Understanding the minimum coherence score in assisted review for Relativity project teams.

Coherence as a Compass: Understanding the 70 Threshold in Assisted Review on Relativity

Let me ask you a quick question: when a team sifts through a massive pile of documents, how do you know everyone is still reading the same map? That’s where a coherence score comes in. In assisted review on Relativity, the coherence score is the little compass that signals whether reviewers are on the same wavelength. It’s not about who has the loudest opinion; it’s about shared judgment, consistent criteria, and trustworthy results. If you’re curious how this works in real-world projects, this guide breaks down what coherence means, why the number 70 matters, and what teams can do to keep reviews reliable without slowing everything to a crawl.

What coherence actually is, and why it matters

Think of coherence as a measure of agreement among reviewers. In a large review task, different people might interpret a document differently—scope, relevance, privilege, responsiveness, or even how to code a given issue. A good coherence score tells you that the judgments are not all over the map; they’re aligned around a shared understanding of the rules and the criteria at hand.

This matters for a couple of reasons. First, it protects the quality of decisions. If reviewers disagree too much, you risk missing important items or wasting time on items that aren’t relevant. Second, it helps keep the process efficient. When people share a common frame of reference, you don’t spend hours re-litigating the same issues or chasing contradictory conclusions. In a project setting—where timelines can be tight and stakes are high—that alignment becomes twice as valuable.

Why 70 isn’t random—it’s a practical balance

On Relativity, the minimum coherence score used in assisted review is set at 70. You might wonder: why not aim higher or take a lighter touch? Here’s the gist.

• A 70 signals a solid level of consensus. It means reviewers are generally agreeing on key decisions, and the team can proceed with confidence.

• It’s a practical floor, not a ceiling. Pushing for a much higher score can slow things down. Calibrating every decision to an extremely tight standard often yields diminishing returns, especially when large data sets are involved.

• It preserves speed without sacrificing trust. In many projects, you want reliable results, but you also want to move. A 70 threshold strikes that balance: enough agreement to be credible, enough flexibility to keep momentum.

In plain terms, it’s like a team sport: you want a good level of collective buying, not a single superstar carrying the whole game.

What happens if you’re just under the line?

If a coherence score sits below 70, that’s a signal to pause and recalibrate. You may see more disagreements, which can translate into longer review cycles, inconsistent coding, or divergent conclusions about what’s important. The risk is not only a slower process; it’s the potential for incorrect inclusions or missed items that matter.

That doesn’t mean you throw out the data and start over, though. It’s a cue to check your review guidelines, ensure everyone has the same definitions, and run a quick calibration.

Balancing act: when is higher really better?

You might guess that a higher coherence score is always better. In practice, though, there’s a nuanced balance to strike. Pushing the threshold higher than 70 can improve reliability, but it can also slow down the workflow—especially with huge document sets or tight deadlines. The key is to align the threshold with project risk, client requirements, and the nature of the data.

For instance, if the data are highly sensitive or involve complex privilege questions, a few extra calibration rounds to nudge coherence above 70 may be worth it. If, on the other hand, time is of the essence and the risk of material errors is comparatively low, keeping the line at 70 often makes more sense. It’s about knowing the landscape and choosing a pace you can sustain.

Ways teams keep coherence strong in practice

So how do groups keep that 70-point line from slipping? A few habits go a long way.

• Clear criteria and coding guidelines. Before any review starts, everyone should be grounded in a shared glossary, decision trees, and examples. If you can’t explain a rule in a sentence, you probably need to rewrite it.

• Regular calibration sessions. Short, focused sessions where reviewers rate the same sample documents help uncover ambiguities before they show up in the workflow. It’s the listening tour of the team portrait.

• Quick dispute resolution. When disagreements pop up, there should be a simple, predictable path to resolve them. Document the rationale, update guidelines if needed, and move on.

• Ongoing audits. Periodically re-check a subset of decisions to ensure consistency hasn’t drifted. A few minutes of audit time can save hours of rework later.

• Role clarity. Make sure reviewers know what success looks like for each role. A shared sense of purpose reduces drift and keeps conversations productive.

A real-world lens: why this matters in a Relativity project

Relativity is a powerhouse for handling large volumes of documents, with features that support organization, tagging, and review workflows. In practice, coherence isn’t just a nice-to-have; it’s a practical necessity.

• When dealing with large datasets, even small variances multiply. A few editors disagreeing about relevance can lead to a mis-timeline or overlooked material.

• In sensitive matters, consistency protects stakeholders. A consistent approach to privilege or confidentiality decisions reduces the risk of leaks or misinterpretations.

• For teams that span locations or time zones, a clear coherence standard keeps the process stitched together. It’s the glue that prevents “great idea, bad execution” syndrome.

Cue the human side: why people care about this stuff

Here’s the thing: numbers matter, but people matter more. A coherence score is a signal—an indicator that the team is aligned enough to deliver trustworthy results. When you’ve got a diverse group of reviewers, you want to feel confident that the ground rules are understood, applied, and shared. That trust is what helps teams stay focused, even when the data sit on the heavy side.

Small tweaks that make a big difference

If you’re part of a team tackling a Relativity project, consider these approachable adjustments to boost coherence without turning the clock backward:

• Start with a concise, practical glossary. Keep it accessible—think quick reference rather than a doctoral dissertation.

• Run bite-sized calibration tasks. A 15-minute exercise with a handful of documents can reveal more than a long meeting ever would.

• Publish a decision log. A running record of why judgments were made keeps everyone honest and helps new reviewers come up to speed quickly.

• Normalize disputes. Treat disagreements as data points rather than obstacles. Each dispute is a chance to refine guidelines and sharpen collective judgment.

Common myths, debunked with a practical twist

• Myth: Higher coherence scores are always better. Reality: They’re best viewed as a tool for balance—reliability without sacrificing momentum.

• Myth: Calibration slows you down too much. Reality: When done in small, regular bursts, calibration saves time by reducing later rework.

• Myth: Coherence is someone else’s problem. Reality: It’s a team sport. Everyone benefits from clear rules and consistent practices.

A few quick reflections to close

As you navigate a Relativity‑driven project, coherence isn’t a dry metric tucked in a dashboard. It’s a living measure of how well your team can turn a big, careful review into credible, actionable outcomes. The threshold of 70 isn’t a magical ceiling; it’s a practical floor that helps teams stay trustworthy while preserving momentum.

If you’re curious to see the concept in action, you can think of it like a well-tuned orchestra. The conductor isn’t asking every musician to play the same note louder; they’re guiding them to align on tempo, dynamics, and phrasing so the music feels effortless, even though it’s complex. In the same spirit, coherence in assisted review weaves together disparate judgments into a cohesive, credible whole.

Bottom line: coherence is more than a number. It’s a compass that guides teams through the noise, helps protect the integrity of decisions, and keeps projects moving forward with confidence. When a Relativity project hits its stride, you’ll notice it in small, steady ways—fewer backtracks, clearer decisions, and a shared sense that everyone is rowing in the same direction.

If you’d like, I can tailor this discussion to reflect the particular kinds of documents you’re working with—case types, data sources, or stakeholder priorities. After all, the best navigation device is the one that fits your map.