NSerio - Relativity Custom Development & Integration Experts | Legal Tech Solutions

How to Optimize Total Cost of Review Management Within Relativity

Written by CaseFlow Team | Nov 5, 2025 1:00:01 PM

You already know where review work can eat up time and money. What most teams need is a simple, steady way to shrink the pile, guide reviewers to the right documents, and keep hosting from creeping back after the case cools. The following plan is simple but fast to set up and respects real-world limits, such as short timelines and fixed budgets.

Start With Scope

Begin before you collect a single file. Sit with legal and the client, and write the real scope in two lines. Include who matters, which dates matter, and which systems matter. That simple act will save you more than any fancy tool later.

To make this concrete, think about a common pattern. For example, you collect ten mailboxes, two shared drives, and a chat export. Processing finishes, and you see 1.2 million items in review.

Now, switch to a smarter pattern.

You interview IT, trim three mailboxes that are noise, cut the shared drives to two key folders, and set a fair date range. You also exclude file types that never hold probative text for this case. Processing finishes, and now only 380,000 items remain for review. It’s the same team and the same issue tags, but it took far less time and far less spend.

Use a short intake checklist each time. You should know:

  • What claims and defenses set the dates and custodians
  • Which systems actually hold the likely hits
  • Which file types add risk if removed and which do not
  • Which sources may hold privilege and should be split out early
  • What will trigger a second collection so you are not over-collecting today

Small choices upstream remove thousands of documents downstream, where “people time” is the top cost driver.

Cut the Total Cost of Review at the Source

Now that the scope is set, protect it. Use a Repository workspace for early case assessment so you can process once, cull hard, and only push the right slice into a Review workspace. If an old Review workspace is sitting idle with a huge dataset, convert it to Repository mode, keep the coding history for reuse, and pay the lower rate for storage while review is paused.

Keep Staging clean. It’s short term, so move what you are not ready to process out, and delete what is no longer needed. Many teams pay overage fees here because no one is responsible for the cleanup. Give one person the job and set a weekly reminder.

When a Matter goes quiet, move the workspace to Cold Storage. If there is no live need, archive it with ARM. Tell the case team how long the retrieval may take so no one is surprised. This simple lifecycle keeps bills from drifting back up after you worked so hard to bring them down, which is the heart of lowering your cost of review.

Use Analytics to Drop the Total Cost of Review

Structured analytics is your first speed boost. Run email threading and work from the inclusive set. An inclusive email contains the full content of the thread, so you can skip the redundant replies and forwards that add nothing new. Build a saved search that returns only inclusive emails and unique attachments, then route that set to your first-pass queue. Keep the rest for QC or later issues.

Next, run near-duplicate analysis. Reviewers make faster and more consistent calls when they see like-with-like. Batch by similarity so a reviewer can carry a decision, with care, across a cluster. This reduces context switching and rework during QC.

Two small rules make the most of analytics:

  • Give full threads or near-duplicate clusters to the same reviewer, so one person has the context.
  • Avoid mixing languages or issues in a single batch, so the reviewer doesn’t need to reset their brain every few minutes.

Turn on Active Learning for Priority-First Review

Active Learning ranks documents by how likely they are to be responsive to a clear “yes” or “no” issue. Use it when your set is large and the team needs the most important documents first. Linear review is for when the set is tiny or the issue is too niche to train.

Set-up is simple: Pull a small richness sample to see how many likely positives exist, seed the model with clear “yes” and “no” examples, turn on continuous Active Learning, and then watch the learning curves in the Review Center.

Early on, you should see two to three times more responsive documents per hour than in a blind queue. As the model learns, the uncertain middle narrows and the low-value ranks grow. That is your cue to slow or even stop reviewing the bottom ranks and focus on QC instead.

After the main pass, use Active Learning for validation. Sample the high-uncertainty ranks, check privilege hotspots, and do an extra pass on any custodian that shows a higher miss rate.

Build Reviewer Speed and Keep Quality Tight

Good reviewers are fast when the work in front of them is clean and consistent. So, track a few simple weekly KPIs and tie them to staffing and coaching. Ensure that there are:

  • Docs per hour by person and by team
  • Overturn rate on QC
  • Agreement rates on overlap samples
  • A small precision and recall proxy, using spot checks from completed batches

If the docs per hour drop and overturns rise, check the batch design before you blame the reviewer. The fix is often to re-thread, re-run near-dups, or split mixed-issue batches.

Coach with facts, not guesses. Sit with a reviewer for fifteen minutes, watch one batch, and look for friction. Are they bouncing between issues in one batch? Are they re-reading for context because the thread is split? Do they lack clear examples for edge cases? Remove the friction, then measure again.

Analytics increase speed and reduce rework by eliminating the need for re-reading. Threading shows the complete story in fewer items. Near-dup grouping keeps similar documents together, so a reviewer can apply the same logic with fewer doubts. That stability reduces mistakes, which cuts QC churn, ultimately lowering the number of hours that contribute to your total cost of review.

Make Costs Visible Inside RelativityOne

Turn on a weekly “no surprises” ritual. Use Cost Explorer to look at hosting growth, user activity, and feature usage by Matter and by client. Ten minutes is enough. If a dataset grows fast, confirm the reason. If a workspace is idle, move it to Cold Storage. If user time spikes on a Matter, ask if the team needs help or if the scope has changed.

Attribution is key. Tag usage to the right Matter and workstream. This way, you can show a straight line from action to spend, which helps leaders approve the next move without debate.

Control Storage After Review

Storage is where costs slowly rise again if no one is watching. Set a simple policy and follow it every month:

  • Keep active reviews in a Review workspace.
  • Move paused Matters to Repository mode.
  • Move quiet workspaces to Cold Storage.
  • Archive truly closed Matters with ARM.
  • Run a monthly “storage delta” check.

Consider moving ARM archives to lower-cost storage, such as your own cloud bucket, to prevent long-term holding from impacting your budget. If the delta is up with no live reason, freeze non-essential uploads, plan archives, and clean Staging. This habit alone ensures months of savings.

Pricing Reality and Why Speed Matters

Vendors sell hosting and review data in tiers, and your contract sets the limits and the rates. While you cannot change the tiers in the middle of a case, you can change your exposure to them. Smaller sets, faster decisions, and clean storage states mean fewer gigabytes billed at higher rates and fewer people hours. That is how you protect the total cost of review while still hitting deadlines.

Set targets that you can see and share each week:

  • Percent reduction in reviewable docs after scoping and analytics
  • Inclusive-only rate for email sets, which shows how well you culled noise
  • Time to first decision from data load to the first coded doc, which shows how fast your engine starts

Playbooks That Your Team Can Reuse

Create these once, then use them on every Matter. Keep them in a shared folder so everyone can grab them without hunting:

  • Intake checklist. This keeps you from over-collecting and locks the real scope. You’ll know who, what dates, which systems, what to exclude, and when to collect more if needed.
  • Analytics runbook. This ensures clear settings for email threading, near-duplicate thresholds, and simple batching rules that match how your team works.
  • Active Learning quick-start. This is how to pull a sample, add seed examples, turn it on, and read the learning curves so you know when to push or pause.
  • Storage lifecycle SOP. This is when a workspace is active, when it moves to Cold Storage, when it archives with ARM, and who signs off at each step.

These playbooks save time at kickoff, cut mistakes during review, and help you finish clean.

Where Automation Fits In: CaseFlow

If your instance has many clients and Matters, the manual steps can eat up your day. This is where CaseFlow helps. It monitors your Relativity instance, finds inactive workspaces based on the rules that you set, and moves them to the correct state. For example, it can shift a workspace to Repository mode, place it in Cold Storage, or schedule archives.

It also warns stakeholders before any archive runs, so you do not take a case offline by mistake. When you need to act across many Matters at once, you can schedule all lifecycle actions in one place.

In plain terms, CaseFlow keeps the lifecycle clean without constant hand-holding. That frees your team to focus on review speed and quality, which is where savings grow fastest.

Try It on a Live Matter Today

Pick one active case, run threading and near-dups, and turn on a small Active Learning project for the main issue. Move any idle workspace to the right storage state, measure two weeks of results, and share them. You will see fewer documents, a faster review of the ones that matter, steadier QC, and lower spend. Then, make this your default way of working.

Stop letting inefficient processes drain your firm's profitability. Every hour spent on manual case management and disorganized workflows is an hour that could be spent generating revenue. Your clients deserve better, and so does your bottom line. CaseFlow delivers streamlined case management, automated workflows, and real-time insights that free your team to focus on what matters most: practicing law profitably. Ready to drive real efficiency and savings? Discover how CaseFlow can transform your firm's operations.