discussion, not a business decision.)  Google hasn’t disclosed the specifics of what happened and has only said their used their tools for the search.

The email had multiple drafts, but only the last one was labeled as “Attorney Work Product” and would have been subject to being withheld from the discovery request.  That means it should never have been produced.  Already the judge in the case has told the Google lawyers that the email is going to be particularly damaging as he has encouraged both sides to negotiate a settlement.

Normally the documents and emails being released to the other side are manually reviewed.  eDiscovery software vendors do a great job with their manual review tools to facilitate this process.  Somehow this didn’t happen and Oracle was given the email.  Normally a miss results in an unimportant document/email being provided, but in this case, the email basically says Google knew it was infringing on the Oracle license and should consider paying.  It is particularly damaging to Google’s case.

Engineers tend to assume that the technology is the answer, but lawyers realize that human review is necessary.  But even then the lawyers can make mistakes by not knowing how the technology works.  One recent lawsuit collected 250,000 pages of electronic documents.  The attorneys did the review and marked 159 as “privileged” so they couldn’t be released.  But the attorneys didn’t realize that marking the documents with an electronic flag wasn’t enough:  the eDiscovery software still copied all of the documents when it gave them to opposing counsel.  The Court ruled that the attorneys process fell “well short of what we would expect for an adequate account of the review procedure.”  The Court ultimately decided that it was the attorneys’ fault for not knowing how the software worked.  More information at:  http://www.rcalaw.com/E-Discovery/Document-Control-Blog/Misunderstanding-of-Vendor-Software-Leads-to-Waiver-of-Privilege.php

When it is all over Google may end with the most expensive eDiscovery mistake in legal history.  Ironic for a company famously committed to “do no evil” and who should know more about the technologies than anyone else.  But Google is smart and can afford to write the check.  Just as Microsoft famously learned from their incriminating emails in their anti-trust lawsuit that led to the company thinking about unstructured data and, in part, leading to the development of SharePoint, look for Google to pay more attention to the eDiscovery market and the human part of search.

 

For more, be sure to follow me on Twitter:  @ElamGuru

Report

Rate Post

You need to log in to rate blog posts. Click here to login.

Add a Comment

You need to log in to post messages. Click here to login.

Comments

Johannes Scholtes

eDiscovery search is very different from Web and portal search

Dan,

Great contribution! I fully agree with your observations. Last year June, I wrote about the limitations of Google's technology for eDiscovery: http://zylab.wordpress.com/2010/06/25/do-we-understand-the-benefits-and-limitations-of-traditional-web-search-engines-such-as-google-when-you-use-their-appliances-and-technology-in-house-for-mission-critical-applications/

It still wonders me why companies are considering Google appliances as eDiscovery solutions and why Google representatives market them a such. Let this case be a warning to all of us that the type of search needed for eDiscovery is very different. Not only does one need a different type of search technology, one also needs clear, transparent, (court)proven methodologies and best practics for defensibility in court!
Report
Was this helpful? Yes No
Reply

Jiri Janecek

Engineers run Google - Lawyers run Oracle and Microsoft - the result - look at market share

"Given the success of Android, a loss to Oracle is expected to result in billions of dollars in damages"
Please learn about OpenJDK.java.net - Dalvik VM.
Android is evolving OS ... Oracle contribution to the Software is high cost and litigation.
Report
Was this helpful? Yes No
Reply
Dan Elam

Tough Argument

Jiri,

I'm a Google fan. I have publicly stated they they have the best business model on the planet and only Facebook may ultimately be able to rival them. That being said, Google, in many ways, looks a lot like Microsoft used to look. In 1991 Gates told Microsoft employees that software patents basically worked against small companies and favored large companies who could "take as much of our profits as they want."

Today, Microsoft makes at least $5 from most Android smartphones thanks to patent licenses. Google makes $0. But Google makes a lot more than $5 from every phone from their search and ads. In fact, Google testified recently at an anti-trust hearing that they make money from iPhones than they do from Android devices because of search and ads.

Oracle is a bit different. They paid $7.4B to purchase Sun and their crown jewel of Java. Everyone else has had to pay to license Java, but Google, it appears, chose not to pay. The entire industry was paying Sun only about $100M/year in 2007. The costs weren't prohibitive - unless you were using a free model. Google could still have chosen to pay the license fee - or make it so that Android didn't have all the Java features that required a license. It is hard to blame Oracle for paying a lot of money to acquire a technology that everyone was licensing and then expect them to give Google the same software for free.

The issue of software patents is a contentious one, especially for Google as it was for Microsoft. Google is losing virtually every patent/trademark lawsuit they are involved in. Google has gotten smart and made the offer for Motorola Mobility in order to get a patent portfolio to defend against Apple and Nokia/Microsoft. They bid on the Nortel patent portfolio and turned down joining the Microsoft/Apple/etc consortium. They also purchased a $1B IBM patent group. All of that is relatively recent as Google has "grown up" as a company. And Google is starting to patent their own ideas, presumably to protect their hard-earned R&D from others profiting from their innovation.

The reason for patents is to encourage innovation. I've spent a lot of time in the ECM industry helping to bust patents. But real innovation takes R&D and that is almost always funded based on the prospect of being able to enjoy a competitive advantage. If the only advantage is low cost labor or materials to produce a new product, the pace of innovation slows considerably. Patent reform is a hot topic across the world, but unless everything goes open source, licenses for patents are likely to remain common as companies look to use the "best idea" instead of merely one that works.
Report
Was this helpful? Yes No
Reply

This post and comment(s) reflect the personal perspectives of community members, and not necessarily those of their employers or of AIIM International