Sample Deliverables
What an Engagement Actually Produces
Not a word cloud and a handshake — written findings you can act on, and a methods appendix that lets you defend them. Here is the format and depth, shown honestly.
Read this first. The samples below are illustrative — drawn from a fully worked engagement with a fictional organization, not from a past client. Client work is confidential, and we would rather show you exactly what you’ll receive than dress the site up with borrowed logos. The format, structure, and rigor shown here are what every real engagement delivers. Evidence over hype applies to us, too.
Sample 1 — The Findings Report
The centerpiece of every engagement: a written report to leadership that answers the question you actually asked. Excerpted below from an illustrative engagement — a community foundation that collected 2,400 open-ended responses across three years of grantee and community surveys, and had read almost none of them.
Excerpt — Executive Summary
“Across 2,400 responses and three survey waves, one theme has more than doubled while overall satisfaction held steady: respondents increasingly describe the application process — not funding amounts — as the barrier… The signal is concentrated among first-time applicants and is strongest in the most recent wave, which is why it has been easy to miss in year-by-year reads… Fifteen direct quotes, each verified in context, illustrate the pattern; the full coded dataset accompanies this report so your team can check any claim against the original text.”
Excerpt — Findings at a Glance
| Theme | Share of responses | Change since Wave 1 |
|---|---|---|
| Application & reporting burden | 31% | ▲ from 14% |
| Staff responsiveness (positive) | 27% | steady |
| Funding flexibility | 22% | ▲ from 17% |
| Capacity-building requests | 18% | new in Wave 3 |
| All other themes | <10% each | — |
Figures are illustrative. In a real report, every theme links to its definition in the codebook, its validation results, and example responses.
What the full report contains
The question as agreed, in one sentence · findings ordered by decision relevance, each with prevalence, trend, and verified example quotes · breakdowns by group, time period, or source where the data supports them · what we looked for and did not find · recommendations only where the evidence supports one · the full coded dataset, so nothing is take-our-word-for-it.
Sample 2 — The Methods Appendix
Every findings report ships with a methods appendix — the document that makes the analysis defensible to a board, a funder, a reviewer, or a court. Most analytics vendors won’t show you one, because most can’t write one. The structure:
Table of contents — Methods Appendix
- 1.The corpus — what text was analyzed, where it came from, what was excluded and why
- 2.The measurement — each signal defined precisely: what counts, what doesn’t, with the full codebook
- 3.The models — which methods and models were used, with versions and settings recorded for every run
- 4.Validation — how machine output was checked against human judgment, on what sample, with agreement statistics
- 5.Known failure modes — where the instrument errs, in which direction, and how that bounds the findings
- 6.Limitations — what this analysis cannot tell you, stated plainly
- 7.Reproducibility — what would be needed to re-run the analysis and get the same answer
For longitudinal contracts, the appendix also documents series continuity: when an instrument changes, how the change was tested, and how comparability with earlier periods is preserved.
Also part of every engagement
The coded dataset
Every document with its assigned codes and scores, in a format your team can open, filter, and re-check — the findings are auditable down to the row.
Live readout
A working session with leadership — and with your board or funder, if useful — walking through findings, methods, and what to do next.
A reusable instrument
The codebook and pipeline are yours to keep. Run the same measurement on next year’s data — with us, or without us.
Want this for your organization?
Tell us the question you’re trying to answer. If text analytics isn’t the right tool for it, we’ll say so.
Get Started