eDiscovery pricing models compared: what each one meters
Four models are in use, and they differ in what they count rather than in what they charge. A meter tied to volume moves when data moves, one tied to headcount moves when the team moves, and one tied to hours moves when the vendor moves. Which meter you are on decides whether a bill can be forecast.
Updated October 6, 2026
Four pricing models are in general use in eDiscovery: per-gigabyte, flat rate, per-user, and hybrid arrangements that combine them. They differ in what they count, not in how much they charge, and that is the distinction worth carrying into a negotiation. A meter tied to volume moves when data moves. A meter tied to headcount moves when the review team moves. A meter tied to hours moves when the provider’s involvement moves. Whether a bill can be forecast depends on which of those you are attached to.
What are the main eDiscovery pricing models?
Four, and the difference between them is the meter rather than the number on it.
| Model | What it meters | What moves the bill | What stays fixed | Who it suits | Check before signing |
|---|---|---|---|---|---|
| Per-GB | Data volume held, processed and produced | New custodians, opposing productions, matter duration | The unit rate | Occasional matters, small and short | Which volume is measured, and when |
| Flat rate | Time, for a bounded environment | The term, and nothing else | Volume and headcount within the boundary | Steady matter load above the crossover | Where the boundary sits and the overage rate |
| Per-user | Headcount holding access | Staffing the review team up or down | Data volume | Small stable teams on large data | Named or concurrent, and how seats are released |
| Hybrid | Two or more of the above at once | Whichever component is the variable one | Whichever component is the fixed one | Most real contracts | Which line dominates at your volumes |
The billing bases behind that table are not invented for it. The Fee Field Guide dataset classifies each of its nineteen fees by mechanism, and the same handful recurs: volume-based one-time, volume-based recurring, headcount-based, hourly, feature-based, threshold-based, per-page and per-event. The models are combinations of those, which is why the last row is so common.
What moves the bill under each model?
The question a price sheet never answers, because it describes a rate at a moment rather than a bill over a matter.
Under a volume meter, the other side moves your bill. Per-GB pricing charges on data held, so a production you did not ask for and cannot refuse raises the recurring charge for every remaining month of the matter. Duration compounds it: processing is charged once and hosting every month, so the forecast depends on a date nobody controls.
Under a headcount meter, your own staffing moves it. Per-user pricing is indifferent to data and sensitive to people, so adding contract reviewers for a document-heavy stretch raises the bill precisely when the team is busiest.
Under a time meter, nothing moves it inside the boundary. Flat-rate pricing trades the upside for that: the provider absorbs variance and prices it in, so light matters subsidize heavy ones.
Under an hourly component, the provider moves it. Project management, collection and support are metered by vendor effort rather than by anything on your side, which is why they are the lines clients most often fail to forecast.
The observation that legacy pricing meters volume and headcount rather than work performed is set out on the Fee Field Guide, fee by fee. This page takes that as read and asks the next question instead.
Are eDiscovery pricing models actually changing?
Measurably, though from a small base, and unevenly across the pipeline.
The Winter 2026 eDiscovery Pricing Survey, conducted by ComplexDiscovery with EDRM across December 2025 through February 2026, closed on a cohort of 53 respondents. Among them, user licensing was described as “in an active state of structural transition”, with 34.0% reporting alternative pricing models, the highest alternative-model proportion in that survey’s processing and hosting section. The direction named is “away from traditional per-seat licensing toward enterprise agreements, volume tiers, and managed service arrangements that bundle access costs into broader contract structures”.
Processing shows the same drift, and shows it growing further down the pipeline: alternative models account for 18.9% of responses at ingestion and 22.6% at completion. Predictive coding and technology-assisted review have moved furthest of all, at 35.8%, the highest such proportion of any per-gigabyte question in the survey, with the analysis concluding that standalone per-gigabyte TAR billing “is becoming the exception”.
Two cautions belong with those figures. Fifty-three respondents is a small base, and every proportion above carries it. And “alternative pricing model” is a residual survey category rather than a named structure, so the numbers show movement away from the traditional meter without saying what replaced it.
Which pricing model suits which firm?
The honest answer is that it depends on two numbers you already have and one you can estimate.
The numbers you have are how much data you hold across all active matters and how many people need access. The one you estimate is how long matters stay open, and it is the one that decides more than the rate does, because a recurring meter multiplies by months.
A firm with occasional matters that close quickly is usually better on a metered model, and paying a fixed rate for headroom it does not use would be a poor trade. A firm carrying several matters at once for years is the opposite case, and the crossover between them is a volume rather than a date. When per-GB pricing stops making sense works through that threshold in detail, and the cost calculator computes it on your own figures instead of an assumed profile.
What should not decide it is the headline rate, because the four models quote different units. Comparing a per-gigabyte rate against a monthly platform fee is a category error until both are expressed as a total over the same matter at the same volume for the same number of months.
What should you check before you sign?
Four questions, one per model, and they are all about the unit rather than the number.
On a volume meter: which volume, measured when. Raw collected data includes everything culling and deduplication will discard. Ask whether reduction happens before or after the measurement, and whether the same volume is used for processing, hosting and production.
On a fixed rate: where the boundary sits. Every flat rate is bounded by something, and a numeric cap with a published overage rate is still forecastable where a fair-use clause is not.
On a seat meter: named or concurrent, and how a seat is released. The same headline rate produces very different totals, and an unreleased seat bills after its holder has left the matter.
On a hybrid: which component varies. Negotiating the fixed floor changes a number you could always predict. Capping or reducing the variable component changes the part you cannot.
Each of those maps onto lines the Fee Field Guide decodes individually, and the glossary defines the terms a price sheet will assume you know.
How this works on a flat rate
Kavarn, the flat-rate eDiscovery service Konvergence Inc operates, is one instance of the third row of that table: a published monthly rate per private instance, with no per-gigabyte processing or hosting charge and no per-user fee. The boundaries are the instance size and stated allowances, and the overage rates are published rather than negotiated case by case, which is the distinction the flat-rate section above calls the difference between a numeric cap and a conversation.
That is a structural claim, not a claim to be cheapest. Below the crossover a metered model costs less, and the pricing page publishes every figure so the comparison can be made rather than asserted. See what the platform includes, or book a walkthrough.
Sources
- ComplexDiscovery, A Complete Analysis of the Winter 2026 eDiscovery Pricing Survey, conducted with EDRM across December 2025 through February 2026 on a cohort of 53 respondents. The survey states it is “not sponsored by, or designed to favor, any vendor, platform, or service category”.
- Billing mechanisms behind the comparison table: the
typclassification in the Fee Field Guide dataset, covering nineteen fees across five categories. - The Sedona Conference, Glossary: eDiscovery & Digital Information Management, for the underlying vocabulary.
- Model definitions: the per-GB, flat-rate, per-user and hybrid glossary entries.
eDiscovery at one flat monthly rate
$3,250 a month on a 12-month agreement. No per-user fees and no per-GB hosting charges.
Common questions
What is an alternative pricing model?
The phrase comes from pricing surveys, where it is the response option for respondents whose arrangement does not fit the per-gigabyte or per-seat bands being asked about. It is a residual category rather than a named structure, so it covers flat rates, subscriptions, enterprise agreements and bundles together. A rising share of respondents choosing it indicates movement away from the band, not agreement about where they went.
Can you negotiate a different pricing model?
More readily than a different rate, in many cases, because a model change can suit both sides. The lever is usually scope: a longer term, a committed volume, or several matters together. The four points listed above under what to check before signing are the ones worth raising, and it helps to ask which structures a provider already supports rather than asking them to invent one.
Why do quotes from two providers not compare?
Because the unit differs before the number does. One quote may meter raw collected volume and another the promoted set, one may count named seats and another concurrent sessions, and one may bundle production while another itemizes it. Normalizing the unit has to happen before any comparison of rates means anything, which is why the Fee Field Guide is organized by fee rather than by vendor.
Which model is cheapest for a small matter?
Usually a metered one, because a fixed rate charges a full period whatever the size of the matter. The crossover is a volume rather than a date, and it moves with how long matters stay open and how many run at once. The cost calculator computes it on your own figures rather than on an assumed profile.
Does a bundled price save money?
It is not measurable from the price sheet, which is the honest answer. Bundling reduces per-matter cost visibility, and whether it is a net advantage depends on the volume actually used against the scope contracted for. A bundle can be cheaper and harder to audit at the same time, and the second of those is the reason to keep track of usage independently.