← Applied Research
Productized service · NDA-ready · 5–10 business days

Qualitative coding for evaluations and research.

Sitting on 20+ interview transcripts and a codebook, with no time to code them? We turn your transcripts into structured coding tables — with verbatim quotes, analytical memos, and inter-coder reliability checks — in 5 to 10 business days. AI-assisted first pass, expert review of every code.

1,000+
Projects evaluated
€50M+
Combined client budgets
17
Countries of fieldwork
4
Languages of coding
01

What you get

A clean, structured deliverable ready for your analysis or donor report — not raw outputs you have to re-process.

Structured coding tables

Excel or Word tables with codes, sub-codes, verbatim quotes, source IDs, and analytical memos. Format agreed at kickoff — NVivo, MAXQDA, ATLAS.ti, Dedoose, or plain Excel.

Inter-coder reliability scores

Cohen's Kappa calculated on a 20% sample with a second human coder. We publish the score in the deliverable rather than hiding it.

Source-traceable evidence grid

Every code links back to a transcript line and respondent ID. Auditors, donors, and triangulation analysis all benefit from full traceability.

Methodology appendix

Documents the codebook used, AI model and version, prompts, expert-review protocol, and reliability sample. Donor-audit ready.

Optional thematic summary

For Coding+Analysis and Full packages — a 5–15 page narrative report with key themes, frequency tables, and visual frameworks.

Optional codebook draft

If you don't have a codebook yet, we can draft one from your research questions, your evaluation matrix, or 5–10 sample transcripts.

02

How we work

Five steps from kickoff to delivery. Most engagements complete in 5–10 business days.

  1. 01

    Kickoff and calibration

    You send transcripts and your codebook (or your research questions if no codebook). Our senior evaluator reviews 1–2 transcripts to calibrate the coding rules and clarify edge cases with you.

  2. 02

    AI-assisted first-pass coding

    Locally-run language models apply your codebook to all transcripts. The AI runs on our infrastructure (no client data sent to commercial APIs, GDPR-compliant). This pass cuts total coding time by ~40%.

  3. 03

    Expert review of every code

    A senior evaluator reviews every machine-applied code, corrects misses and false positives, adds analytical memos for ambiguous cases, and resolves any cultural-context flags. AI accelerates — humans interpret.

  4. 04

    Inter-coder reliability check

    A second human coder independently codes a 20% sample. We calculate Cohen's Kappa, document any disagreements, and refine the codebook if needed.

  5. 05

    Delivery in your format

    Final coded files in your preferred CAQDAS format, plus the Excel evidence grid, ICR report, and methodology appendix. Optional debrief call to walk through findings.

03

Why us

Three positions in this market: cheap solo freelancers (uneven quality), AI-only tools (cultural blindness, hallucinations), and academic services (not built for evaluation contexts). We sit in the strategic middle.

Typical competitor
IMACON
Format
Solo freelancer
Specialised evaluation team
Specialisation
Generic qualitative analysis
M&E + evaluation + applied research
AI use
Hidden, or none, or AI-only
Open hybrid model with published ICR scores
Languages
English only
EN, RU, PL, BY
Methodology
Implicit
Published methodology — six books with ISBN
Trust signals
PhD title or platform rating
1,000+ projects, €50M+ budgets, UN/EU/USAID/EBRD clients
Confidentiality
Cloud AI uploads (privacy risk)
NDA per project · GDPR-safe · no sensitive data to commercial AI
Try-before-buy
Pay first
Free 2-page trial of your data
04

Pricing

Transparent per-transcript packages. Volume discounts for 15+ transcripts. Custom-scoped engagements priced separately.

Coding only
$150–300 / transcript

You provide transcripts and codebook. We deliver coded tables.

  • Codebook applied across all transcripts
  • Verbatim quotes with source IDs
  • Excel / NVivo / MAXQDA / ATLAS.ti format
  • Inter-coder reliability on 20% sample
  • Methodology appendix
  • Turnaround 5–10 business days
Full analysis
$500–700 / transcript

Full pipeline from codebook design to evaluation-grade report.

  • Codebook drafted from research questions or eval matrix
  • Everything in Coding + analysis
  • Triangulation across qualitative + secondary sources
  • OECD-DAC criteria mapping (if evaluation use case)
  • Recommendations section with senior evaluator review
  • Optional in-person debrief workshop
Hourly consulting
$50–100 / hour

For codebook design, ICR audits, methodology reviews, or training.

  • Codebook design and refinement
  • Inter-coder reliability audit on existing coded data
  • Methodology review for grant submissions
  • Coding-team training (in-house or remote)
  • Senior evaluator availability for ad-hoc consultations

Volume tiers reduce per-transcript cost: 1–5 transcripts $150–250 · 6–15 transcripts $120–200 · 16–30 transcripts $100–170 · 30+ transcripts $80–150. Transcription (audio→text) priced separately if needed.

05

Tools and methodologies

Software
  • NVivo
  • MAXQDA
  • ATLAS.ti
  • Dedoose
  • Quirkos
  • Excel
  • MS Access
Methodologies
  • Thematic analysis
  • Framework analysis (Gale et al.)
  • Content analysis
  • Deductive coding (a-priori codebook)
  • Inductive coding (grounded)
  • IPA (interpretative phenomenological)
  • OECD-DAC criteria mapping
  • Theory-of-change-aligned coding
06

Use cases

Programme evaluation (OECD-DAC)

Mid-term, final, ex-post evaluations of EU, UN, USAID, EBRD, bilateral programmes. Coding mapped to relevance, coherence, effectiveness, efficiency, impact, sustainability. Theory-of-change-aligned thematic synthesis.

Academic research

Thematic analysis, IPA, grounded theory, content analysis. NVivo / MAXQDA / ATLAS.ti deliverables ready for thesis appendix or journal submission. Multilingual transcripts welcome.

Market and UX research

Verbatim coding, sentiment tagging, codeframe development for survey open-ends. Survey-end analysis with cross-tabulation by respondent segment.

Mixed-methods studies

Qualitative component of mixed-methods designs. Codes mapped to quantitative survey variables for joint display tables and convergent-divergent integration.

NGO and CSO research

Stakeholder interviews, beneficiary feedback rounds, KAP-survey open-ends. Direct integration into donor reporting templates.

Internal organisational research

Employee interviews, customer satisfaction studies, change-management assessments. NDA-protected handling for internal-only use.

07

Trust signals

A specialised team — not a solo freelancer — with verifiable institutional track record.

Track record

1,000+ projects evaluated. Combined client budgets exceeding €50 million. 25+ years of cumulative qualitative-research practice across the founding team.

Institutional clients

UNDP (incl. Independent Evaluation Office), UNICEF, UNFPA, UNIDO, USAID, EuropeAid, EBRD, IFC, IFRC, DVV International, Renovabis, SIDA-funded programmes, BMZ, Polska Pomoc.

EBRD-prequalified consultancy

Prequalified for the EBRD Business Advisory Services programme since 2018, requalified Vilnius 2023. EBRD compliance procedures embedded in our project workflow.

Published methodology

Six methodological books with ISBN: IMACON Model, Organisational Viability Toolkit, LFA Guide, NORAD Planning Guide, and others. Free PDFs in our Publications section.

Confidentiality and GDPR

NDA signed for every project. GDPR-compliant data handling. We do NOT upload sensitive beneficiary data to cloud AI tools — all AI processing happens locally on our controlled infrastructure.

Professional memberships

Polish Evaluation Society. Minnesota Evaluation Association. European Evaluation Society. Transformation Thinkers Network (GIZ / Bertelsmann Foundation).

08

Frequently asked questions

Can you work with my CAQDAS format (NVivo, MAXQDA, ATLAS.ti, Dedoose)?
Yes. Tell us your tool at kickoff. We deliver native files (.nvp, .mx20, .atlproj, .dedoose) plus an Excel evidence grid as a backup. We can also work directly inside your project if you grant us access.
What languages do you code in?
English (native), Russian (native), Polish (professional), Belarusian (native). For other languages we work via verified transcript translations or partner with native-speaker coders. Multilingual codebooks are common — we maintain a master codebook in English with code labels in each transcript language.
How do you handle confidential or sensitive data?
NDA signed before any data exchange. GDPR-compliant data handling. AI processing happens on our local infrastructure — your transcripts are never sent to OpenAI, Anthropic, Google, or any third-party API. For sensitive evaluation work (vulnerable populations, prisoners, GBV survivors), we apply additional anonymisation and "do no harm" protocols.
What if we don't have a codebook yet?
We can draft one. Send us your research questions or evaluation matrix and 5–10 sample transcripts. We deliver a structured codebook in 3–5 business days, with examples and inclusion / exclusion rules. Once you approve, coding proceeds.
How does your AI assistance work? Can I see the human review?
AI runs locally on our infrastructure to apply your codebook in a first pass — typically achieving 70–85% accuracy on inter-coder reliability vs. expert-only baseline. Every machine-applied code is then reviewed and corrected by a senior evaluator. The methodology appendix documents which codes were AI-applied vs. expert-applied so you can audit. We publish the final Cohen's Kappa score with a second human coder.
How long will it take for X transcripts?
5 business days for up to 15 transcripts. 10 days for 16–40 transcripts. 15+ days for larger volumes or where we also draft the codebook. Rush projects (50% premium) can compress these timelines.
Do you offer a free sample of your work?
Yes. Send your codebook and a sample transcript (any length). We will code 2 pages free so you can assess quality, formatting, and the level of memo detail before committing. Most projects start this way.
How do you price multilingual or mixed-language transcripts?
Same per-transcript rate. Mixed-language transcripts (typical in evaluation contexts where interviews shift between English, the local language, and a regional lingua franca) are coded by a coder fluent in all relevant languages. No surcharge for multilingual capability — it's our default.
Can you integrate with our existing M&E plan?
Yes — this is our default for evaluation clients. Send your evaluation matrix, theory of change, and Logframe indicators. We map every code to evaluation questions and OECD-DAC criteria, so the deliverable plugs directly into your evaluation report structure.
Do you do transcription too?
Yes, as an add-on. $1.50–$2.50 per audio minute depending on language and audio quality. Transcripts delivered in your preferred format (.docx, .vtt, .srt). We can also work with audio directly using local Whisper-based transcription if you prefer.
Who actually does the coding?
A senior evaluator with at least an MA-level qualification in social research and 5+ years of qualitative analysis experience leads every project. The IMACON team includes 5+ qualified coders who collaborate on larger volumes. The lead evaluator signs the methodology appendix.
Do you sign DPAs (data-processing agreements)?
Yes. We have template DPAs aligned with EU GDPR Article 28 and can review your standard DPA. For UN agency clients we comply with UN data-protection principles. EBRD compliance procedures are standard.

Start with a free 2-page trial.

Send your codebook and a sample transcript. We code 2 pages free so you can assess quality, formatting and memo detail before committing. Most engagements start this way and convert within a week.