agentleFS
styfinitystyfinity/claude-marketing-department/skills/customer-deep-dive/SKILL.md

customer-deep-dive

Skill6 starsChanged 4 months agoMITView on GitHub

Builds the full customer picture in one pass — avatar, verbatim language bank, pains ranked by what they cost, desired outcomes, and objections sorted stated versus real — then writes it into the customer section of marketing-brief.md. Use before writing any copy or building any offer, or whenever the team describes the customer in company words instead of the customer's own.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Write the results into The customer section of marketing-brief.md — ICP notes, avatar, top 5 pains, desired outcomes, objections, language …

    line 24

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Customer Deep-Dive — the buyer, in their own words

Every weak headline, dead ad, and ignored email traces back to the same root: writing about the customer instead of from them. This skill mines real customer material into the assets copy actually runs on — a verbatim language bank, pains ranked by what they cost, and the objections that live behind the polite ones. One pass, written straight into the brief.

Inputs

  • Raw customer material — call notes, reviews (yours and competitors'), support tickets, community threads, win/loss notes — pasted or pointed to: $ARGUMENTS
  • From marketing-brief.md if present: the ICP and any existing customer section to sharpen rather than restart.

Do this

Inside Claude Code, steps 2–6 map to the customer-research agents by name — voc-miner, avatar-builder, pain-mapper, desire-mapper, objection-collector — and can run as delegated passes. Inline, work through them in order.

  1. Audit the raw material first. If there is little or none, do not invent a customer — name the fastest real sources (your reviews, competitor reviews, the communities where buyers complain) and build a provisional version clearly marked as such.
  2. Mine the language. Harvest verbatim phrases into a language bank grouped by pain, desire, and objection. Verbatim means verbatim — keep the typos, the profanity, the odd metaphors. That texture is what makes copy land.
  3. Build the avatar. Day in the life, pressures from above and below, watering holes, buying triggers, dream outcome — assembled from the evidence, in their words, one avatar per run.
  4. Rank the pains. Score each on intensity, frequency, and willingness to pay to remove it. Attach the cost of leaving each one unsolved — that cost is a future headline.
  5. Map the desires. What "solved" looks like in their words, the status and identity payoff behind it, and the before/after gap copy will have to bridge.
  6. Sort the objections. Stated versus real ("too expensive" usually means "I don't believe it will work for me"). For each real objection, name what has to be true in the buyer's head and the proof that installs it.
  7. Write the results into The customer section of marketing-brief.md — ICP notes, avatar, top 5 pains, desired outcomes, objections, language bank — marking every element that came from inference rather than evidence.

Output

The customer dossier, written into marketing-brief.md under The customer. Close with the three weakest spots in the evidence and the single fastest source that would strengthen them — usually five customer calls or one competitor-review sweep.

Rules

  • An invented quote is a landmine. Copy built on fabricated customer language fails silently and expensively — mark every inference as inference.
  • Distinguish what buyers say from what you conclude. Both are useful; confusing them is how personas become fiction.
  • One avatar per run. A blended avatar describes nobody and writes copy for nobody.
  • Weight fresh material higher. A three-year-old review describes a market that may no longer exist — date the bank.
  • Competitor reviews are fair game. Buyers describing a rival's failures are describing your opening, in their own words.
  • The language bank is the crown jewel. When in doubt, spend the extra time there — everything downstream draws on it.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.