Findings your board can act on, and anyone can check.

Send us interviews or open-ended survey responses, at any scale. Our team returns answers to your questions and recommendations for what to act on, with every sentence linked to the responses behind it.

Led by Andrew Katz, PhD

Book a scoping callSee how we count

What you receive

  1. Answers to your questions, every sentence cited and counted.
  2. Recommendations, marked as our judgment and set on their evidence.
  3. The evidence itself: each finding linked to the responses it came from.

Answers where every sentence cites its responses.

Bring our team a question and you receive a memo. Each sentence carries a count of the responses that mention it and opens to their own words. A point the data doesn’t support is struck out, not written. Here, a question about returning to the office, answered from 1,806 responses.

Question

What’s driving the pushback on returning to the office?

Asked of all 1,806 responses · Return to the office (simulated)

1,028 of 1,806 responses mention this

The cumulative expense of daily fuel, transit fares, and mandatory parking fees effectively erodes the salary increase I anticipated from this role transition.marketing coordinator, under a year
1 of 4 shown · 1,028 in all

413 of 1,806 responses mention this

204 of 1,806 responses mention this

83 of 1,806 responses mention this

38 of 1,806 responses mention this

28 of 1,806 responses mention this

Staff are organizing a petition against the policy.

Not supported by your data. 0 of 1,806 responses mention a petition, a union, a protest or organizing together, so it stays out of the answer.

126 of 1,806 responses mention this

Tap any sentence to see who said it.

Illustration: Return to the office, 1,806 simulated responses, none written by a real person. Each count is the number of responses that mention the point, with every match read by a reviewer. Every quote shown was read and confirmed.

Recommendations a board can audit, down to a single response.

The memo ends with what to act on. Each recommendation is marked as our judgment and kept apart from the evidence under it: the finding, how it was counted, who mentions it, and their words. A reader can follow one down to the responses, or start from a single response and see every recommendation it supports.

This one response is evidence for 2 recommendations.

Our recommendation (judgment)

Strength: share that mention it. Strong 20%+, moderate 10–19%, limited under 10%.

Evidence: finding

Evidence: group test

Evidence: response

Select anything, in any column, to trace it both ways.

Re Return to the office: what to act on

Basis 1,806 responses, read and counted; every figure below opens to its evidence

  1. Our recommendation (judgment, not a finding)

    Address the burden of the commute, starting with the hours.

    More than half of all responses mention the commute, its hours most often, then the energy it takes and what it costs.

    Evidence: strong(57% mention it)Strength is the share of responses that mention it: strong 20% or more, moderate 10–19%, limited under 10%.

    Evidence: the finding

    1,028 of 1,806 responses (57%) mention the commute: the hours, the cost, the energy it takes.

    • Hours586
    • Energy383
    • Cost171

    What the mentions are about; one mention can name more than one.

    82 of these mentions are positive or neutral (they live close, or don't mind it): a mention is not always a complaint.

    Evidence: how it was counted

    Responses that mention any of: commut*, traffic, parking, gas money, fuel, transit. Every match was read. A mention counts what people talk about, not what matters most to them.

    Evidence: who mentions it

    • Live close to the officeMore: 69% mention it, vs 55% of everyone else (155 of 224 vs 873 of 1,582)
    • Sharing with roommatesMore: 65% mention it, vs 56% of everyone else (152 of 234 vs 876 of 1,572)
    • Parents of young childrenLess: 47% mention it, vs 58% of everyone else (111 of 234 vs 917 of 1,572)
    • Caring for an aging parentLess: 47% mention it, vs 58% of everyone else (103 of 221 vs 925 of 1,585)

    Only clear differences are shown (each group against everyone else, Fisher’s exact test, Holm-corrected).

    Evidence: in their words

    Hours
    Now, I will be spending an extra ninety minutes every day stuck in traffic or waiting for transit.
    operations supervisor, under a year
    Energy
    The daily trek through transit hubs and traffic is physically draining.
    software engineer, 10+ years
    Cost
    The cumulative expense of daily fuel, transit fares, and mandatory parking fees effectively erodes the salary increase I anticipated from this role transition.
    marketing coordinator, under a year

  2. Our recommendation (judgment, not a finding)

    Give caregivers more control over their hours.

    Family and caregiving come up in about a quarter of responses, and in most responses from parents of young children.

    Evidence: strong(23% mention it)Strength is the share of responses that mention it: strong 20% or more, moderate 10–19%, limited under 10%.

  3. Our recommendation (judgment, not a finding)

    Let people choose where they do focused work.

    Responses that mention noise and interruptions point both ways: some at the open office, some at home.

    Evidence: moderate(11% mention it)Strength is the share of responses that mention it: strong 20% or more, moderate 10–19%, limited under 10%.

Open any recommendation to follow it down to the responses.

Illustration: the same 1,806 simulated responses, none written by a real person. The findings are illustrative; the recommendations show the format, not advice, and predict nothing about what an action would achieve.

Strength is the share of responses that mention it: strong 20% or more, moderate 10–19%, limited under 10%. Group differences: each group against everyone else, Fisher’s exact test, Holm-corrected; only clear differences are shown.

Three kinds of analysis, each returned with its evidence.

Every engagement is carried out by our team. We do the analysis and hand you the results, with the responses behind each finding and a plain account of how it was made.

  • Inductive Coding

    Extract-based thematic analysis that maintains qualitative nuance while providing quantitative insight.

    Used for

    • Large-scale feedback analysis
    • Customer experience research
    • Program evaluation
    • Policy impact assessment

  • Decision Network Analysis

    Map and analyze complex decision-making patterns to understand how choices are made and influenced.

    Used for

    • Customer journey mapping
    • User experience optimization
    • Investment decision modeling
    • Product adoption analysis

  • Causal Belief Networks

    Understand how people think about cause and effect in complex systems.

    Used for

    • Educational outcome analysis
    • Healthcare intervention planning
    • Risk perception modeling
    • Behavioral change research

How an engagement works

  1. A scoping call

    You tell us the decision and the data you have. We say what an analysis can and can’t answer before any work starts.

  2. The analysis

    Our team reads and codes your responses. Every count we report is checked against the responses behind it.

  3. The delivery

    You receive the answers, the recommendations and the evidence, each finding linked to the responses it came from.

Start with the decision your data has to inform.

Tell us what you have and what it needs to answer. Book a scoping call, or send one quick question first.

Book a scoping call

Just have a quick question?

Or email [email protected]

We reply by email.