Workshop 4

Growth & Social Media Acquisition

~3 hours · 6 phases · 3 sheets
Workshop 4 of 7

Growth & Social Media Acquisition

From "we built it" to "people are using it" · 3 sheets · 6 phases
What this workshop is about: In W3 you shipped a working MVP. The product exists. Now the hardest question isn't "does it work?" — it's "who sees it, and why would they click?" Today you will define your exact customer, learn how targeting platforms actually work, write real ad copy, and leave with a complete 2-week test campaign plan.
3
fillable sheets
6
phases
1
campaign plan to submit

Workshop flow

1
ICP — Who exactly is your customer?
Drill from "young people" to a precise, targetable persona. Fill Sheet 1 A+B.
20 min
2
Platform Selection — Where do they spend time?
Platform mindset matrix + decision. Complete Sheet 1 Section C.
15 min
3
Targeting Mechanics — How the algorithm decides
The Targeting Triangle. Custom audiences, lookalikes, learning phase. Complete Sheet 1.
45 min
4
Ad Creative — Write 3 variants that stop scrolling
Hook science. HPSC formula. Creative testing. Complete Sheet 2.
35 min
5
Budget & Unit Economics — What does a customer cost?
CAC vs LTV calculator. The €50 test framework. Sheet 3 budget section.
20 min
6
Campaign Plan — Build your 2-week test
Campaign structure. Week-2 decision gate. Submit Sheet 3.
25 min

Your 3 sheets today

S1
Sheet 1 — ICP & Targeting Card
Phases 1–3 · Define your exact audience in platform-native language
↗
S2
Sheet 2 — Ad Creative Brief
Phase 4 · 3 ad variants with hook, pain, solution, CTA, visual direction
↗
S3
Sheet 3 — Campaign Plan & Budget
Phases 5–6 · Budget calculator, KPI targets, 2-week plan, submit
↗
Tip All sheets auto-save in your browser. You can close and reopen them — your answers persist. To submit, use the Print to PDF button inside each sheet.
"The best targeting in the world can't save a bad product. But the best product in the world dies without targeting."— growth marketing principle
Phase 1 of 6 · 20 minutes

ICP — Who Exactly Is Your Customer?

Sheet 1 · Sections A & B·Goal: a persona your targeting platform understands
Goal: Transform your W3 user assumptions into a precise, platform-native customer profile. "Young entrepreneurs" is not a target audience. "22–28-year-old males in Vilnius who follow Y Combinator and Lex Fridman and have made an in-app purchase in the last 3 months" is.
20 minutes — specificity is not optional. Vague audiences waste money.

The Specificity Ladder

↓
Every step down costs less per click

Each rung makes your audience smaller but more relevant. Relevance = lower cost per click = lower cost per customer.

Avoid"Young people" — 2 billion people. Useless.
Better"18–25 year olds interested in startups" — better, still broad
Target"Women 23–28, Vilnius + Kaunas, follow productivity accounts, made an in-app purchase in last 6 months" — targetable

The three layers of an ICP

Demographics — WHO

Age, gender, location, language, income. The platform's most basic filter. Necessary but not sufficient on its own.

Psychographics — WHY

Interests, values, identity. What they read, follow, and believe. This is where the algorithm gets smart.

Behaviour — WHEN

Purchase history, app usage, engagement patterns. The most powerful layer — it catches people mid-decision.

Action Open Sheet 1 and complete Sections A (Demographics) and B (Psychographics). Return here when done.
S1
Sheet 1 — ICP & Targeting Card
Complete Sections A and B now
↗
Phase 2 of 6 · 15 minutes

Platform Selection — Where Do They Spend Time?

Sheet 1 · Section C (Platform)·Goal: one primary platform for your test campaign
Goal: Each platform rewards a different type of ad. Choosing the wrong one doesn't mean your ad fails — it means your ad can't possibly work regardless of quality.
15 minutes — pick ONE platform for the test. You can expand later.

Platform mindset matrix

IG / Meta
Instagram / Meta
Entertainment & connection

Best for: emotional, visual, lifestyle, and consumer products. Widest demographic range. Strong interest + behaviour targeting. Works best with authentic-feeling creative.

Ad formats: Stories, Reels, Feed posts, Carousels

TT
TikTok
Discovery & entertainment

Best for: under-30 audience, viral potential, entertainment-native products. Ads must feel like organic content — native, fast, and without corporate polish.

Ad formats: In-Feed videos, TopView, Spark Ads (boost organic posts)

LI
LinkedIn
Professional growth

Best for: B2B products, professional tools, high-income decision makers. Most expensive CPM (~€20–30) but highest-intent professional audience.

Ad formats: Sponsored posts, Message Ads, Lead Gen Forms

G
Google Search
Intent — actively searching

Best for: products people already know they need. Highest intent of any platform — user typed your keywords. But requires someone to already be looking for your solution.

Ad formats: Text search ads, Shopping ads, Display

The platform trap most student teams fall into

Choosing the platform you personally use most, rather than where your ICP actually is. If your target customer is 35–50 year old business owners, TikTok is probably wrong regardless of its reach. If your product is a consumer app for Gen Z, LinkedIn is expensive and mostly wasted.

Action Open Sheet 1 Section C now and select your platform. Write a one-sentence justification — "we chose X because our ICP is in [mindset] mode when they'd encounter our solution."
Phase 3 of 6 · 45 minutes

Targeting Mechanics — How the Algorithm Decides

Sheet 1 · Complete·Goal: finish your full targeting spec
Goal: Understand the three layers of paid targeting, how platforms learn who to show your ad to, and how to use your W3 users as a seed for finding more people like them.
45 minutes — the core of today. Theory first (15 min), then complete Sheet 1 (30 min).

The Targeting Triangle

Demographics

Age, gender, location, language, income. The platform's coarse filter. Reduces waste but doesn't find intent.

Psychographics

Interests, values, pages followed. Tells the algorithm who they identify as. The richest signal on Meta.

Behaviour

Purchase history, app installs, engagement patterns. The most powerful — it catches people mid-decision.

The overlap of all three = your high-converting audience

Custom Audiences & Lookalikes — your unfair advantage

CA
Custom Audiences

Upload your existing data — email list, website visitors (via pixel), app users — to Meta or TikTok. The platform matches them to its user profiles. Now you can: (a) retarget people who already know you, (b) exclude them from cold outreach, (c) use them as a seed for Lookalikes.

Even 50 users from W3 is a usable Custom Audience seed
LA
Lookalike Audiences

Tell the platform: "find me people who look statistically similar to this Custom Audience." Meta analyses hundreds of signals (interests, behaviour, demographics) of your seed audience and finds 1–10% of its total user base that matches the pattern.

A 1% Lookalike of your 50 most engaged users is often the highest-converting cold audience you will ever run. This is why collecting real user data in W3 matters.

1% Lookalike = ~500k–2M people who "look like" your best customers

Interest targeting — how to find the right keywords

On Meta Ads Manager — search for these

Instead of broad categories ("technology"), target specific pages, behaviours, and interest clusters your ICP would actually follow:

High signal interests

  • Specific influencers in your space
  • Competitor products/apps
  • Industry publications & podcasts
  • Professional associations

High signal behaviours

  • "In-app purchasers (past 30 days)"
  • "Business page admins"
  • "Frequent international travellers"
  • "Early technology adopters"
Action Complete Sheet 1 now — Sections B (psychographics), C (behaviour), D (platform), E (audience statement), and F (reality check). Aim for an audience size of 100k–2M for a first test. Smaller = too expensive to learn. Larger = too diffuse.
Phase 4 of 6 · 35 minutes

Ad Creative — Stop the Scroll

Sheet 2 — Ad Creative Brief·Goal: 3 distinct ad variants written & tested
Goal: Write 3 ad variants using the Hook–Problem–Solution–CTA framework. Each variant should test a different angle. The algorithm will find the winner — your job is to give it 3 real options to test.
35 minutes — write fast. You'll iterate after seeing real data.

The HPSC formula — Hook → Problem → Solution → CTA

H
Hook (first 1–3 seconds / first line)

Interrupt the scroll. Create tension, curiosity, or recognition. The three types that work best: pain hook ("Stop losing 2 hours a day to this"), curiosity hook ("What if your morning routine took 8 minutes?"), or bold claim ("Most student apps fail because of this one mistake").

P
Problem (1–2 sentences)

Name the pain specifically enough that the right person thinks "that's me." Vague pain ("it's hard to stay organised") has no power. Specific pain ("you open 5 apps every morning and still forget the one thing that mattered") creates recognition.

S
Solution (1–2 sentences)

One clear mechanism. Not features — the outcome the user gets. Not "our app has AI-powered scheduling" — "your morning routine planned in 8 minutes, every day."

C
CTA — Call to Action

One specific next step. Match the temperature of the ask to the relationship: cold traffic = "Try free" or "Learn more." Don't ask strangers to buy before they've seen any value. If you have a free tier, lead with that.

Hook tester — rate your hooks before writing the rest

H

Does this hook stop the scroll?

Enter a hook above and press Rate it to get feedback.

Ask Claude to write your 3 variants

Copy this prompt, fill in the blanks, paste into Claude

I am writing social media ads for my MVP. Product: [one sentence — what it is and what it does] Target customer: [paste your audience statement from Sheet 1] Platform: [Instagram / TikTok / LinkedIn / Google] Goal: [awareness / clicks / app installs / sign-ups] Write me 3 distinct ad variants. For each: 1. A hook (≤125 characters) — try a DIFFERENT angle per variant: Variant A: pain-first ("Stop doing X") Variant B: curiosity/question ("What if you could...") Variant C: bold social proof or result ("Teams like X use this to...") 2. Two sentences naming the specific problem 3. One sentence showing the solution mechanism 4. One CTA phrase (4–6 words) 5. A one-line visual direction Format each variant clearly as A, B, C. Do not explain your reasoning — just write the ads.
Warning The thumb-stop test: Before submitting Sheet 2, read each hook out loud as if you're a tired person scrolling at 11pm. Would YOU stop? If not, rewrite the hook — that is always the problem, not the targeting.
Action Open Sheet 2 now and write your 3 variants. Use the Claude prompt above if you get stuck. The sheet has character count guidance built in.
Phase 5 of 6 · 20 minutes

Budget & Unit Economics

Sheet 3 · Budget section·Goal: know your CAC target before spending a cent
Goal: Understand the math that determines whether your campaign can ever be profitable. The goal is not to spend as little as possible — it's to spend enough to get meaningful data, and no more.
20 minutes — set your numbers before you touch the platform.

The key metrics — know these before touching Ads Manager

CAC — Customer Acquisition Cost

Total ad spend ÷ number of new customers acquired. This is the number you are trying to minimise. If you spend €200 and get 10 customers: CAC = €20.

LTV — Lifetime Value

Average revenue per customer over their entire relationship with you. If customers pay €5/month and stay 6 months on average: LTV = €30. Rule of thumb: CAC should be ≤ LTV ÷ 3 for a healthy business.

CPM — Cost per 1000 impressions

What you pay for 1,000 people to see your ad. Meta average: €6–12. TikTok: €4–8. LinkedIn: €20–35. Google: varies by keyword.

CTR — Click-through rate

Percentage of viewers who click. Social ad benchmark: 0.8–2%. If your CTR is below 0.5%, the problem is almost always the creative — specifically the hook.

Live CAC Calculator — run your numbers now

€50 Test Calculator

Adjust the inputs to model what your test campaign might produce.
Typical: Meta €6–12, TikTok €4–8, LinkedIn €20–35
Benchmark: 0.8–2.0% for social ads
% of clicks that become sign-ups or customers
Average revenue per customer over their lifetime
Daily budget€3.57
Estimated impressions6,250
Estimated clicks94
Estimated conversions3
Estimated CAC€17
Enter your numbers above
Action Open Sheet 3 now and complete the Budget Calculator section using numbers from above. Then set your KPI targets — what CTR, CAC, and conversion count will tell you the campaign is working?
Phase 6 of 6 · 25 minutes

Campaign Plan — Your 2-Week Test

Sheet 3 · Complete & submit·Goal: a plan your team can execute next week
Goal: Write a complete 2-week campaign plan. When it's done, you should be able to hand it to anyone on your team and they could run it without you.
25 minutes — finish Sheet 3. Submit before the session ends.

Campaign structure — how platforms organise ads

1→
Campaign → Ad Set → Ad (this is how Meta structures it)

Campaign: your objective (traffic, conversions, awareness) and total budget. One per product.

Ad Set: your targeting audience + schedule + placement. For your test: 2 ad sets — one narrow (your full targeting spec), one broad (just demographics) to see which the algorithm prefers.

Ad: the actual creative. 3 variants per ad set = 6 total ads for your test.

Structure: 1 Campaign → 2 Ad Sets → 3 Ads each = €50 total test

The Week-2 decision

CTR > target AND CAC < target

Campaign is working. Double the budget. Expand to a 2% Lookalike audience. Start testing a new creative iteration.

CTR good BUT CAC too high

People click but don't convert. Problem is the landing page or onboarding — not the ad. Fix the post-click experience before adding budget.

CTR below target

Creative problem — specifically the hook. The audience is right but the message doesn't land. Rewrite hooks, keep the same targeting, retest.

The 2-week timeline

Action Complete Sheet 3 now — fill in the Week-2 decision gate and the timeline rows. Then get team sign-off and submit.
S3
Sheet 3 — Campaign Plan & Budget
Complete & submit before end of session
↗
"Half the money I spend on advertising is wasted; the trouble is I don't know which half."— John Wanamaker, 1800s. Today's platforms tell you exactly which half — if you set up tracking correctly.