eDiscovery costs can be huge, making up 20 to 50 percent of all costs in federal civil litigation, according to some estimates. In some cases, the discovery process can end up costing clients hundreds of thousands, even millions of dollars. A notable example is the long-running patent litigation between Samsung and Apple, where Samsung, in a single case, wracked up over $13 million in discovery costs -- not including its document review fees, which generally dwarf other discovery expenses.
But it’s not just major disputes that can leave litigants with gargantuan discovery bills. Discovery surrounding "Bridgegate" has cost New Jersey over $2 million. In an appeal currently before the Eighth Circuit, H&R Block is seeking to recover $250,000 in eDiscovery costs, out of a total of more than $450,000 -- for a case that was decided on summary judgement.
Here’s a quick, cautionary look at how parties ended up with such massive bills.
Twenty-Cent Fees Soon Add Up to Eight-Figure Bills
Samsung’s million-dollar eDiscovery costs are some of the most stunning we’ve seen. What’s more, those costs were accrued during one of the smaller patent lawsuits between the two phone giants, Apple v. Samsung II, in which Samsung was awarded a mere $158,000 for its offensive patent claims.
Before that award, Samsung paid $13,100,960.35 to its eDiscovery vendors for 20 months of discovery work. That discovery process involved the collection and processing of more than 11 million documents, or about 3.6 terabytes of data. Of that, only about 8 percent, or 880,000 documents, were produced.
The vast majority of the $13 million Samsung spent on discovery costs went to data processing, production, and online hosting. The exact breakdown of costs isn’t clear, as Samsung’s vendor invoices were redacted (you can see all 399 pages of them here). To reiterate, that $13 million doesn’t account for the cost of document review, either, which by some estimates makes up to 80 percent of all discovery costs. But it does include:
- Hundreds of thousands of dollars in license fees, charged per gigabyte
- More than 1,500 individual uploading charges, ranging from $0.20 to $1306
- Thousands of dollars in reviewer training
- The occasional $20,000-plus “hibernated subcollection fee,” whatever that might be.
Add it all up and you’re on the fast track to millions in eDiscovery fees -- about $82 for every dollar awarded by the trial court, to put it in perspective.
While Samsung’s litigation pitted two of the world’s biggest companies against each other in a major patent battle, the H&R Block case represents litigation of a different nature. That case, Perras v. H&R Block, Inc., involved a putative consumer class action suit over a $4 “compliance fee” charged during tax preparation services.
The H&R Block dispute is important for two main reasons. First, in its appeal to the Eighth Circuit, H&R Block is arguing that it is entitled to recover, under 28 U.S.C. § 1920(4), all costs of discovery related to data duplication -- essentially, all processing fees. That would entitle it to much more than the $11,175.70 in discovery costs awarded by the district court. If the Eighth Circuit agrees with H&R Block's interpretation of the law, the case could also significantly shift how discovery burdens are apportioned. (We take a deep dive into the issues raised by the appeal here.)
But more importantly, for our purposes here, H&R Block’s attempts to recover costs offer another glimpse into the often opaque world of eDiscovery billing. Seeing eDiscovery costs is rare; such invoices usually only see the light of day, as they have here, during litigation over cost recovery. That lack of transparency often leaves buyers of eDiscovery software services at a significant disadvantage, as they’re unsure of how that $550-per-gigabyte fee for metadata extraction compares to the rest of the market.
It also leaves judges in the dark as to the true costs of eDiscovery, which can complicate proportionality decisions, where judges must make cost-benefit analyses as to the amount of discovery to be done. If eDiscovery costs are by and large unknown, how do you make assessments as to the validity of parties’ assertions regarding these costs?
If eDiscovery costs are by and large unknown, how do you make assessments as to the validity of parties’ assertions regarding these costs?
So how do H&R Block’s discovery costs look?
The tax preparation company says it expended more than $450,000 to process, host, and Bates stamp hundreds of thousands of pages over a two year period. The company paid one vendor $32,000 for pre-processing / data filtering (at $75 per gigabyte) and TIFF processing with metadata extraction ($450 to $550 per gigabyte).
It paid a second eDiscovery company $418,000 for hosting ($25/GB), data ingestion ($100/GB), data enrichment ($250/GB), and Bates branding ($0.015/page). Other costs included $350/GB for “data loading: enrichment only” and $250/hour for project management.
Here are the costs from one exemplary invoice:
|Data Ingestion (per GB)||779.688||100.00||77,968.80|
|Data Enrichment (per GB)||46.856||250.00||11,714.00|
|Hosting (per GB)||341.79||25.00||8,544.75|
|Nearline Storage (per GB)||784.77||25.00||19,619.25|
|TIFF Creation (per page)||44||0.02||0.88|
|Project Management (per hour)||21.4||250.00||5,350.00|
Big eDiscovery Bills Aren’t a Thing of the Past
There’s an important caveat with the above numbers, of course. The costs in the Samsung and H&R Block cases were incurred years ago. The first invoices in the H&R Block litigation were issued in May of 2012, for example.
If we want a glimpse at more current costs, though, we just need to look to New Jersey. The Garden State offers us another, much more recent example of the huge costs of eDiscovery.
By last February, the state had paid out nearly $2.3 million in eDiscovery costs stemming from the investigation into the closure of lanes on the George Washington Bridge -- a.k.a. “Bridgegate.”
That total included nearly $1 million for data hosting, over $33,000 for PDF generation, charged at $0.02 per page, and $400 each for 19 two-terabyte hard drives. (The savvy consumer can pick up similar hard drives for under $80 at Staples today.)
Here’s a more complete breakdown:
|Unit Cost||Total Units||Total Cost|
|Data Hosting (GB/month)||$30||31,153||$934,590|
|Standard Processing (per GB)||$250||431||$107,750|
|Bates Stamping (per page)||$0.01||358,145||$3,581.45|
|Branding and Watermarking (per page)||$0.01||1,353,581||$13,535.81|
|Hard Drives (each)|
|Load File Delivery (each)||$75||13||$975|
|Native File Delivery (each)||$350||30||$10,500|
|PDF Generation (per page)||$0.02||1,650,766||$33,015.32|
|TIFF Generation (per page)||$0.03||191,271||$5,738.13|
|Compact Discs (each)||$35||4||$140|
|Thumb Drives (each)||$65||2||$130|
(Costs for travel, postage, software and other service fees, including data ingestion, are not included above.)
These costs are significant not just because New Jersey is getting charged $400 for a hard drive or $35 for a single CD, but because, unlike Samsung and H&R Block’s costs, these costs are recent.
Thus, while costs may have dropped since Samsung and H&R Block’s litigation began over five years ago, New Jersey’s bill of costs shows us that million-dollar eDiscovery bills are still regrettably common.