Arrow keys, scroll or swipe

Persona Based Creative Strategy

From the Meta account to the idea list, step by step.

The four stages of the process

01

Meta account analysis

Which ads are running, which ones are efficient, and which audience the account reaches today.

Output · Account findings
02

Research

Verbatim customer language from reviews, forums, Reddit and ad libraries.

Output · Coded quote library
03

Buyer ID™

Personas built from the quotes, with eight evidenced layers and a purchase score.

Output · Persona cards with score
04

Ideas

One idea per persona and pain, each with evidence and a score. A strategist makes the selection.

Output · Idea list, 50 and more

Top ads by spend and by efficiency, side by side

Top ads by spend

The ads with the highest spend over the last 30 to 180 days.

Ad A
Ad B
Ad C
Ad D
Ad E

Top ads by efficiency

The same ads, ranked by CPA and CPM.

Ad D
Ad A
Ad E
Ad B
Ad C

Flag the outliers

Ad D gets little spend and has the best CPA. Ad C gets a lot of spend and delivers little. We look at the message and the hook behind each one.

Read at ad set level

Meta shows one person several ads in a row and credits the purchase to the last one. That is why we evaluate ad sets first, then individual ads.

Build buckets

We group the ads by product, creator, hook tactic and message, and see which group carries the account.

Place every ad by attention and by purchase

Purchase, strong at the top (conversion rate, CPA)

Ugly Duckling

Few people watch, but those who do buy. The message works, the hook is weak. We write new hooks for this message.

Winner

High attention, many purchases. We keep angle, hook and format as winner elements.

Loser

Low attention, few purchases. We record angle and hook as loser elements.

Compare ad and landing page

Many clicks, few purchases. We check whether the landing page delivers on the promise of the ad.

Attention, strong on the right (hook rate, hold rate, CTR)

Measure attention

How many are still watching after 3 seconds (hook rate), how many stay until second 15 (hold rate), how many click (CTR).

Measure purchase

How many buy (conversion rate) and what a purchase costs (CPA), measured 7 days after the click.

Count cold campaigns only

Retargeting shows better numbers because people already know the brand. Only ads to cold audiences go into the matrix.

Analyzing the winning ads

01

Transcribe

We write out every winning ad, spoken words and on-screen text.

02

Pull the core messages

We note which claims the ad makes. Meta reads these claims and decides who gets to see the ad.

03

Build a parts inventory

Hook, scenes, proof, CTA, creator and setting go into a list. New ads are built from these parts.

04

Tag the psychological format

Every ad gets its format: Recognition, Social Mirror, Reframe, Specificity or Earned Confidence. This shows which format runs twice and which one is missing.

05

Read the comments

We read every comment under the top 20 ads, the angry ones included. Objections in the customers' own words go into the quote library.

06

Write the findings per ad

For every ad we record which angle, hook, message and format carries it. This becomes the list of winner elements and gaps.

The running ads reach one part of the audience

Reached by the current messageBuys the category, sees themselves in no ad
1Persona
In most accounts, every winning ad speaks to the same person, in the same situation, with the same argument.

Find the missing buyers

We compare the people in the ads with the people in the reviews. Whoever buys but appears in no ad gets an idea of their own.

Why growth stalls

Meta groups similar ads together and throttles them. New reach comes from a new persona, a new situation or a new argument.

The search starts at the product and ends at the category

Category
Competitors
Product

Product, inner ring

Own reviews, comments under own ads, post-purchase surveys, support tickets.

Competitors, middle ring

Competitors' 1-star reviews, ad library, landing pages, offers.

Category, outer ring

Reddit, YouTube comments, forums, Google page one, the study base.

The research process, from the product outward

Internal

Account, own reviews, customer data.

Reputation

Google page one, press, social, founder story.

Competitors

Ad library, landing pages, 1-star reviews.

Audience

Reddit, YouTube comments, forums.

Evidence and compliance

For health, beauty and food. Study base per claim.

Name the decision before every search

Before every search we write down which decision it is meant to enable.

Save verbatim, with link and date

Every finding is stored in its original wording, with link and date. These sentences later become hooks.

Search until nothing new comes up

We collect until nothing new appears, usually beyond 100 quotes. If a source delivers fewer than 20, we flag it as thin data.

The platforms and what we look for there

01

Own shop reviews

All reviews from Yotpo or Judge.me, every star rating. Each review gets a quality score from 1 to 5.

02

Customer data

Post-purchase survey ("why now"), support tickets, quiz answers.

03

Google page one

We google the brand plus "reviews", "test" and "scam" and note what stops a first-time buyer.

04

Competitor reviews

The 1-star reviews of competitor shops. That is where the others' broken promises are written down.

05

Ad libraries

Meta Ad Library and TikTok Creative Center, Foreplay for the running time of ads.

06

Reddit

The category's subreddits, in German and English. We look for first-person stories, the search for alternatives, and hate.

07

YouTube comments

Comments under review, test and debunking videos about the category.

08

Forums

Depending on the category: onmeda, paradisi, motor-talk, gutefrage or Stack Exchange.

What each star rating of a review delivers

5 stars
The result in the customer's words
4 stars
Result with a caveat
3 stars
Expectation against result
2 stars
Broken promises
1 star
Fear, objections, anger

Each review gets a quality score from 1 to 5. We take quotes from score 4 upward.

5 stars

Customers describe the result in their own words. This is the material for hooks.

1 and 2 stars

Own reviews hold the fear before buying, competitor reviews hold the broken promises. Material for objection ads.

3 stars

Customers write what they expected and what arrived. Material for ideas that explain the mechanism.

Reading ad libraries by running time and repetition

Running time of an ad in the ad library

Switched off after about 8 days
Fail
Active for under 30 days
Caution
44 days and longer
Winner
60 days and longer
Mature market

We cannot see other brands' numbers. The evidence is repetition: the same hook three or four times, the same creator more than once.

Build a list of 10 to 30 accounts

Direct competitors, brands with the same audience from other categories, one or two role models.

Note the gaps

Audiences nobody addresses. Results nobody promises. Objections nobody answers.

Compare the advertised with the real customers

We compare the people in a competitor's ads with the buyers in their reviews. Buyers without a matching ad become an idea.

The market's sophistication stage and the rule for each stage

1

New market

Nobody has heard the promise.

RuleThe simple, direct claim is enough. Say what the product does.
2

First imitators

The claim is known, the market still responds.

RuleEnlarge the claim. Faster, more, longer than the others.
3

Claims worn out

Big promises have been running everywhere for 60 days and longer.

RuleThe mechanism. Explain how and why the product works.
4

Mechanisms known

Everyone explains a mechanism. German supplements sit here.

RuleExtend the mechanism and set it apart with diversifiers. Bioavailability, lab testing, a new form.
5

Everything heard

Claims and mechanisms no longer pull.

RuleIdentification. The person, their situation, their problem and their experience with the product carry the ad.

How we determine the stage

We read in the ad library which claim stage has been running at competitors for 60 days and longer. The stage above that is the one we need.

The stage describes the whole market

The stage applies to the whole category. It tells us which kind of message still works there. Which single message is free or taken is shown afterwards by the Desire Map.

The awareness pyramid and the messaging for each stage

Most aware

Offer, reason to buy now, guarantee. Keep it short.

Product aware

Proof, comparison, answers to objections.

Solution aware

The path to the solution and why our mechanism is better.

Problem aware

Name the problem, why it occurs, what it costs.

Unaware

A situation and a symptom from everyday life, without the product.

At the bottom, many people and a lot of explaining. At the top, few people and a short path to purchase. Every quote and every idea gets its stage.

Coding every finding as a record

quoteverbatim, unabridged, in the original language sourcelink to the finding, date of the post, platform, star rating bucketpain, trigger moment, objection, transformation, striking language, wished-for solution, price talk, search for alternatives, doubt whensituation with place and moment whythe emotion behind the problem, positive or negative, intensity 1 to 5 whowhat this person shares with others affected desirewhich core desire the sentence touches awarenessunaware, problem, solution, product, most aware phasetrigger, exploration, evaluation, purchase

The situation is mandatory

"She wants to lose weight" is a fun fact. "She gets changed in the bedroom before the holiday" is an insight.

Count per bucket

We count per category. What is said most often gets the highest priority.

Group into narrative patterns

We cluster the findings into narrative patterns: problem first, before and after, revelation. The hooks grow out of these.

The core desires behind every purchase

Survival, health

Anti-aging, prevention

Food, enjoyment

Taste, ritual

Freedom from fear and pain

Worry gone, safety

Attraction, partnership

Attractiveness, relationship

Comfort

Relief, home

Being superior, keeping up

Performance, comparison

Protecting loved ones

Family, children, pets

Approval, belonging

Avoiding shame, showing up

Eight innate core desires. Every ad attaches to one of them.

Being informedCuriosityCleanlinessEfficiencyConvenienceReliabilityBeauty, styleSaving, valueBargains

Nine learned wants, useful as a secondary motive.

One product serves several desires

A red light panel serves health, freedom from pain, attractiveness and comfort. Each desire is its own field for ads.

Assign a desire to every finding

Every quote gets its dominant desire. When in doubt, the field stays empty.

Mark the desires nobody plays

A desire with many findings that no competitor plays. That is where the first ideas start.

Findings per desire set against what competitors offer

The bars show how many findings a desire has in the quotes. Next to it, how many competitors play it. Schematic with sample values.

Tranquility
6 competitors
Recognition
0 competitors
Status
4 competitors
Family
1 competitor
Order
5 competitors
Saving
7 competitors

Orange marks desires with many findings and little supply.

Two directions per motive

Away from a feared state, towards a desired one. Every persona has both.

Motives with plain names

On the cards, the motives are called Tranquility, Recognition, Status, Family, Curiosity, Order, Saving or Independence.

Bars with a data base

Without coded findings, a motive counts as a hypothesis and gets no bar.

The eight layers of a Buyer ID™

Identity

How they see themselves, where they belong, what they believe.

Deeper wish

What she really wants underneath the wish for the product.

Trigger now

Why right now: a life event, a social comparison, a tipping point or a break in the daily routine.

Purchase context

What she has already tried. Which objections she brings along.

Micro-moment

Place, time, emotion. In private or in front of others, acute or chronic.

Conditioning

How she grew up, how she handles money, whom she believes.

User and buyer

Who uses, who pays, who decides.

Reachable future self

The image in the ad is at most two steps ahead of where she is today.

Evidence every layer

Every layer has a verbatim quote or the label "hypothesis".

Name by the pattern "role, core state"

"Woman 45+, mature skin", "Nurse, shift work".

What the one-liner lacks

"Working mom, tired" has no situation, no trigger, no objection.

Purchase likelihood from proximity, heat and lead motive

Purchase proximity

How close the persona is to buying, read from the awareness stage of her quotes.

Most aware, product aware100
Solution aware70
Problem aware40
Unaware25
+

Heat

How much the problem hurts and how much is at stake, evidenced with quotes.

Very high100
High85
Medium55
Low40
+

Lead motive

How strong this persona's most important desire is, intensity 1 to 5 times 20.

Intensity 5100
Intensity 480
Intensity 360
Intensity 240
÷ 3

Purchase likelihood

Sample calculation

0%
Purchase proximity70
Heat85
Lead motive82

Heat map before the score

Before scoring, we check every persona for urgency and stakes. Only personas with strong leverage move on.

What the score does

The score decides which persona is addressed first. It says nothing about conversion.

The finished Buyer ID™ as a card

P1 · Woman 45+, mature skin 1

Problem awareapprox. 45 to 60Formal address2
74%
Purchase likelihood, modeled3
Away from
Tranquility fear of the next flare-up in the morning before work
Towards
Recognition facing the camera on a video call without explaining herself
Towards
Independence hypothesis from quotes, not coded
4

Portrait in prose with verbatim quotes woven in. At the end, one line that sets her apart from the neighboring persona.5

1Name by "role, core state", short code, gender.
2Awareness, age range, form of address.
3Purchase likelihood from proximity, heat and lead motive.
4Motive bands. Lead motive in orange, length equals number of findings.
5Portrait in prose, the eight layers as a checklist behind it.

Every moment of the product story can be an ad

01The problem
02The trigger
03The solution
04The proof
05The guarantee

We write down the story of the product and split it into moments. Every moment is an ad of its own.

An idea consists of a core sentence and a hypothesis

Core sentence, in the customer's words

"Your dermatologist ruined your skin."

Hypothesis

Speaks to women whose skin is worse after cortisone than before. The evidence is 1-star reviews with the line "it has been worse since the cream".

Every idea connects a pain with a Buyer ID™. The core sentence sounds the way the customer would say it to a friend.

Failed solution

What she has already tried and why it did not help.

Own mechanism

Why the product solves the problem differently from everything before.

Awareness stage

Which awareness stage the idea is written for.

Market maturity

In a saturated market the idea needs a mechanism. A promise alone is not enough there.

Claims traffic light

Every health-related statement is checked for compliance before it goes into an ad.

Evidence

Every idea has a verbatim quote as evidence or is marked as a bet. Without either, it gets cut.

Building the relevance score from frequency, heat, whitespace and traffic light

1to 5
Relevance score per idea, reasoning in one line

Frequency

How often the pain appears in the quotes.

Heat of the Buyer ID™

How urgent it is for the persona.

Whitespace

How few competitors run the message.

Claims traffic light

Whether the statement is compliant, a borderline case or blocked.

Check against earlier tests

Every idea gets the status winner, loser, open or new. Winners are iterated, losers are cut.

Allow wildcards

Unusual connections are welcome. They are marked as wildcards and get the same score.

Include repositionings

Ideas that carry the product into an entirely different audience, with an estimate of how much spend they can support.

The rest goes to the backlog

Ideas not selected wait, with their score, for the next cycle.

Every Buyer ID™ gets its own message

P1, current messageP2P3P4P5P6
0Buyer IDs™
Each with its own situation, its own objection, its own motive, two to three ideas.

Meta reads the ad as targeting

Who appears in the ad, which setting is shown and which words are spoken decide whom Meta shows the ad to.

Six messages for one product

Every persona sees her argument in her language. The reach adds up.

One touchpoint lifts one or two awareness stages

Unaware
Problem aware
Solution aware
Product aware
Most aware
Ad
Funnel
Example. The ad picks up a problem-aware person and makes her solution aware. The funnel picks her up there and takes her to product aware, and to most aware if needed.

At least one stage

A touchpoint that leaves the person on the same stage has explained nothing.

At most two stages

More education in one step overwhelms. The rest belongs in the next touchpoint.

Ad and funnel interlock

The funnel starts on the stage where the ad ends. It can carry more than the ad alone.

A strategist makes the selection of ideas

01Account findings
02Quote library
03Desire Map
04Buyer ID™ with score
05Idea list, scored
06Selection by strategist
07Funnel per idea
08Concepts, hooks, scripts
09Test, learnings fed back

A strategist selects from the list

The score sorts. Which ideas go into the funnel is decided by a strategist.

A strategist thinks along at every step

Findings, quotes, personas, ideas: every stage is checked and refined by a strategist.

Funnel before creatives, learnings fed back

First the page, then the creatives. Every test writes winners and losers into the brand's memory.