The eDiscovery Airline Problem: Why Data Hosting Became a Commodity
Private equity kept seeing a technology business. What they were actually buying was something much older and much harder to scale.
A Note on Perspective
I came to discovery before anyone had added the “e.” As a young associate, my introduction to the work was a warehouse, pallets of paper boxes, and a room full of colleagues doing by hand what software would eventually do faster and cheaper. I have been in eDiscovery proper since 2008, which means I have had a front-row seat to nearly every phase of the industry’s transformation: the infrastructure buildout, the platform wars, the private equity wave, and the AI reckoning now underway. What follows is my attempt to name what I think I have been watching.
There is a sentence every eDiscovery professional has heard from a corporate client, usually delivered with the flat confidence of someone who has just Googled the spot price of an AWS gigabyte:
“Hosting is hosting. Why are we paying this much?”
The instinct is understandable. It is also correct, in the narrowest possible sense, and missing the point entirely. The eDiscovery industry spent twenty years building toward that conversation. Understanding why it happened, and why the confusion persists, tells you something important about what this business actually is.
How barriers to entry disappeared
Two decades ago, the eDiscovery market had real structural protection. Significant capital was required to compete: processing clusters, proprietary tools, massive storage, physical document conversion operations. The barriers were high enough that only a handful of vendors could play at scale, and pricing reflected that scarcity.
Then three forces converged and dismantled the architecture almost simultaneously.
Cloud infrastructure commoditized compute and storage. Anyone with AWS credits can now spin up the processing power that once required a purpose-built data center. Platforms like Relativity standardized the software stack, transforming what had been a source of differentiation into a licensed commodity that nearly every vendor runs on top of. And corporate legal departments, staffed with increasingly sophisticated operations professionals, learned to negotiate. They learned to ask the question the client above was asking: what exactly are we paying for here?
Once those conditions existed simultaneously, the economics became almost textbook. When supply is easy to create and differentiation is weak, price competition becomes the dominant strategy. The result is what happens to every commodity market: a race to the floor.
Data processing and hosting became airlines. Expertise became the cockpit.
Why private equity kept getting this wrong
The misreading has been remarkably consistent. Investors looked at the eDiscovery sector and saw a genuinely compelling narrative: massive data growth, accelerating regulatory pressure, rising litigation complexity, a corporate legal market with no choice but to spend. The demand side of the equation was, and remains, real.
The mistake was assuming that demand would translate into software margins.
It doesn’t, because the structural conditions required for software-style pricing are absent. Law firms switch vendors with relatively low friction. Procurement teams negotiate aggressively and have become better at it each year. The Relativity ecosystem flattens differentiation almost by design. And the largest matters, the ones that generate the most revenue, are episodic rather than subscription-like: they end, the relationship resets, and the next engagement goes back out for pricing.
The revenue stream behaves less like SaaS and more like consulting plus infrastructure rental. Consulting businesses have a ceiling. Revenue scales with people and judgment, not with software leverage.
The M&A consolidation cycle follows almost inevitably from this misread. Private equity acquires several vendors, consolidates infrastructure, attempts to cross-sell, targets margin improvement. Sometimes it works temporarily. But the underlying market structure hasn’t changed, so pricing pressure returns. That is why the industry has experienced waves of consolidation for twenty years. It is not a strategy failure so much as an equilibrium state: the predictable outcome of a commoditizing service industry absorbing capital that expected something else.
Where pricing power actually lives
The things clients will consistently pay for are not gigabytes. They are defensible processes, sound early case assessment, expert project management on complex matters, analytic judgment about what data actually matters, and testimony that holds up under scrutiny. In other words: risk mitigation and judgment.
Corporations and law firms do not pay premiums to store documents. They pay premiums to not lose cases or sanctions motions. That distinction is the entire business model, and it has been hiding in plain sight for years.
The firms that have built durable margins understand this intuitively. They do not compete on processing rates. They compete on the quality of the people who design the workflows, manage the matters, and stand behind the defensibility of the work product.
What AI changes, and what it doesn’t
AI tools create genuine uncertainty about which direction the chessboard moves next. Predictive review, generative search, and automated summarization could further commoditize the service layer, pushing more of the work into software and removing more of the human judgment that currently commands a premium. That is the pessimistic scenario for the professional services side of this industry.
The more likely outcome, in my view, is the opposite. Courts are not indifferent to process. They care deeply about defensibility: how decisions were made, what methodology was applied, whether the work product can withstand challenge. AI accelerates output but does not eliminate the need for an expert who can stand behind it. If anything, it increases the value of professionals who understand how to design and defend AI-assisted workflows in litigation.
The cockpit remains occupied. The plane just got faster.
The structural question this leaves open
The firms best positioned in this environment are not the ones chasing processing volume. They are the ones moving toward what might be called an intelligence model: becoming data analysis specialists for legal risk rather than discovery vendors. Investigations, compliance analytics, regulatory monitoring, insider-trading reviews. These engagements widen the moat and shift the conversation away from per-gigabyte pricing entirely.
The platform owners occupy a different and more durable position. Relativity has software economics because it owns the layer that everyone else runs on top of. Platform economics beat service economics in almost every industry, and this one is not an exception.
For everyone else competing in the middle of the market, the question is not how to win the commodity war. It is how to get out of it.