#24 How to build an evidence based ICP in hours, not months
A playbook for marketers to define their ideal customer when primary data is scarce with a AI system.
I once heard a founder proudly declare their ICP: “Our customer is everyone with WiFi.”
That’s not an ICP. That’s a dream. And dreams are fine. Until someone has to turn them into a marketing plan and realizes you can’t target “everyone with WiFi” on LinkedIn Ads.
Why ICPs matter
Everything in marketing starts with knowing who you are talking to. Without a clear picture of your ideal client, strategy is guesswork. And guessing in marketing isn’t cheap; it shows up later as wasted budget. Spoiler: the CFO always notices.
An ICP anchors everything: positioning, messaging, and campaigns. But building one based on solid evidence has always been slow work: surveys, interviews, and waiting for enough “real” customers.
The common ICP problem
What if your product is new, your audience is small, or your users don’t have time to give feedback? Maybe the marketing team is stretched thin, and time is scarce.
The roadblock is simple: not enough primary data to learn from.
So the big question is: how do you build a precise ICP based on evidence when primary data is missing?
That’s exactly the problem AI can help solve.
ICPs in hours, not months
Instead of waiting weeks, AI now lets marketers:
Collect huge volumes of data
Analyze it in depth
Turn it into action in hours, not days
This is exactly what
explains in this episode.He shows how to continue evolving the ICPs for Hubspot, breaks them into micro-audiences, and turns them into personalized campaigns at scale.
Why this matters to you
I adapted Kieran’s prompt system for my newsletter use case, and in this article, I’ll show you how to do the same.
This guide is for anyone who needs to define an audience without endless surveys or large datasets: startup marketers, solo creators, growth managers, product marketers, or community builders.
The focus is on building ICPs from secondary sources, external signals, and evidence you can act on. The goal is a profile that’s precise, usable, and ready to guide strategy. I’ll share how I applied this approach and give you 10 practical methods you can use right away.
The magic of well-designed prompts
Kieran’s method isn’t just about using prompts. It’s about designing prompts with a few key principles baked in:
Evidence first. Every field ties to a source (or is marked as TBD).
Recency matters. Focus on the last 45–90 days to capture current pains and language.
Deterministic format. Fixed blocks make outputs easy to compare and automate.
Separation of roles. ICP → Micro-audiences → Campaigns, each step with its own “hat.”
Bias controls. Conflicts resolved by freshness or intent strength; sources must be cited.
Actionability over academia. Every output is immediately usable, whether it’s a card, a CSV, or campaign copy.
These design rules are what make the system powerful. And they can be adapted to many other use cases.
Applying the method: my newsletter as an example
Here’s how I applied this approach to my newsletter. I don’t have thousands of readers yet, but I do know the kind of person I want to reach: curious marketers who want to understand AI through practical, human examples.
Instead of relying only on my small subscriber base, I look outward.
I study the audiences of top creators in my niche, people like Kieran Flanagan or Neil Patel. They publish daily, spark engaged conversations, and build communities that likely overlap with the readers I want to reach.
By analyzing their content and the signals from their audiences, I can extract insights that help me shape a clearer, more actionable ICP for my newsletter.
Step 1: Tailor your template
ICP templates are not one-size-fits-all. The profile for a B2B SaaS company looks very different from one for a B2C e-commerce brand, and different again from what you would need for a newsletter audience.
That’s why starting with the right structure matters. If you don’t already have a clear template, you don’t need to build one from scratch. You can ask ChatGPT to generate a tailored framework, for example:
“Give me a tailored ICP template for a newsletter or blog reader.”Step 2: Use the Super Kieran Prompt
Pick one or two top creators in your niche.
Open Perplexity (I tested ChatGPT 5 and Perplexity; Perplexity performed better here).
Adapt external sources and the ICP template to your case and paste the prompt. Here is mine.
# You are my ICP-Builder assistant.
Your task: assemble a complete Core ICP template for a **Blog reader** by analyzing the public content of a top creator, **Neil Patel**.
Deliverables:
- Return **(1)** the finished template in the exact format shown under **OUTPUT FORMAT** and **(2)** a final **one-pager summary**.
- If any field cannot be found, insert **“TBD.”**
- Use plain language. Avoid jargon in pain statements.
- Where helpful, cite specific posts/videos in parentheses with **Title — URL (Month YYYY)**.
---
## 1) EXTERNAL SOURCE SCOPE (last 90 days)
Analyze only content published in the **past 90 days** from:
- **YouTube Channel** (videos/shorts): https://www.youtube.com/@MATGpod
- **Blog** (articles): https://www.kieranflanagan.io/
If there’s insufficient volume within exactly 90 days, expand to the nearest immediately previous items and label older items as **[>90d]** in parentheses.
---
## 2) METHOD (Field-Mapping Logic)
Extract concrete signals from the last 90 days to populate each section:
- **Reader Persona Snapshot** → Infer role/seniority and “self-story” from who posts address (e.g., “if you run an ecommerce store…”) + CTAs + examples used. Derive archetypes from tone (teacher/coach vs. hacker/optimizer).
- **Topic & Outcome Map** → Cluster recurring themes, outcomes promised, and “before → after” transformations stated or implied in headlines, intros, and conclusions.
- **Reading Context & Routine** → Estimate typical session length and cadence tolerance from article length, posting cadence, and time-to-value patterns (TL;DR blocks, checklists).
- **Discovery & Acquisition** → Identify first-touch channels referenced (search, YouTube, LinkedIn, X), plus referral vectors and lead magnets (templates, checklists).
- **Content Preferences** → From formats actually used (how-to, frameworks, case studies, visuals) and proof style (screenshots, benchmarks, step-by-steps).
- **Value & Motivation** → Translate promises into functional/emotional/social value; write success criteria as “It’s worth it when ___ within ___ days.”
- **Frictions & Risk Reversal** → List top hesitations implied by “objection handling” sections (e.g., complexity, time, cost) and the reframes used (key takeaways, no-fluff policy).
- **Engagement & Retention Behaviors** → Define activation trigger (first save/reply/template use), 2-week habit loop (cue → routine → reward), and “signals of love.”
- **Monetization Fit** → If applicable, note preferred model and upgrade triggers mentioned (courses, premium community, office hours).
- **Segment KPIs** → Propose practical metrics for this reader segment (acquisition, activation, engagement, retention, advocacy, monetization).
- **Signals & Data** → Specify quant/qual you’ll collect from email/blog analytics and VoC prompts.
- **Anti-ICP** → Define disqualifiers from content style and channel preferences (e.g., wants news-only or video-only).
When uncertain, write **TBD**. Keep every field concrete and skimmable.
---
## 3) STYLE GUARDRAILS
- Write in clear, concise English; bullet where appropriate.
- Keep each subsection tight (1–5 bullets max).
- Parenthetical citations: **(Post/Video Title — URL, Month YYYY)** for the 3–5 most load-bearing claims across the doc.
---
## 4) OUTPUT FORMAT (return exactly this block)
# ---------- CORE ICP TEMPLATE ----------
based on the info collected, you have to fill
1) Reader Persona Snapshot
Name & 1-liner: example: “Time-poor marketer seeking practical AI tips”
Archetype(s): Explorer • Achiever • Pragmatist • Creator • Skeptic (pick 1–2)
Primary outcome: “I want ___ so I can ___”
Life/work context: role, seniority, industry, career stage
Identity cues: “People like me read/follow ___”
2) Topic & Outcome Map
Core themes: (e.g., AI workflows, creator economy, demand gen)
Job-to-be-done: “Help me ___ without ___”
Level: beginner • intermediate • advanced
Must-solve problems: top 3 pains your content should relieve
Desired transformations: before → after (bullets)
3) Reading Context & Routine
When/where: commute • morning coffee • lunch break • late night
Cadence tolerance: daily • 2–3×/week • weekly • monthly
Session length: <3 min • 3–7 min • 8–12 min • 12+ min
Device posture: one-hand mobile • desktop deep-dive • inbox-only
Consumption style: scan headlines • save for later • reads line-by-line
4) Discovery & Acquisition
First touch channels: search • X/LinkedIn • TikTok/YouTube • referrals • Product Hunt • Reddit
Creators/brands they trust: (3–5)
Referral vector: friend forward • social share • blog → email capture
Lead magnet sensitivity: checklists • templates • case studies • swipe files
5) Content Preferences (Format, Tone, Proof)
Formats loved: quick tips • deep dives • playbooks • teardown threads • interviews • case studies • visuals
Tone palette: warm • witty • concise • expert • contrarian • community-led (pick 2)
Proof they need: screenshots • benchmarks • step-by-steps • live examples • code/snippets
Length comfort: 300–600 • 800–1,200 • 1,500–2,500 words
CTA acceptance: reply prompts • surveys • tool trials • community join
6) Value & Motivation (Why They Stick)
Functional value: saves time • makes money • reduces risk • learns faster
Emotional value: confidence • belonging • inspiration • control
Social value: “signals I’m ahead” • “gives me talking points”
Success criteria: “It’s worth it when ___ happens within ___ days”
7) Frictions, Objections & Risk Reversal
Top objections: “too long,” “too basic,” “salesy,” “not actionable,” “paywall anxiety”
Reframes: TL;DR block • key takeaways • no-fluff policy • examples first
Trust levers: social proof, testimonials, public roadmap, content samples
Privacy stance: data use, tracking transparency, unsubscribe ease
8) Engagement & Retention Behaviors
Activation trigger: first “save,” first “reply,” first “uses a template”
Habit loop: cue → routine → reward (map first 2 weeks)
Signals of love: 50%+ read rate • clicks to tools • forwards • replies
Community fit: comments • AMA attendance • Slack/Discord activity
9) Monetization Fit (if applicable)
Preferred model: free • sponsor-supported • freemium • paid tier • bundle
WTP band: €___ / month (anchor with value math)
Upgrade triggers: exclusive templates • expert office hours • premium community
Sponsor resonance: categories your reader actually buys
10) Segment KPIs (Set per Persona)
Acquisition: sub conversion rate by channel, cost/sub
Activation: % reading 2 of first 3 sends; first reply/save
Engagement: avg read time, CTR, scroll depth (blog), WAU/MAU
Retention: D30/D90 active rate, churn, “last 3 sends read”
Advocacy: shares/100 readers, referral rate, NPS
Monetization: free→paid %, ARPU, sponsor CTR
11) Signals & Data You’ll Collect
Quant: open/read time, CTR, topic clicks, device, send-time heatmap
Qual (VoC): reply prompts, micro-surveys, comment themes, DM quotes
External: search trends, subreddit threads, competitor newsletters, ad libraries
12) Anti-ICP (Disqualifiers)
Content misfit: only wants news headlines; hates how-to/longform
Channel mismatch: never uses email; inbox-zero zealot; prefers video-only
Value mismatch: ultra price-sensitive to any paid tier; expects daily newswire
Support burden: demands 1:1 consulting replies at scale
---
One-Page Fill-In (copy/paste)
Persona name & 1-liner:
Archetype(s):
Primary outcome (JTBD):
Core themes:
Level:
Must-solve pains (top 3):
Reading routine & cadence:
Device & consumption style:
Discovery channels (top 3):
Trusted creators/brands:
Formats & tone:
Proof needed:
Objections + your reframes:
Activation trigger & habit loop:
Monetization fit & upgrade triggers:
Segment KPIs:
Anti-ICP flags:
# ---------- END TEMPLATE ----------
The Outcome
You’ll get a surprisingly strong starting point.
The Growth-Minded Marketer: A Modern Reader Persona
Reader Persona Snapshot
Name & One-Liner
Growth-minded marketer eager to leverage AI for smarter results
Archetype(s)
Explorer - Achiever
Primary Outcome
“I want actionable AI and growth playbooks so I can scale my impact and stay ahead.”
Life/Work Context
Team leads, heads of marketing, senior individual contributors
SaaS and growth-stage companies
Mid- to late-career professionals
Identity Cues
“People like me read/follow Kieran Flanagan, Marketing Against the Grain, HubSpot, Zapier, Lenny’s Newsletter”
Topic & Outcome Map
Core Themes
AI in marketing workflows
Building growth engines
Content that stands out
Playbooks for changing search/social
Productivity with AI
Job to Be Done
“Help me use AI to get measurable growth without wasting months on failed experiments”
Level
Intermediate–advanced
Must-Solve Problems
Figuring out where AI/automation actually works (not hype)
Standing out as AI floods content channels
Planning/operating when traditional channels (SEO, paid) are less reliable
Desired Transformations
Overwhelmed with new AI tools → Confident, practical daily workflows
Unsure how to apply AI → Runs experiments that get noticeable results
Reactive campaign cycles → Systematic, high-leverage playbooks
(“My Daily AI Playbook: Use Cases, Tools and Prompts” — kieranflanagan.io, Aug 2025 )
Reading Context & Routine
When & Where
Morning routine
Pre-work scan
Campaign planning windows
Cadence Tolerance
2–3×/week
Session Length
3–7 min
Device Posture
Desktop deep-dive or mobile quick-check
Consumption Style
Scans headlines/TL;DR, reads line-by-line if actionable
Discovery & Acquisition
First Touch Channels
X/LinkedIn - Substack discovery - YouTube
Trusted Creators/Brands
HubSpot, Zapier, Marketing Against the Grain, Lenny’s Newsletter
Referral Vectors
Social share
Friend forward
Blog → email capture
Lead Magnet Sensitivity
AI prompts/templates - Execution frameworks - Checklists
Content Preferences
Formats Loved
Deep dives - Stepwise playbooks - Tool-tutorials - Briefings/case studies
Proof Needed
Screenshots - Benchmarks - Step-by-steps
Length Comfort
800–1,200 words
CTA Acceptance
Tool downloads - Reply prompts - Template sign-ups
Value & Motivation
Functional Value
Makes growth experiments easier
Learns new AI workflows
Increases team productivity
Emotional Value
Confidence to lead on “new” marketing
Feels at the tech forefront
Social Value
“Signals I’m an AI-native leader”
Success Criteria
“It’s worth it when I launch a working AI-based process within 14 days”
Frictions, Objections & Risk Reversal
Top Objections
“Too complex”
“Not concrete”
“Too much AI hype”
“Can’t apply as a small team”
Reframes
TL;DR summary blocks
Step-by-step recipes
Real examples, not just theory
Trust Levers
Audience comments - Testimonials - Practical samples
Privacy Stance
Clear data use - Easy unsubscribe
Engagement & Retention Behaviors
Activation Trigger
First AI prompt/template use
Habit Loop
Cue: sees new post/issue
Routine: checks/takes action
Reward: sees workflow or idea work in practice
Signals of Love
50%+ read or forward rate
Tool/template adoption
Post comments
Community Fit
Commenting - AMAs - Event/podcast attendance
Monetization Fit
Preferred Model
Free, sponsor-supported, limited paid templates/tools
WTP Band
€0–€29/month
Upgrade Triggers
Exclusive AI workflows - Peer masterminds - Expert office hours
Sponsor Resonance
Martech - SaaS tools - AI platforms
Segment KPIs
Acquisition: Signup/conversion rate by channel
Activation: % using templates/tools in first 2 weeks
Engagement: Avg. read time, template usage, CTR
Retention: D30/D90 email open, “last 3 sends read”
Advocacy: Forwards/share rate, comments/referrals
Monetization: Free→paid %, ARPU, sponsor CTR
Signals & Data You’ll Collect
Quantitative
Open/read time - Template CTR - Device type - Send timing
Qualitative (VoC)
Quick polls - Comment themes - Prompt replies
External
Search/social topics - Other AI/growth newsletters - Tool usage trends
Anti-ICP (Disqualifiers)
Content misfit: Wants only news, hates frameworks/how-to
Channel mismatch: Never checks email/doesn’t use LinkedIn
Value mismatch: Refuses to try AI tools, demands daily inspiration
Support burden: Expects 1:1 guidance.
10 external sources to enrich your ICP
The outcome is a starting point, something useful and practical when primary data is limited. Here are 10 external sources you can use to enrich your ICP more and adapt to your needs. Combine them, adapt the prompt to your context, and keep refining. The possibilities are endless.
Competitor and peer content mining
Analyze blog posts, case studies, webinars, or YouTube content from competitors and category leaders. Map out who they’re addressing, what pains they mention, and what objections they answer.
Rationale: Content strategy reflects actual targeting and sales narratives, revealing implicit ICP segments.
Social listening and sentiment analysis
Monitor LinkedIn, Twitter (X), Reddit, forums, and review sites for keywords, competitor mentions, and buyer complaints. Analyze sentiment and patterns by job role, industry, or urgency.
Rationale: Real conversations show in-market frustrations and triggers for action.
Technographic and firmographic data aggregation
Use tools like BuiltWith, Wappalyzer, or LinkedIn Sales Navigator to map company size, funding stage, tech stack, and hiring trends.
Rationale: Tech and org stage are strong predictors of readiness for your product.
Job posting and team analysis
Scan job boards or company career pages for roles like “RevOps” or “Head of Automation.”
Rationale: Hiring signals point to scaling pains or maturity levels that match your solution.
Ad and campaign monitoring
Track paid ads from competitors in Meta Ad Library, LinkedIn Ads, or Google. Look at offers, copy, and audiences.
Rationale: Ad budgets are spent on audiences with high ROI—reverse engineer them.
Review mining and voice-of-customer aggregation
Collect reviews from G2, Capterra, or TrustRadius. Cluster job titles, industries, and repeated pain points.
Rationale: Reviews provide real-world, data-driven insights into who buys and why.
Sales leader and rep interviews
Listen to podcasts, Q&As, or interviews with sales leaders in your space. Note anecdotes about “best customers” or common deal patterns.
Rationale: Sales stories surface ICP realities marketers often miss.
Community and event analysis
Map speaker panels or attendee lists from webinars, summits, or niche Slack/Discord groups.
Rationale: Active participants often represent your “super ICP.”
AI-driven lookalike modeling
Feed your best customers into B2B data platforms or simple ML models to find statistical lookalikes.
Rationale: Expands and validates ICP hypotheses with less bias.
Modern competitor comparison
Compare your ICP matrix with competitor positioning. Look for segments they ignore or serve poorly.
Rationale: Helps you find white space where your product fits best.
Final Thoughts
So, no, an ICP is not “everyone with WiFi”. It is not “millennials with disposable income.” And it is definitely not “people who breathe oxygen,” or “humans with a pulse.”
Some people love to dream big. “Our product is for everyone!” Great, so is water, and yet even Evian has a target market.
Let’s stop guessing, stop dreaming, and for everyone’s sanity (including our CFO’s), stop pitching “everyone with WiFi” as a strategy.
Use evidence. Use signals. Use prompts. Because, unlike dreams, those actually convert.
If this article made you laugh, nod, or rethink your “everyone is our customer” pitch, pass it along. Consider it your good deed for the day.
If you want to continue diving into how to build ICPs and personas, let’s have a look at this interesting article I recommend to you from
.
