Content analysis checklist: before, execution, after

This content analysis checklist covers the full project from setup through the final brief. Use it as a working document — copy it into your project notes, tick off each item as you go, and add comments where the project deviates from the standard flow. The checklist assumes a typical product research dataset of a few hundred to a few thousand units; for larger datasets the same items still apply, but plan more time for piloting and reliability checks.

Before

• Write down the specific research question content analysis should answer
• Choose the analytic logic (inductive, deductive, or summative) and document the choice
• Pull the dataset into one place and clean it (deduplicate, anonymize, fix encoding)
• Decide explicitly what is in scope (date range, channel, language, product area) and what is out
• Read 20–30 random units end to end without coding to absorb the tone and vocabulary
• Define the unit of analysis (word, sentence, paragraph, post, transcript) and write the rule down
• Draft the first version of the codebook with definitions, inclusion rules, exclusion rules, and examples
• Pilot the codebook on 5–10% of the data and refine until disagreements are edge cases

Execution

• Code the full dataset in a consistent order using the locked codebook
• Keep a memo file open for unfit units, new patterns, and notable quotes
• Stop and re-pilot if the codebook needs to change mid-pass; do not silently drift
• Tag out-of-scope units explicitly rather than leaving them blank
• For multi-coder projects, double-code 10–20% of units and calculate intercoder reliability
• For solo projects, re-code a random 5% sample after a one-week gap as a self-audit
• If using AI auto-coding, spot-check 5–10% of the output and document systematic errors
• Lock the coded dataset and version-control any post-coding edits

After

• Build the frequency table per category, with counts and percentages
• Build cross-tabulations across the dimensions that matter (segment, channel, time, version)
• Write a one-paragraph summary per category with common subpatterns
• Pull 3–5 illustrative quotes per major category covering the range
• Note the limitations of the dataset (sample bias, language coverage, time window)
• Write the insight brief (3–8 pages) with research question, top findings, category summaries, and recommendations
• Schedule a stakeholder readout and walk through the brief, not a deck of bullets
• Archive the codebook, the coded dataset, and the brief so the next analysis can build on this one