Levi Steede

If ads still address the same macro persona, pain point, promise, awareness level and core desire, they call it fake diversity and link it to flat spend — the account is fishing in the same pond.

• Real diversity changes one of four things: who the ad addresses (persona or job-to-be-done), why they would act (desire versus fear, speed versus certainty, convenience versus status), where they sit in awareness and objection state, or what is promised.
• Visual and structural differences can affect how Meta classifies ads under separate entity IDs, but they are a distribution mechanic; the goal is incremental reach.
• Volume belongs to the same argument: launching only two or three ads a week starves the system of signals, but different hooks on one idea produce volume without new demand.

The original contrasts what brands say — “let's make 15 new hooks”, “let's swap the creator” — with what stays identical across those ads, then notes that most brands rely almost entirely on product-aware messaging and ignore the largest audience pools.

Read original · Original 2

Most accounts they audit repeatedly do the second job and call it strategy.

• The failure mode: cutting demand generation when performance dips.
• Capture numbers look strong for a few weeks, then stall because the pipeline has no top.
• Steede also observes that unaware and problem-aware messaging is underused relative to product-aware ads, even though the larger pools sit there.

“Kill your demand generation when performance dips and your demand capture numbers will look great for a few weeks.

• Then they won't.
• The budget isn't the problem.
• The range is.”

Read original · Original 2

Steede argues this matters most when every DTC brand is competing for the same attention and consumers are sceptical of polished scripts and AI-written angles.

• They report that creator-led partnership ads carry a disproportionate share of spend in many accounts.
• The process: recruit creators, track which audiences respond, then whitelist top performers' content as Meta partnership ads for cold traffic.
• The channel needs runway; trying to build it during a peak week leaves no time to earn trust.

A fashion brand was six weeks out from Black Friday with nothing in place.

• Affiliates and creators were recruited through UpPromote, links went live, response was tracked, and the top performers' content ran as Meta partnership ads whitelisted through the creators' handles.
• Steede reports click-through rate up 40%, cost per click down 20%, conversion rate up 25%, and cost per acquisition from €41 to €26.

Read original · Original 2

The point: most brands think they need new ads when they need a better way to decide which ads to make. The agency handles ads management, creative strategy, filming, launching and reporting; the client approves concepts and keeps inventory current.

Steede's introduction post to the ADF Data × Creative Scaling System describes three phases — Foundation, Engine, Scale — and offers free growth audits to brands that apply.

Read original · Original 2

The old default started with original UGC, which meant weeks of planning and production before an account produced any learning.

• The new default starts with proven creative — the client's own working material or a close analog from a similar account — and invests in new footage once there is signal.
• The principle is unchanged; the tradeoff is faster learning through iteration at the cost of immediate originality.

Steede says around 70% of new-client concepts used to be UGC-led, with a new creator, brief and shoot taking about five weeks before the first ad ran. After the change, the same accounts produce real signal in days instead of weeks.

Steede describes a change in the agency's approach to the start of a client relationship, not a replacement for original production.

Read original

Consumers are more sceptical, and polished scripts and similar angles are everywhere, so authenticity can matter more than production quality. The point aligns with their partnership-ad argument: creator-led content cuts through that scepticism.

The five-lessons post closes with “ugly often wins”, contrasting raw, native storytelling with polished creative.

Read original

By contrast, designers and editors earn less while contributing across the production process.

• They endorse hiring a strategist who can assess where an account is underleveraged, verify that with a growth strategist, then write the hypothesis, copy and design the whole ad.
• Buyers should stress-test whether they are paying for strategy or another AI subscription.

The agency put a senior designer through Stefan Georgi's RMBC2 course — a named third-party copy course — and reports that the designer could then take an account from diagnosing underleveraged areas through to a finished ad.

Read original

A useful starting point before the more detailed system posts.

Steede introduces themselves as the founder of Adfluent, a paid social agency for ecommerce brands, and says it has generated over €200M in sales for clients.

• They argue that most brands think they need new ads when they actually need a better system for deciding which ads to make and scaling the ones that work.
• They call that system the ADF Data × Creative Scaling System.

For brands planning a campaign

Frame an agency or in-house evaluation around the decision system, not just ad production: ask how creative choices are made, tested, and scaled.

Start here for the operational shape of Steede's offer.

Steede describes Adfluent's three-part system.

• The Foundation phase examines the brand's numbers, category, and customers, down to why each person buys.
• The Engine is a data-driven creative testing system in which every ad launches with a clear hypothesis and every result informs the next round.
• The Scale phase expands winners across creative formats and micro-personas within an account structure built for Meta's Andromeda algorithm.
• Steede says the agency handles ads management, creative strategy, filming, launching, and reporting.

For brands planning a campaign

Use the three phases as a checklist when building a creative testing plan or assessing a vendor: foundation research, a hypothesis-driven testing engine, and a scaling structure.

A compact diagnostic for anyone planning paid social creative.

Steede says most DTC ad accounts they audit repeat the same creative, which mostly converts people already close to buying.

• They argue that an account needs to do two jobs at once: generate demand by raising awareness of a problem the viewer has not fully named, and capture demand by converting existing intent.
• Demand generation is harder, colder, and usually the first thing cut when performance dips.
• Cut it, and demand capture numbers look strong for a few weeks, then collapse.

For brands planning a campaign

Allocate explicit budget and briefs to problem-aware and unaware angles alongside product-aware conversion ads, so the account does not run out of demand to capture.

Steede argues that most brands mistake more hooks, different creators, or changed edits and captions for creative diversity.

• If ads share the same persona, pain point, promise, awareness level, and core desire, that diversity is fake.
• True diversity changes the persona or job-to-be-done, the core motivation (for example, desire versus fear or speed versus certainty), the awareness level or objection state, or the primary promise.
• Visual and structural differences can affect how Meta classifies ads, but incremental reach comes from demand differentiation.

For brands planning a campaign

Brief for variation across persona, motivation, awareness stage, and promise, not just visual changes. This helps ads expand coverage instead of competing with each other.

A quick read for benchmarks and operating principles.

Steede lists five lessons from generating over €52M in revenue from €24M in ad spend in 2025.

• First, increase creative volume; only two or three new ads per week starves the system.
• Second, volume alone is not enough: if ads are just different hooks on the same idea, Meta groups them under the same entity ID, so messaging and awareness levels need to change too.
• Third, unaware and problem-aware messaging is underused.
• Fourth, partnership ads reduce scepticism and often carry a disproportionate share of spend.
• Fifth, raw, native-looking ads often win, and storytelling remains the most scalable skill in paid media.

For brands planning a campaign

Increase creative volume, add creator-led partnership ads, and brief raw storytelling formats for cold audiences rather than relying only on polished product-aware ads.

Steede says they used to make 70% of new-client concepts UGC-focused, with a new creator, brief, and shoot before any ad went live.

• The process took about five weeks before the first ad ran, leaving the account without new learnings.
• They changed their approach: they still create UGC, but start by iterating on creative already proven for the client or a close analog from a similar account.
• The same principle applies, but execution is faster.
• The bigger win is faster learning cycles, with real signal in days instead of weeks.

For brands planning a campaign

At the start of an engagement, use existing winners or comparable-account creative to get signal quickly, then invest in original UGC shoots once hypotheses are validated.

Steede contrasts brand ads, which ask someone to trust you, with partnership ads, which borrow existing trust.

• In a fashion brand case six weeks before Black Friday Cyber Monday, they used UpPromote to recruit affiliates and creators, get links live and track which audiences responded, then ran top performers’ content as Meta Partnership Ads, whitelisted through the creator’s handle.
• They report CTR up 40%, CPC down 20%, CVR up 25%, and CPA falling from €41 to €26.

For brands planning a campaign

Start affiliate and creator recruitment at least six weeks before a peak period, collect performance data, and whitelist top creators' content into Meta partnership ads. Starting late leaves no runway to earn trust.

They call that fake diversity and tie it to spend plateaus.

• Real diversity, in their framing, means sending distinct signals: a different persona or job-to-be-done, a different core motivation (desire versus fear, speed versus certainty, convenience versus status), a different awareness or objection state, or a different primary promise.
• They acknowledge visual and structural differences can affect how Meta classifies ads under separate entity IDs, but say classification is not the objective — incremental reach is.

Four planning dimensions: who the ad addresses (persona or job-to-be-done), why they would act (core motivation), where they sit (awareness and objection state), and what is promised. Visual and structural differentiation is treated as a distribution mechanic, separate from demand differentiation.

For brands planning a campaign

When briefing this creator or an agency working this way, specify which dimension each concept changes.

• A request for "ten new hooks" gives the system variation without new demand.
• Ask for concepts separated by persona, motivation, awareness level, or promise—the dimensions this work is built to address.

Examples: Source 1 · Source 2

They split the work into demand generation — cold ads that name a problem the viewer has not fully articulated, with no product pitch — and demand capture — ads that convert existing intent and therefore show low CPA and strong ROAS.

• In their observation, accounts over-index on capture and call it strategy, then cut generation when performance dips; capture numbers look strong for a few weeks and then stop.
• They also flag that unaware and problem-aware messaging is underused relative to product-aware ads, even though those larger pools are where the most scalable growth sits.

For brands planning a campaign

A media plan for acquiring new customers rather than cheaper conversions should fund both jobs and identify which concepts serve each.

• To reach new customers, a campaign needs problem-naming concepts aimed at cold audiences, not only social proof, objection handling, and product-aware angles.
• Ask a creative partner what share of output serves each job.

Examples: Source 1 · Source 2

Their second point qualifies the first.

• If the ads are different hooks on the same idea, Meta groups them under the same entity ID, so volume alone does not produce new demand.
• They pair volume with changes in messaging and awareness level.

Volume plus diversity, reviewed as one decision rather than two: more ads are useful only alongside a check that the underlying concept is not repeated.

For brands planning a campaign

Agencies and in-house teams that promise velocity should be asked what changes between concepts, not only how many launch. For a brand evaluating the pitch, the useful questions are how many genuinely distinct ideas launch each month and how the team verifies that.

Examples: Source 1 · Source 2

The Foundation learns the client's numbers, category, and customers, down to why each person buys.

• The Engine is a data-driven testing system where every ad launches with a clear hypothesis, each result feeds the next round of creative, and insights carry forward.
• The Scale step extends winners across creative formats and micro-personas inside an account structure built for Meta's Andromeda algorithm.
• The agency positions itself as end-to-end — ads management, creative strategy, filming, launching, and reporting — with the brand sending product, approving concepts, and keeping inventory current.
• Free growth audits are offered to brands that meet the agency's requirements, by application.

A hypothesis attached to every ad launch, with results feeding the next creative round rather than single-winner selection.

For brands planning a campaign

This is a framework a brand can test directly.

• Ask what hypothesis was attached to a past ad, what the result changed, and how micro-persona scaling differs from simply duplicating a winner.
• The agency describes a light client-side workload, suited to brands able to approve concepts quickly and keep stock and reporting data flowing.

Examples: Source 1 · Source 2 · Source 3

Previously, around 70% of concepts for a new client were UGC-led — new creator, new brief, new shoot — and roughly five weeks passed before the first new ad was live, during which nobody learned anything.

• The agency still does UGC but now balances speed of learning against new footage: instead of starting from a blank page, the team first iterates on creative that is already proven, either the client's own working material or a close analog from a similar account.
• They frame it as the same principle with different execution and report seeing real signal in days instead of weeks.

Start new engagements with proven or analogous creative, then layer in new production once there is signal.

A documented change of approach over time, not a fixed rule: it applies at the start of a client relationship and to slow-starting accounts, rather than replacing original production.

For brands planning a campaign

For brands with fixed windows, such as a seasonal peak with a few weeks of runway, iteration-first is the faster route to readable data. Ask which existing assets will be reused or adapted, and reserve new shoots for concepts that earn them.

Examples: Source 1

They say this matters most when every DTC brand is competing for the same attention.

• They report that creator-led, native, authentic partnership ads carry a disproportionate share of spend in many accounts because they cut through consumer scepticism toward polished scripts and AI-written angles.
• Their example is a fashion brand six weeks out from Black Friday with nothing in place: the team used UpPromote to recruit affiliates and creators, got links live, tracked which audiences responded, then put the top performers' content into Meta partnership ads whitelisted through their handles.
• They report CTR up 40%, CPC down 20%, CVR up 25%, and CPA from €41 to €26.

Affiliate or creator recruitment first, performance tracking second, then whitelisting the best-performing content into Meta partnership ads for cold traffic.

For brands planning a campaign

Recruiting and testing creators takes runway, so the channel has to be built before peak season rather than during it.

• A creator-recruitment plan should include who recruits creators, how response is tracked, and which handles get whitelisted into paid.
• Ask for the underlying client-case data supporting the reported lift figures.

Examples: Source 1 · Source 2

They set this against a market where consumers are more sceptical and AI-produced scripts and similar angles are everywhere.

For brands planning a campaign

The campaign direction should say what story an ad is telling and to whom, rather than only specifying production quality, casting, or format. Over-specifying polish works against the quality this creator says performs.

Examples: Source 1

They earn roughly €4,000–6,000 a month for output that, Steede says, requires little real thinking.

• Designers and editors typically earn less despite contributing across the production process.
• To test this, the agency put a senior designer through Stefan Georgi's RMBC2 course and trained them to assess where an account is underleveraged, verify that with the growth strategist, then write the hypothesis and copy and design the full ad.
• Steede advises buyers to stress-test whether they are paying for creative strategy or another AI subscription.

For brands planning a campaign

Brands hiring creative talent should ask who actually produces the work and what that person can do beyond scriptwriting: develop hypotheses, write copy, design, and spot where an account is underleveraged. That distinguishes the staffing model Steede endorses from prompt-and-paste output.

Examples: Source 1

One post says the agency has worked with 100+ seven- and eight-figure ecommerce brands over five years and offers free growth audits.

View example

Ask whether Steede has a commercial relationship with UpPromote.

View example

Ask whether Steede has a commercial relationship with Stefan Georgi.

View example

Both treat paid social as a system for finding what converts rather than a single campaign, and both favour creator content over brand-produced polish.

Schroën's posts cover an editor hiring call and a piece on how creator and affiliate content feeds AI product recommendations.

• They focus on building the team and content supply, framing the argument around hiring and lead generation.
• Steede's posts are more consistently explanatory: planning dimensions for concept diversity, the split between demand generation and capture, and how to sequence a testing system.
• Readers looking for the reasoning behind the ads get more from Steede; those interested in production and team-building get more from Schroën.

Read original · Original 2

His audit example walks through an account where most creative targets existing customers and warm, product-aware audiences. He argues for creative spanning unaware, problem-aware, solution-aware and product-aware levels—the same awareness-level logic Steede applies when splitting demand generation from demand capture.

Wasim writes for individuals building a practice as much as for brands: one post is a client-acquisition and positioning guide for new marketers, and another is a first-person account of winning a creative-strategy engagement through an audit.

• Steede's posts operate at the agency level, covering testing systems, onboarding changes, staffing and a named framework.
• They are aimed at brands and teams deciding how to structure creative work, rather than at people learning to sell it.

Read original · Original 2

The collagen example builds an ad from research into the account, product, competitors and personas, then reaches for a mechanism strong enough to convert a stranger — recognisably the same cold-audience problem Steede addresses with demand-generation creative and partnership ads.

Yanko lays out a specific copy architecture: establish the problem's unique mechanism, agree with the buyer's scepticism, then explain the solution's mechanism, deciding on format last.

• He reports tracking net new customer acquisition cost rather than blended ROAS and also publishes on using AI skill files to run DTC creative work.
• Steede focuses on the decision and testing layer: which dimension a concept changes, how volume becomes signals, and how the agency's system works.
• Yanko offers more detail on copy mechanics.

Read original · Original 2

In a collagen example, Nativ describes researching the ad account, ingredients, competitors, and personas, then building a mini-VSL around a customer’s skepticism, a distinct solution mechanism, and relevant proof.

• The post also says to assess new-customer acquisition rather than rely on blended ROAS.
• That gives a brand a concrete route from audience insight to an ad concept.

This is a specific example, not an end-to-end service covering media buying, production, and reporting.

• It offers less evidence than Levi’s posts about an ongoing system for iterating existing winners, balancing demand generation and capture, and scaling concepts across formats and micro-personas.
• Ask for the data behind the reported performance claims; they are not independently verified here.

Examples: Source 1

Waleed describes auditing a hair-care brand’s ads, finding that only about a quarter targeted new customers, and reframing the problem as a missing creative funnel across awareness levels. The work can help a brand check whether its creative speaks to people beyond those already familiar with the product.

The reviewed example focuses on diagnosis and creative direction, not the full-service combination of filming, launch, reporting, and ongoing iteration described in Levi’s offer.

• Waleed’s post says the audit led to a client engagement but does not spell out its scope or results.
• Ask what the engagement included and what results it produced; the account finding is an example, not a benchmark for other brands.

Examples: Source 1

Jim’s post describes creating UGC ads and proposes that creator and affiliate content may be cited by AI tools when they recommend products. This is more relevant if the campaign centers on creator content and affiliate reach than on expanding an existing Meta account’s creative testing system.

Levi’s posts do not show a comparable demand-generation framework, a process for testing distinct customer motivations and awareness levels, or results for the same paid-social scaling task.

• The AI-discovery argument is a claim in a promotional post, not independently established.
• Levi is the closer match when the brand wants a system for diagnosing and expanding paid-social demand.

Examples: Source 1