Services / Survey analysis
Survey analysis that tells you what your respondents actually said
Staff surveys, customer research, consultations and needs assessments come to us when the raw counts are in and someone has to say what they mean. A named statistician weights the sample, tests what is real and what is noise, and writes it up so a board or funder can act on it.
- Named statistician assigned before you commit
- Weighted and unweighted results, side by side
- Every published comparison significance-tested
- Follow-up questions included
Survey analysis is turning questionnaire responses into findings an organisation can act on: weighting the sample to the population it represents, testing which differences between groups are real rather than noise, coding open-text answers, and reporting all of it so a board, funder or executive can read it without a statistics background.
What is included
What you get with survey analysis
Weighting to your population
If women, night-shift staff or one region answered less often than they exist in the population, the raw figures are wrong. We weight to what you know about the whole group and show you the difference it made.
Significance, not just percentages
A 6-point gap between two teams is a finding if each team has 300 people and noise if each has 9. Every comparison you want to publish is tested and the base sizes shown.
Scales and constructs
Likert items combined into reliable scales — engagement, satisfaction, trust — with reliability reported, so you are measuring a thing rather than averaging questions.
Open-text coding
Free-text comments coded into themes, counted, and reported alongside the numbers. The why behind the what.
Year-on-year and benchmark comparison
Whether this year moved from last year, and whether your figure differs from a sector benchmark — with the uncertainty stated instead of hidden.
A report people will read
Tables and charts ready for a board paper or funder return, a plain-English summary of what matters, and a technical appendix for anyone who wants to check.
Who it is for
Two kinds of client, one standard of work
- Service-user and beneficiary surveys for funder reporting
- Public consultations with a known population to weight to
- Needs assessments and community surveys
- Staff and volunteer surveys where the response rate was uneven
- Customer satisfaction, NPS and market research surveys
- Employee engagement surveys with subgroup reporting
- Agency fieldwork that needs an independent statistical write-up
- Tracking studies — is this wave different from the last one?
The work itself
What survey analysis actually involves
Most surveys arrive as a spreadsheet export and a deadline. The percentages are easy — any survey platform produces them. What the platform cannot tell you is whether the people who answered look like the people you meant to ask, whether the differences you can see would survive a second survey, or what to do about the 40% of respondents who left a comment.
So the work starts with the sample. We compare who responded against what is known about the whole population — headcount by department, customers by region, residents by ward — and weight where the mismatch is large enough to matter. Then every comparison you intend to publish is tested. Then the open text is read and coded. Then it is written up in the order a reader needs it, not the order the questionnaire asked it.
The result is a report that says what the survey found, how sure you can be, and where the evidence is too thin to lean on — which is the part most survey reports leave out, and the part a board or funder will ask about.
Typical work
Surveys organisations bring us most often
If your survey resembles one of these, we have done it before and can quote quickly.
Staff and engagement surveys
Weighted to headcount, scales built and checked, and subgroup gaps tested before they go in front of the executive team.
Customer and market research
Satisfaction, NPS and tracking studies — including whether a movement between waves is real or sampling noise.
Public consultations
Responses weighted to the resident population, with the self-selection made explicit rather than ignored.
Beneficiary and outcome surveys
Before-and-after and matched designs for funder returns, with effect sizes and confidence intervals reported.
Needs assessments
Estimating prevalence and demand for a service from a sample, with the margin of error stated.
Second-opinion review
An independent check of survey analysis done in-house or by an agency, before it is published.
How to compare us
Questions to ask of any survey analysis
These are the questions we would put to any provider, including us. If the answers are vague, the report will be too.
| What to ask | Why it matters | Our answer |
|---|---|---|
| Was the sample weighted, and to what? | Unweighted survey figures assume responders match the population. They rarely do. | Weighted to what you know about the whole group; both versions reported. |
| Are the subgroup differences tested? | A percentage-point gap means nothing without the base sizes behind it. | Every published comparison tested; base sizes shown in every table. |
| What was done with the free text? | Comments are usually skimmed for quotes. That is not analysis. | Coded thematically and counted, reported alongside the numbers. |
| Can the analysis be reproduced? | If it cannot be re-run, it cannot be checked next year. | Syntax or code delivered with the report. |
| Who did the work? | Survey agencies often subcontract the statistics. | A named statistician; you correspond with them directly. |
Security and governance
How your data is handled
Usually the first question from NHS, university and local authority buyers — reasonably so.
If your organisation has a standard supplier assurance process, send it with the enquiry and we will complete it during scoping rather than after you have committed.
Pricing
What it costs, and why it is quoted rather than listed
The honest answer is that price depends on the state of the data and the scope of the question. A clean dataset with a well-specified question costs a fraction of a project that begins by reconciling three incompatible systems. So we look first and quote after.
Typical agency model
- Hourly billing against a scope that expands
- A senior name sells it, a junior delivers it
- Output without the code behind it
- Revisions billed as new work
How we work
- Fixed price agreed before anything starts
- The statistician named on the quote does the work
- Syntax and method note delivered as standard
- Follow-up questions included in the price
For most organisations this is materially cheaper than the alternatives — a permanent analyst hire for intermittent work, or a large consultancy billing hourly.
How it works
Three steps, entirely in writing
Send what you have
Attach the data, or just describe the project. No account, no mandatory call, and a five-field form rather than a fifteen-field one.
Approve a fixed quote
Usually within one working day, naming the statistician assigned, the deliverables and the deadline.
Receive the work
Output, syntax and a plain-English explanation — with follow-up questions answered at no extra cost.
Client feedback
What clients say about this work
“The survey findings were presented with useful charts and meaningful insights. Exactly what I needed for my report.”
“Very detailed analysis of my questionnaire responses. They identified trends that I had completely overlooked.”
Proof
This work, on real projects
Clients are not named and no identifying detail is published, so what is described is the statistical problem — which is the part that shows whether a consultancy knows what it is doing.
Factor Analysis and Ordinal Regression on a National Survey
Reducing dozens of Likert items to reliable constructs, then modelling what predicted them across a population-scale sample.
Consultation · Framework analysisFramework Analysis to Compare Stakeholder Groups
Open-text consultation responses coded and compared across stakeholder groups, so the differences in view were evidenced rather than asserted.
Workforce · Thematic analysisReflexive Thematic Analysis of Staff Interviews
What staff actually said about a change programme, analysed to a published standard and reported to the executive.
Your statistician
Who does this work
Member of the Royal Statistical Society
Econometrics, forecasting and causal inference for policy evaluation, business analytics and operations.
This page was reviewed for statistical accuracy by Saad Tariq on 14 August 2026.
Common questions
Answers before you ask
Can you weight our results to the population we surveyed?
Yes. If you can tell us the known make-up of the population — from HR records, a census, a customer database — we weight the responses to match it and report both weighted and unweighted figures so the effect of weighting is visible.
Our response rate was low. Is the survey still usable?
Often, yes, but the answer depends on who did not respond rather than how many. We check whether responders differ from the population on what is known about both, weight where that is defensible, and tell you plainly where a finding should be reported as indicative rather than firm.
Will you tell us which differences are real?
That is the core of the work. Every subgroup comparison you want to publish is tested, with the sample sizes behind it shown, so a five-point gap between two departments of twelve people is not reported as if it were a finding.
Can you analyse the free-text comments as well?
Yes. Open responses are coded thematically and reported alongside the numbers, so the report explains not just what changed but what respondents said about why.
What do we receive?
A written report with the tables and charts ready to drop into a board paper or funder return, a technical appendix with the method and the weighting scheme, and the syntax or code so the analysis can be reproduced.
How is the price set?
Quoted against a written scope once we have seen the data — number of respondents, number of questions, how many subgroups you need compared, and whether free text is included. Fixed once agreed.
Related
Often needed alongside this
Statistical consultancy
When the survey is one input to a bigger question — evaluation, forecasting, a business case.
From £950Qualitative coding
Interviews and focus groups analysed to a published standard, alongside the survey.
Fixed quoteSample size & power
How many responses you need before you send the survey, not after.
From £180Send the data. Get a fixed quote.
Attach the survey export or just describe the study. A named statistician replies with a price and a deadline, usually within one working day.