Per-GB pricing, review markups, and vendor secrecy: how law firms, service providers, and software platforms each profit from opaque eDiscovery pricing — and what a public benchmark would change.
The invoice on the litigation support manager's desk was not itemized. It was not, technically, an invoice at all — more of a settlement figure: a single number, six figures, for "eDiscovery services" on a mid-size commercial dispute. She approved it, as she had approved a hundred like it, because the alternative was interrogating a line item no one in her building could decode. The vendor had processed the data. The lawyers had reviewed it. The money had gone somewhere. Nobody, she later admitted, could have told you where.
That single number is the business model of an entire supply chain. eDiscovery's opacity is not a bug. It is a condition three parties separately profit from — vendors, service providers, and law firms — each extracting a margin from the fog. Understanding who gains from the dark is the first step to seeing what daylight would cost them.
Start with the software platforms. Relativity — still the dominant large-matter platform — prices through hosting partners and subscriptions with per-matter totals that swing from $50,000 to $500,000, a tenfold range no two quotes will reconcile. When it launched aiR, its generative-AI review product, it built billing on per-document "aiR Units" — and, as the University of Waterloo and Elevate researchers documented, declined to disclose the allowance in subscriptions or the overage rates. A market that cannot see the price cannot comparison-shop. That is not a documentation accident. It is a pricing strategy.
Emergen Research's sizing explains why: services — processing, hosting, managed review — hold 61.4 percent of the eDiscovery market in 2025, roughly $9 billion of the $14.82 billion total. These are the metered businesses. Logikcull's calculator showed the same 25-gigabyte matter priced at $3,125 on a flat per-GB rate and $28,875 through a traditional provider — a 9.2x spread on identical work, with the difference attributable to pricing structure, not quality. When DISCO moved in February 2026 to a single all-inclusive per-gigabyte rate, it was effectively conceding the point: the metered bundle had become harder to defend than a flat rate customers could check.
The second tier profits more quietly. Thomson Reuters' Alternative Legal Services Providers 2025 report puts the ALSP market at $28.5 billion — an 18 percent compound annual growth rate from 2021 to 2023 — with e-discovery and document review the leading segment at roughly 46 percent of ALSP revenue. These are firms like UnitedLex, Elevate, and Axiom: they buy platform capacity at wholesale, staff reviews with contract attorneys and offshore teams, and sell the packaged outcome at retail.
Fifty-seven percent of corporate law departments now use ALSPs, mostly to escape law-firm review rates. But the ALSPs' own rates are negotiated bilaterally, matter by matter, with no published reference anywhere. A managed review billed at $60 an hour might cost the provider $25 in blended labor. The client has no way to know. Laura Clayton McDonnell, president of Thomson Reuters' corporates business, called the current moment a transformation driven by generative AI. Notice the direction: AI is making them faster, which widens internal margins unless client pricing falls with it. Opacity lets the provider keep the gain instead of passing it on.
The third beneficiary is the law firm itself. Am Law 100 firms grossed a record $178.95 billion in 2025, up 13 percent, on billing-rate growth of 9.6 percent — with productivity flat at 1,589 hours per lawyer. When rates rise and hours don't, the firm needs somewhere to book the spread. Disbursements — vendors, ALSPs, court reporters — are that somewhere.
The mechanics are mundane and the math is old. A firm pays a subscription or per-matter vendor fee, marks it up as a "litigation support" charge, and passes it to the client. Logikcull's own blog walked firms through a clean version: $60,000 in annual software spend across 60 matters becomes a $1,000-per-matter client fee, pure pass-through. The clean version is not the common one. The common one is a vendor invoice buried inside a monthly legal bill, marked up or simply unexamined, approved by a client who has no independent number to check it against. The Thomson Reuters report found law firms with their own ALSP affiliates far more likely to use independent ALSPs too — 62 percent versus 23 percent — because once the firm controls the pipeline, every dollar of opacity is a dollar of optionality.
Nobody in this chain is defrauding anyone. That is what makes it durable. Each margin is individually defensible: the platform invested in the product, the provider managed the review, the firm bore the risk. Opacity doesn't require villainy. It just requires that nobody publish the price.
Here is the counterweight, and it is real. Clients are not as helpless as the model assumes. Corporate legal departments have been writing billing guidelines for two decades: the modern ones cap disbursement markups, demand preferred-vendor rates, and push fixed-fee arrangements that move pricing risk back onto the firm. And 40 percent of corporate respondents in the Thomson Reuters report said GenAI-leading ALSPs are more attractive — which reads as interest in the technology, but functions as price pressure: the client who believes review now costs half as much will not pay last year's rate.
And the vendors are competing, whatever their rate cards hide. Logikcull's $40 flat per-GB pitch exists because it wins deals. DISCO's all-in per-GB bundle exists because clients asked for one number. Everlaw built its growth on the argument that review should feel close to free. The market is not a cartel. It is a set of firms exploiting an information asymmetry that competition is slowly eroding — and every vendor that publishes a real price makes the asymmetry a little harder to maintain.
But bilateral competition is not a benchmark. A GC who beats one vendor down 20 percent has won a negotiation, not a market. Without a shared public reference — real processing rates, real hosting fees, real review costs, by matter size — every deal starts from the vendor's number. The information asymmetry survives every individual discount.
A public benchmark would rearrange all three margins at once. Vendors would lose the uncheckable unit — the undisclosed aiR allowance, the $135-per-gigabyte quote nobody can compare — and compete on total cost of matter. Service providers would see their wholesale-retail spread compressed toward transparent brokerage margins, exactly what happened to travel agents and stockbrokers. Law firms would find disbursement markups visible in a way billing guidelines never achieved, because a benchmark turns "trust us" into an auditable claim.
None of the three will build this voluntarily. Why would they? The fog is worth billions — $42.1 billion a year in review spend alone, per the ABA estimate, flowing through pricing no one can verify. The benchmark, if it comes, will come from outside the supply chain: from a buyer, a publisher, or a data business with no margin to protect. From someone whose only product is the number itself.
The litigation support manager approved the invoice; the matter needed to move. But she kept a copy, and she has started a spreadsheet of her own: vendor, matter size, price, outcome. Row by row, she is building the thing the industry never would. Billions of dollars of margin are riding on whether she has to do it alone.
Secure transfer via Escrow.com · info@primedomaincapital.com