Evidence that holds up when someone checks the method.
Survey design and programming, public opinion research, fieldwork, analytics and program evaluation — built for scrutiny.
The buyer's challenge
Activity is not evidence.
Programs are increasingly required to show evidence, not activity. In the United States the Foundations for Evidence-Based Policymaking Act obliges federal agencies to build evidence for their programs; funders and boards apply similar pressure everywhere else.
The failure mode is familiar: a survey written by the team that owns the program, fielded to a convenience sample, and reported as a result. It does not survive a methods review, and by then the reporting deadline has passed.
What we do
Capabilities in this practice.
Survey design
Instrument design, cognitive testing and question wording that does not lead the respondent. Sampling frames and weighting documented before fieldwork starts.
Survey programming
Complex logic, multi-language instruments, accessibility-conformant question rendering, and quality checks built into the instrument itself.
Public opinion research
Population studies with documented sampling, response-rate reporting and margin-of-error statements that reflect the design actually used.
Fieldwork & panel management
Fielding, quota management, data quality screening and removal of fraudulent or inattentive responses, with the exclusion rules stated in advance.
Analytics
Descriptive and inferential analysis, segmentation and dashboards — with the analysis code retained so a result can be reproduced.
Program evaluation
Logic models, outcome measurement and process evaluation designed so the finding can be attributed to the program rather than to the environment.
What you receive
Deliverables.
- Research design and methodology statement
- Tested survey instrument with logic map
- Sampling plan, weighting scheme and field report
- Cleaned dataset with a documented data dictionary
- Reproducible analysis code
- Findings report written for a non-technical decision-maker
- Technical appendix that survives a methods review
How we deliver.
Define the question
We write the decision the research is meant to inform, in one sentence, and get it agreed. Studies that cannot pass this step do not get fielded.
Design & pre-register
Method, sample, weighting and analysis plan documented before data collection — so the analysis cannot drift toward a preferred answer.
Field & quality-control
Live monitoring of quotas, completion and data quality, with the exclusion criteria fixed in advance.
Report both ways
A plain-language report for the decision-maker and a technical appendix for whoever checks the method. Both, always.
Methods & technologies
What we build with, and what we build to.
Standards listed here are frameworks we design and document against. They are a description of how we work, not a claim to be certified against them — our actual certification status is published in full on the certification roadmap.
Standards we work to
- AAPOR Transparency Initiative disclosure elements
- OMB statistical standards
- GDPR / PIPEDA data handling
Evaluation
- Logic models
- Quasi-experimental design
- Process evaluation
- Outcome measurement
Analysis
- R
- Python
- SQL
- Reproducible notebooks
Outcomes & sectors
What changes.
Stated as capability, not as a percentage. We do not publish performance figures we cannot attribute to a named engagement.
- A finding that states its own confidence and limitations
- A dataset another analyst can pick up and reproduce
- Disclosure that meets recognised transparency standards
- Evidence a funder or oversight body will accept
Sectors we deliver into.
- Government
- Nonprofit & philanthropy
- Health & human services
- Education
- Financial services
North American public-sector codes
- NAICS 541910
- PSC B505
- PSC R701
Questions
Why is a technology company doing survey research?
Because the hard part of modern research is instrument engineering, data pipelines and reproducibility — and because measuring whether a system worked is the last step of our delivery loop, not a separate business.
Can you field in multiple languages?
Yes, including instrument translation and culturally adapted question wording. Translation is treated as a design task subject to testing, not a text substitution.
Will you publish a finding we do not like?
We will report what the data supports, including null and unfavourable results. Agreeing that in advance is part of scoping, and it is the reason the finding is worth anything.
The rest of the loop
This pillar is one part of a delivery loop.
Have a requirement in Data, Research & Surveys?
Send the solicitation, statement of work or role description. You will get a direct answer on fit, including when the honest answer is that we are not the right supplier.