delveo’s practical guide to inclusive process design: qualitative benchmarks for 2025

Inclusive process design is a discipline that promises better outcomes for everyone, yet many teams struggle to know if they are actually making progress. Without clear qualitative benchmarks, efforts can feel aimless—or worse, performative. This guide from delveo offers a set of practical, observable indicators for 2025, grounded in real-world facilitation and organizational change work. We focus on what you can see, hear, and feel in a room (physical or virtual), not on numbers you cannot verify.

Who needs qualitative benchmarks and what goes wrong without them

Qualitative benchmarks are for anyone who designs or facilitates processes that involve multiple stakeholders—facilitators, product managers, HR leads, community organizers, and consultants. Without them, teams often fall into one of several traps. The first is metric fixation: they chase easy-to-count outputs (number of attendees, survey response rates) while ignoring whether participants felt heard or able to contribute. The second is performative inclusion: adding a feedback form or a diversity statement without changing how decisions are actually made. The third is burnout: well-meaning facilitators overcorrect by trying to accommodate every possible need in real time, leading to exhaustion and inconsistency.

Consider a composite scenario: a mid-sized tech company wants to redesign its quarterly planning process to be more inclusive. The team adds a pre-meeting survey, invites more junior staff, and appoints a facilitator. On paper, it looks inclusive. But in practice, the same senior voices dominate, the survey results are ignored, and junior participants leave feeling tokenized. Without qualitative benchmarks—like observing who speaks first, whose ideas get built upon, or how disagreements are handled—the team cannot see what is broken.

Another common failure is relying on a single data point. A team might run one retrospective where participants say they felt included, and declare success. But inclusion is contextual and dynamic; a process that works for a homogeneous group may fail when the group becomes more diverse. Qualitative benchmarks provide a continuous check, not a one-time checkbox.

Finally, without shared benchmarks, team members may have conflicting ideas of what inclusion looks like. One person thinks it means everyone speaks equally; another thinks it means decisions are made by consensus; a third thinks it means underrepresented groups get veto power. These unspoken differences create friction and undermine trust. Qualitative benchmarks offer a common language to discuss what inclusive process design actually requires.

Who this guide is for

This guide is written for practitioners who are already familiar with basic facilitation concepts but want to move beyond surface-level inclusion. It is not an introductory primer on diversity or equity; it assumes you have some experience and are looking for more nuanced ways to evaluate your work.

Prerequisites: what to settle before you start benchmarking

Before you can apply qualitative benchmarks, you need to establish a few foundational elements. First, define what inclusive means in your context. Inclusion is not a universal state; it depends on the power dynamics, cultural norms, and historical patterns in your organization. A benchmark that works for a startup with 20 people may not fit a multinational with 10,000. Spend time with stakeholders to articulate what inclusion looks like for this specific process—not in abstract terms, but in observable behaviors. For example, instead of “everyone feels valued,” you might define inclusion as “decisions are explained, and dissent is acknowledged without penalty.”

Second, secure commitment from decision-makers. Qualitative benchmarks require time and attention to observe and discuss. If leaders expect a simple numeric score, they may resist the ambiguity. Prepare a brief elevator pitch: explain that qualitative benchmarks reveal patterns that numbers miss, and that they can be aggregated into themes for reporting. Without this buy-in, your benchmarking efforts may be dismissed as “too soft.”

Third, train observers or facilitators to notice specific signals. Not everyone is naturally attuned to micro-dynamics like who interrupts whom, or how silence is interpreted. Create a simple observation guide with examples of what to look for. For instance, “notice whether questions are answered directly or deflected” or “track whether side conversations are invited into the main discussion.” This training does not need to be formal—a 30-minute walkthrough before a meeting can suffice.

Fourth, establish a feedback loop. Benchmarks are useless if they are not acted upon. Decide how observations will be shared—anonymously, in aggregate, or as part of a retrospective—and who is responsible for making changes. Without a feedback loop, the benchmarking becomes a surveillance exercise rather than a learning tool.

Finally, acknowledge your own biases. As observers, we all have blind spots. A facilitator might over-index on verbal participation and miss non-verbal cues, or might assume that silence means agreement. Consider having multiple observers from different backgrounds, or rotating observers across sessions. This reduces the risk of a single perspective dominating the interpretation of what is happening.

Common prerequisites checklist

• Context-specific definition of inclusion for the process
• Leadership buy-in for qualitative methods
• Observation guide with concrete examples
• Feedback mechanism (e.g., debrief after each session)
• Multiple observers or rotation to reduce bias

Core workflow: how to develop and use qualitative benchmarks

This workflow is designed to be iterative and adaptable. It consists of four phases: identify signals, collect observations, interpret patterns, and adjust the process. Each phase builds on the previous one, but you can loop back as needed.

Phase 1: Identify signals

Start by listing observable behaviors that indicate inclusive or exclusive dynamics. These signals should be concrete and context-specific. For a brainstorming session, signals might include: “ideas from junior members are built upon by senior members,” or “the facilitator explicitly invites contributions from those who haven’t spoken.” For a decision-making meeting, signals could be: “the rationale for a decision is shared before a vote,” or “dissenting opinions are summarized and addressed.” Aim for 5–10 signals per process. Avoid vague terms like “respect” or “safety”; instead, describe what respect looks like in that setting.

Phase 2: Collect observations

During the process, have one or more observers note occurrences of each signal. This can be done with a simple tally sheet or a digital tool like a shared document. Observers should also capture critical incidents—moments that felt particularly inclusive or exclusive—with enough context to understand what happened. For example: “During the budget discussion, when Ana proposed a new approach, three people interrupted her within 30 seconds. She did not finish her point.” This level of detail helps in the interpretation phase.

It is important to collect observations in real time, not from memory. Memory is unreliable, especially for subtle interactions. If real-time observation is not possible, record the session (with consent) and review later. But real-time observation allows for immediate intervention if something harmful occurs, which is a separate benefit.

Phase 3: Interpret patterns

After the session, the observers and facilitator (and optionally a few participants) review the observations. Look for patterns across signals. For instance, you might notice that interruptions are concentrated on certain speakers, or that ideas from one department are consistently ignored. Discuss possible explanations: is it a power dynamic, a communication style difference, or a structural issue like time pressure? Avoid jumping to conclusions; treat patterns as hypotheses to test in future sessions.

This phase is also where you compare observations against your definition of inclusion. If your definition includes “decisions are explained,” but you observe that most decisions are announced without rationale, that is a clear gap. Document these gaps and prioritize which to address first based on impact and feasibility.

Phase 4: Adjust the process

Based on the patterns, make concrete changes to the process. This might be as simple as adding a round-robin check-in to ensure everyone speaks, or as complex as redesigning the agenda to allocate more time for divergent thinking. The key is to make one or two changes at a time, then observe again in the next session. This creates a continuous improvement loop. Over time, you will build a set of benchmarks that are specific to your team and context.

Tools, setup, and environment realities

Qualitative benchmarking does not require expensive software. A simple spreadsheet or a shared document can serve as your observation log. However, there are a few tools and setup considerations that can make the process more effective.

Observation guides and templates

Create a template that lists your signals with columns for tally marks, notes, and a timestamp. This keeps observations consistent across sessions. You can also include a section for critical incidents with prompts like “What happened? Who was involved? What was the outcome?” Free tools like Google Docs or Notion work well. For remote sessions, consider using a second screen or a chat-based bot to capture observations without disrupting the flow.

Environmental factors

The physical or virtual environment affects inclusion. In a physical room, seating arrangements matter: who sits at the head of the table, who is in the back row. In a virtual meeting, the chat function can be a great equalizer, but only if it is moderated and acknowledged. Note environmental factors as part of your observations. For example, “the only woman in the room was seated farthest from the projector, making it hard for her to see the slides.” These details often reveal hidden power structures.

Time and attention constraints

Observation is demanding. If you are also facilitating, it is nearly impossible to observe effectively. Ideally, have a separate observer. If that is not possible, record the session and review later, or ask a participant to take on a dual role (with clear instructions). Another option is to focus on just one or two signals per session rather than trying to capture everything. Over time, you can rotate which signals you track.

Technology for remote sessions

For remote meetings, tools like Zoom or Teams offer features like breakout rooms, polls, and non-verbal reactions. These can be used as built-in benchmarks: are participants using the raise-hand feature? Are poll results evenly distributed? But be cautious—technology can also introduce bias. For instance, participants with slower internet may be left behind. Include connectivity and platform familiarity as part of your observations.

Variations for different constraints

Not every team has the resources for a dedicated observer or the time for a lengthy debrief. Here are variations for common constraints:

Small teams or limited budget

If you cannot spare an observer, use a rotating participant-observer model. In each session, one participant volunteers to observe and takes notes on a specific signal. They share their observations at the end of the meeting. This spreads the load and builds observation skills across the team. Another low-cost approach is to use a check-in and check-out ritual: at the start, ask each person to state one thing they need to feel included; at the end, ask how well that need was met. These qualitative data points can be tracked over time.

Large groups or multi-session processes

For large groups, it is impractical to observe every interaction. Instead, sample sessions or use anonymous feedback forms with open-ended questions like “Describe a moment when you felt heard (or not heard) in today’s session.” Aggregate responses and look for themes. You can also train a small team of observers to cover different breakout rooms or tracks. For multi-session processes, track benchmarks over time to see trends—for example, does the frequency of interruptions decrease after a new facilitation technique is introduced?

High-stakes or sensitive contexts

In contexts where power imbalances are extreme (e.g., a meeting between executives and entry-level staff), direct observation may feel intrusive. In such cases, consider confidential third-party observation or post-session interviews with a cross-section of participants. Another variation is to use decision audits: after a decision is made, trace how it was reached and whose input was considered. This can be done without naming individuals, focusing on process steps.

Cross-cultural or multilingual settings

Inclusive process design must account for cultural differences in communication. For example, in some cultures, direct disagreement is seen as disrespectful, so silence may indicate dissent rather than agreement. In multilingual settings, language fluency affects participation. Adjust your benchmarks accordingly: instead of “speaks up,” use “communicates agreement or disagreement through culturally appropriate channels.” This requires understanding the cultural norms of participants, which itself is a benchmarking exercise. Consider having a cultural liaison who can help interpret signals.

Pitfalls, debugging, and what to check when it fails

Even with the best intentions, qualitative benchmarking can go wrong. Here are common pitfalls and how to address them.

Pitfall 1: Over-relying on a single observer's perspective

One person's observations are inevitably filtered through their own biases. If you notice that your benchmarks consistently point to the same few issues, consider whether the observer is focusing on certain participants or behaviors. Solution: rotate observers, or have two observers compare notes after a session. If they disagree, that is a useful data point—it reveals that the signals are ambiguous and need clarification.

Pitfall 2: Confusing observation with judgment

Observers sometimes slip into evaluating whether a behavior is “good” or “bad” rather than neutrally noting it. For example, noting “the facilitator interrupted a participant” is different from “the facilitator rudely interrupted a participant.” The latter introduces bias. Solution: train observers to use descriptive language and separate observation from interpretation. In the debrief, you can interpret together, but the raw data should be as objective as possible.

Pitfall 3: Ignoring structural factors

Sometimes the process itself is the problem, not the facilitation. If you observe that certain voices are consistently missing, it may be because the meeting time excludes people in different time zones, or because the agenda assumes a level of prior knowledge that not everyone has. Before blaming the facilitator or participants, check the process structure: Is the agenda realistic? Are materials accessible? Is there a clear decision-making framework? Qualitative benchmarks should include structural signals, not just interpersonal ones.

Pitfall 4: Benchmark fatigue

If you track too many signals, observers burn out and participants feel surveilled. Keep the list to 5–7 signals per session. Rotate signals across sessions to cover different dimensions over time. Also, communicate to participants that the purpose is learning, not evaluation—this reduces anxiety.

Pitfall 5: Not acting on observations

If you collect observations but never change anything, participants will lose trust. Make sure there is a clear feedback loop: after each session, share a summary of patterns (anonymized) and announce one or two changes for the next session. Even small adjustments—like starting with a round-robin check-in—signal that you are listening.

Debugging checklist

• Are observers trained and rotated?
• Are observations descriptive, not evaluative?
• Have we examined structural factors?
• Are we tracking too many signals?
• Is there a feedback loop with visible changes?

Frequently asked questions about qualitative benchmarks for inclusive process design

This section addresses common questions that arise when teams start using qualitative benchmarks.

How do we know if our benchmarks are good?

A good benchmark is specific, observable, and relevant to your context. Test it by asking: “If I saw this happen, would I be confident that it indicates inclusion or exclusion?” If the answer is ambiguous, refine the signal. Also, check that your benchmarks are not biased toward a particular communication style. For example, valuing only verbal participation may disadvantage introverts or non-native speakers. Include signals that capture different forms of contribution, such as written input or non-verbal agreement.

Can we combine qualitative benchmarks with quantitative data?

Yes, but be careful not to let the quantitative data override the qualitative. Use numbers to complement, not replace, observations. For instance, you might track the number of times each person speaks (quantitative) alongside observations about whether their ideas were engaged with (qualitative). The quantitative data can highlight patterns, but the qualitative data explains why those patterns matter.

How often should we review benchmarks?

Review after each session if possible, or at least after every few sessions. The goal is to catch patterns early and adjust quickly. If you review too infrequently, you may miss the chance to intervene before exclusion becomes normalized. For long-term processes, do a deeper review monthly or quarterly to assess overall trends.

What if participants resist being observed?

Transparency is key. Explain that the observations are about the process, not about individuals, and that all data will be anonymized. Offer participants the option to opt out of being observed, though this can complicate the data. In practice, most people are fine with observation if they understand the purpose and see that it leads to improvements.

How do we handle disagreements about what the observations mean?

Disagreement is a sign that your benchmarks need clarification. When observers or participants interpret the same event differently, use that as an opportunity to refine your signals. For example, if one person thinks a long pause indicates reflection and another thinks it indicates confusion, you may need a signal that distinguishes between the two (e.g., “participant asks for clarification after a pause” vs. “participant builds on the pause to formulate a response”).

What are next steps after we have a set of benchmarks?

Once you have a stable set of benchmarks, embed them into your regular process documentation. Share them with new facilitators or team members as part of onboarding. Also, consider sharing your benchmarks with other teams or communities to contribute to a broader understanding of inclusive process design. Finally, revisit your benchmarks periodically—at least once a year—to ensure they remain relevant as your context evolves.