M2.L6 · Customer Research
Synthesis: from raw notes to Ideal Customer Profile
What you'll be able to doSynthesise research inputs into a one-page, evidence-linked Ideal Customer Profile and revisit the M2.L1 assumption audit.
Learn
A folder of notes is not research
You now have raw material: two or three interviews' worth of notes, ten verbatim quotes, perhaps survey results. Left as a folder, this is worth nothing except the feeling of diligence. Synthesised into one page, it becomes the highest-leverage artefact in your entire Plan — the page that decides your positioning in Module 3, your content topics in Module 5, your keywords in Module 6, your ad targeting in Module 9.
Synthesis is a craft with actual moves. This lesson teaches them: group, name, count; separate what you observed from what you inferred; choose one beachhead; and then face the five beliefs you wrote down in Lesson 2.1.
The craft: group, name, count
Spread everything out — index cards, sticky notes, a spreadsheet, it doesn't matter. One finding or quote per card. Then:
Group. Move cards that seem to say the same thing next to each other. Don't force it; let clusters form and let orphans stay orphans.
Name. Give each cluster a name in customer language, not marketing language. "Breakage worry" if that's what they said — not "logistics concerns". The moment you rename a cluster in your own vocabulary, you start losing the data.
Count. Write next to each cluster how many independent sources feed it. Two quotes from the same interview are one source. This number is your protection against the biggest hazard of small-scale research: promoting an anecdote to a pattern. The working rule: a theme needs about three independent sources before it earns a line in your profile. One mention is a lead. Two is interesting. Three, across different methods if possible, is a pattern you can build on.
With a handful of interviews you cannot do statistics, and you shouldn't pretend to. What you can do is triangulate: when an interview, a mined review and a survey answer all point at the same thing, three weak signals become one strong one. That is how small-n research earns its conclusions.
Observed, inferred, assumed — keep the three visible
The subtlest synthesis skill is refusing to let these blur.
Observed: what someone actually said or did. "One buyer found us through a wedding-gift guide." Inferred: what you conclude from observations. "Gift guides may be a watering hole for our buyers." Assumed: what you still believe without evidence. "Gift-buyers will pay more than self-buyers."
All three are allowed in your profile. What's not allowed is disguise — an inference wearing an observation's clothes, or an assumption presented as a finding. That's why every line of the profile you build today carries a source tag, and why "assumption" is a permitted tag that stays visible. An honest "assumption" today becomes a research question tomorrow. A disguised one becomes a wasted quarter.
Ostara, synthesised
Mateus's inputs: two interviews, ten mined quotes from Lesson 2.3, and the 63-response survey from Lesson 2.5. Grouped and counted, three clusters dominate:
"Gift-buying with taste anxiety" — 9 independent sources "Instagram sameness fatigue" — 6 sources "Breakage worry" — 5 sources
Look at what happened to Ana's belief. The pack records it: buyers are women 28–45 in Portugal, Spain and France who buy for their own homes or as gifts, and find us on Instagram. The evidence says the segment that keeps appearing is older and wider — roughly 30–55 — and, decisively, that the gift occasion isn't a sideline; it's the centre. People buying wedding and new-home gifts, anxious to show taste without risking the recipient's, actively avoiding "everyone's Instagram plates", worried the parcel arrives in pieces.
So the beachhead Mateus chooses is the considered gift-buyer, 30–55, buying for a wedding or a new home — not the self-purchase design lover Ana assumed. Not because self-purchasers don't exist or aren't welcome, but because the gift-buyer has the strongest pain (taste anxiety has 9 sources), a nameable trigger, and reachable watering holes. The ripples run through the whole Program: gift-guide content and search terms in Module 6, "arrives safely, beautifully" messaging in Module 3, occasion-based email flows in Module 8.
Two of Ana's five Lesson 2.1 beliefs did not survive. That is the module working.
An ICP is not a persona
You may have seen personas: a stock photo, a first name, a list of hobbies. Decoration. An Ideal Customer Profile is a decision tool, and it has seven load-bearing lines:
- Who — only the details that change your marketing
- Trigger — the event that starts the buying journey
- Job — the progress they're hiring you for, carried from Lesson 2.2 and revised to fit the evidence
- Pains — the top two or three, in their words
- Alternatives — what they'd actually do instead, always including the non-purchase option
- Objections — what nearly stops them
- Watering holes — where they already look or gather when the trigger fires
Every line carries a source tag. The trigger deserves special respect: "a wedding invitation lands" or "a new client signs the engagement letter" tells you when and where to show up, which is worth more than any demographic. And the profile describes one segment — the beachhead you'll centre the strategy on first, chosen on pain, reach and economics. The other segments remain welcome buyers. They just don't get to steer.
Now revisit your two channels
In Lesson 1.3 you shortlisted two channels — deliberately before you had research, because commitment couldn't wait. That was a loan. Today it comes due.
Put your shortlist next to your watering_holes lines and ask: knowing where these people actually look when the trigger fires, would I still pick these two? For Ostara: email keeps its slot without argument — an owned list of buyers, and occasion-driven gift-buying is exactly what email flows serve. But the evidence about gift guides and gift-idea searches puts a named challenger next to social organic: content and search. Mateus's honest verdict is "confirmed, challenger noted" — Instagram stays for now, and gift-guide content is the first thing Module 6 will test. Confirming a channel with reasons is as valid an outcome as swapping one. What's no longer valid is a shortlist that has never met your evidence.
Kill your darlings
Last move. Your five beliefs from Lesson 2.1 come back on screen, and you stamp each one: Confirmed, Revised, or Killed.
This is the emotional core of the module, and the temptation is to be gentle with yourself. Resist it. A killed belief is not a mistake — it's a wasted euro you'll never spend now. An audit that returns five Confirmed and no notes usually means the research was too polite to threaten anything, and it will be queried. Somewhere in your notes there is probably a sentence that contradicts something you were sure of in Lesson 2.1. Find it. Stamp it. That stamp is what you came to this module for.
Now build the page.
Do
Exercise 2.6.1 — Your Ideal Customer Profile
Build your Ideal Customer Profile. Every line carries a source tag — "assumption" is allowed but stays visible. Then judge your five original beliefs, and re-test the two channels you chose in Module 1. This is the Module 2 Project.
Write these down — in your plan document, or on the worksheet at the end of this lesson. Tag every line of the profile with its source: interview · desk research · survey · own data · assumption.
| What to write | Guidance |
|---|---|
| Key findings — five to twelve | Your evidence table. Each: the finding, 5–60 words, verbatim where possible and anonymised (roles, not names), plus its method — interview 1, 2 or 3 · desk research · survey · own data. At least two findings from interviews to pass; three interviews represented is the distinction path |
| Segment name | e.g. "The considered gift-buyer" |
| Who | 20–60 words. Firmographics (B2B) or life situation (B2C). Only details that change your marketing |
| Trigger | An event, not a demographic. What happens that starts the buying journey? |
| Job statement | Carried over from Lesson 2.2, revised to fit the evidence |
| Top pains — two or three | In their words |
| Alternatives — two or three | Include the non-purchase option — doing nothing, DIY, a voucher |
| Objections — two or three | What nearly stops them |
| Watering holes — two to four | Where they already look or gather when the trigger fires |
| Verdicts on your five Lesson 2.1 beliefs | Confirmed · revised · killed, each with a short note. All five confirmed with empty notes? Really — what threatened them? |
| Channel verdict | Both M1.L3 channels confirmed · replacing one · replacing both. Put your Lesson 1.3 shortlist next to your watering-holes lines — the comparison is the exercise |
| Channel note | 20–80 words. Argue it from the watering holes. "Confirmed, with a challenger noted" is a legitimate verdict |
When you're done, count the share of lines tagged something other than assumption — aim for 60% or more.
Where this goes: section 3 — Customer — of your Marketing Plan, in full: section 3.1 gains its verdict column, and the ICP becomes section 3.5. If your channel shortlist changed, update section 2.3 too. And reopen your Module 0 brief: if it guessed your buyer was X and your research says Y, revise it — that revision is the Program working, not a correction.
Check
Rubric
Mark your own work against these criteria, scored 1–10.
| Criterion | 8–10 | 5–7 | 1–4 |
|---|---|---|---|
| One beachhead | A single, pictureable segment, chosen on pain, reach and economics | One segment, but fuzzy at the edges | "Everyone who…", or three segments in a trench coat |
| Evidence linkage | 60%+ of lines carry non-assumption tags that genuinely support them | Tags present but some are decorative | Untagged claims, or assumptions dressed as findings |
| Trigger quality | A specific event that starts the buying journey | A vague circumstance ("has back pain") | A demographic posing as a trigger |
| Real alternatives | Includes what people actually do instead — with the non-purchase option | Competitors only | Missing, or "no real alternatives exist" |
| Honest verdicts | At least one Revised or Killed, substantiated, visibly reflected in the ICP | Verdicts given, notes thin | All Confirmed, no notes, no threat to any belief |
Pass: 5+ on every criterion. Distinction: 8+ on all five.
Quiz
Four questions. Pick an answer to see whether you were right.
1. What makes an ICP different from a classic marketing persona?
- a) The ICP has a photo and a memorable name
- b) The ICP is a decision tool where every claim is evidence-linked; personas often decorate assumptions ✔
- c) ICPs are only for B2B
- d) Nothing — the terms are interchangeable
Why: the test of an ICP line is "what decision does this change, and how do I know it?" A stock photo and a list of hobbies answer neither.
2. Why choose ONE beachhead segment when three segments buy from you?
- a) GDPR limits how many segments you may target
- b) The other segments must be refused as customers
- c) Focus: strategy, message and channel choices sharpen around one segment first — others remain welcome buyers ✔
- d) Because surveys can only measure one segment
Why: a message aimed at three segments at once is aimed at none of them. The beachhead steers the strategy; it doesn't police the till.
3. An interviewee said: "I found you through a wedding-gift guide." Which reading is correct?
- a) Observation: gift guides are our best channel
- b) Observation: one buyer arrived via a gift guide. Inference: gift guides may be a watering hole — check it against other sources ✔
- c) Observation: gift-buyers don't use Instagram
- d) Nothing usable — one person proves nothing
Why: keep what was observed separate from what you infer from it. One data point is a lead, not a law — but (d) overcorrects; a lead is exactly what tells you where to look next.
4. A theme appears in one interview and nowhere else. What belongs in the ICP?
- a) The theme, as a headline finding — interviews are the highest-grade evidence
- b) Not yet: a theme needs roughly three independent sources to earn an ICP line — log it and look for corroboration ✔
- c) Nothing — single mentions are noise and should be deleted
- d) The theme, once a survey confirms it with n≥100
Why: patterns over anecdotes. A single mention is neither a finding nor noise — it's a candidate. If it must appear now, it appears tagged for what it is, not promoted.
Advance
Module 2 complete. You started this module with five guesses. You end it with an evidence-linked profile of one real segment — their trigger, their job, their words — plus a channel shortlist that has finally met your customers, and at least one belief honestly buried.
Next: Module 3 — Positioning, Brand & Offers. Six lessons. Your ICP's alternatives line becomes the raw material: positioning starts from what your customer would do without you.
Mark your own work
| Good | Not yet | |
|---|---|---|
| Grouped, named, counted | Clusters named in customer language, with source counts | Themes named in marketing-speak, counts absent |
| Nothing disguised | Observed, inferred and assumed are visibly separate | Inferences and assumptions wearing evidence tags |
| One beachhead | A single segment you could pick out of a crowd | "Everyone who…" |
| Channels re-tested | Shortlist confirmed or changed with reasons from watering holes | Shortlist never re-examined |
| A belief died | At least one Revised or Killed, and the ICP shows it | Five polite Confirmed stamps |
Worksheet
THE SCHOOL OF NET MARKETING
Lesson 2.6 — From raw notes to Ideal Customer Profile
STEP 1 — GROUP, NAME, COUNT
Cluster name (their words) independent sources
1 ______________________________ _____
2 ______________________________ _____
3 ______________________________ _____
Rule of thumb: ~3 independent sources = a pattern.
1 = a lead. Same interview twice = one source.
STEP 2 — THE PROFILE (tag every line:
I interview · D desk · S survey · O own data · A assumption)
Segment name: _________________________________
WHO (only what changes your marketing) tag
____________________________________________ ___
TRIGGER (an event, not a demographic)
____________________________________________ ___
JOB (from Lesson 2.2, revised)
When ______________ I want to ______________
so I can ___________________________________ ___
TOP PAINS (their words)
1 __________________________________________ ___
2 __________________________________________ ___
ALTERNATIVES (include the non-purchase option)
1 __________________________________________ ___
2 __________________________________________ ___
OBJECTIONS
1 __________________________________________ ___
2 __________________________________________ ___
WATERING HOLES (where they look when it fires)
1 __________________________________________ ___
2 __________________________________________ ___
Non-assumption tags: ____ of ____ lines (aim: 60%+)
STEP 3 — CHANNEL RE-TEST (your two from Lesson 1.3)
Channel 1 ______________ ☐ confirmed ☐ replaced by ______
Channel 2 ______________ ☐ confirmed ☐ replaced by ______
Because (from the watering holes): ____________________
_______________________________________________________
STEP 4 — VERDICTS on my five Lesson 2.1 beliefs
1 ☐ Confirmed ☐ Revised ☐ Killed note: ______________
2 ☐ Confirmed ☐ Revised ☐ Killed note: ______________
3 ☐ Confirmed ☐ Revised ☐ Killed note: ______________
4 ☐ Confirmed ☐ Revised ☐ Killed note: ______________
5 ☐ Confirmed ☐ Revised ☐ Killed note: ______________
Five Confirmed and no notes? Look again.
SELF-CHECK
☐ Clusters counted, in customer language
☐ Observed / inferred / assumed kept separate
☐ One beachhead, with an event trigger
☐ Channels re-tested against watering holes
☐ At least one belief Revised or Killed
Next: Module 3 — Positioning, Brand & Offers.
theschoolofnetmarketing.com/learn/positioning-the-context-that-makes-value-obvious