Last week our team spent a day at Chief Marketer SF, listening to operators from brands like ManyChat, Ridge and FIXD, plus the agencies and tool makers who sit behind some of the fastest growing ad accounts on Meta. We went in expecting the usual mix of channel tactics, platform updates and attribution debates.
That is not what we got. Session after session landed on the same conclusion, from different speakers with different businesses and different data. Media buying is no longer where brands win. Creative is.
For an agency that has spent years building its reputation on performance media, that is not a comfortable message. It is also the right one. This post covers what we heard, what we expect to see across DTC marketing in 2027, and exactly how Pennock is changing because of it.
Part one: what we heard
Most ads fail, and that is the system working
The most cited number of the day came from Motion's 2026 Creative Benchmarks report. Motion analyzed 578,750 Meta creatives across 6,015 advertiser accounts and $1.29B in spend, launched between September 2025 and January 2026. Only about 5% to 8% of those ads became winners. Motion defines a winner as an ad that spends at least 10x the account's median ad and at least $500.
Read that again. More than nine out of ten ads that a brand launches will not become a meaningful driver of spend. Roughly half get turned off before they reach 28 days. And the small group of winners captures about 55% of all spend in the dataset.
This changes how a brand should think about creative. The goal of any single ad is not to be good. The goal of the creative program is to find the few ads that are great, as cheaply and quickly as possible. Nobody in the room claimed they could reliably predict which ad would win before launch, including teams spending millions a month. The only dependable path to more winners is more qualified attempts.
Volume separates the top quartile
The benchmark data broke this down by monthly Meta spend, comparing all accounts with the top 25% of accounts in each tier. Motion reports volume as new creatives launched per week.
| Monthly Meta spend | Weekly creatives, all | Weekly creatives, top 25% | Winners per month, all | Winners per month, top 25% |
|---|---|---|---|---|
| Under $10K | 2.80 | 4.83 | 0.00 | 0.00 |
| $10K to $50K | 4.10 | 8.09 | 0.25 | 0.50 |
| $50K to $200K | 6.67 | 15.95 | 0.75 | 2.00 |
| $200K to $1M | 11.24 | 31.11 | 1.75 | 5.99 |
| $1M and above | 18.85 | 54.64 | 3.99 | 10.48 |
The pattern holds at every tier. Top quartile accounts launch roughly 2x to 3x as many creatives and find roughly 2x to 3.5x as many winners. Look at the $50K to $200K tier, where many growing beauty brands sit (see our Meta advertising cost guide for beauty brands). The average account launches about 7 new ads a week and finds fewer than one winner a month. The top quartile launches about 16 a week and finds 2 winners a month. Same budget band. More than double the hits.
Motion is careful to call this an association, not proof of cause. Bigger, better resourced teams can test more for many reasons. But no speaker disputed the direction, and the brands on stage made it concrete. One slide showed live Meta ad counts for well known DTC and software brands: Cursor above 1,000, Wispr Flow above 800, Victor above 600, Base44 above 500, Ridge above 470, AG1 above 400, Caraway above 200 and True Classic above 190. Ridge described running 1,400 creators and 3,200 live ads on Meta at once.
Another speaker shared a smaller example that makes the same point. Of 200 ads run to drive event applications, the top 10% drove 80% of applicants. Winners are rare, and they carry the account.
Creative is the targeting
The second theme of the day explains why volume matters so much. On Meta today, the ad itself is the targeting.
For years, performance teams won by building audiences: interest stacks, lookalikes, exclusions and careful account structure. Platform automation has absorbed most of that work. Meta's delivery system now reads the creative to decide who should see it. Several speakers referenced the Andromeda era of Meta delivery, where the ad's content does the job that audience settings used to do.
One session named the two highest signal elements for getting Meta to serve an ad to the right person. First, the opening 3 seconds, which both qualify and disqualify the viewer. Second, who is on screen, meaning their age, setting and life stage. If a 50 year old woman in a bathroom talks about texture changes in her skin, the system learns who that ad is for. If a 24 year old talks about breakouts in a dorm room, it learns something different.
That is why hooks dominated the conversation. One speaker's list of creative fundamentals opened with a blunt claim: hooks are 80% of the game. The rest of the list followed from it. Creative is the targeting. Become a professional doomscroller. The algorithm rewards variety. Shoot more shots.
Persona research is the most important step
If creative is the targeting, then knowing exactly who you are making each ad for becomes the most important strategic work a brand does. Multiple speakers shared research processes, and they were remarkably consistent.
One team walked through a four step creative strategy research process. Step one is a reputation analysis of the brand and its category. Step two is customer review mining across the brand's own site, Amazon and retail partners. Step three is persona and desire segmentation, which the speaker called the most important step after Andromeda. Step four is a mission document that turns the research into direction a creative team can act on.
The persona work stood out for its specificity. FIXD, a car diagnostics brand, showed four personas pulled from reviews, each with a one line core story. "At the Mechanic's Mercy" people walk into a shop already braced to get scammed, and have been more than once. The "Dashboard Light Panic Club" sees a warning light and has no idea whether the fix costs $50 or $5,000. "One Repair Away From Broke" drivers treat every unexpected bill as a small financial crisis. "Dealer Lot Gut Check" buyers need proof before they sign.
None of those personas describes a demographic. Each describes a moment and a fear. That is what makes them useful for creative. A brief built for the Dashboard Light Panic Club writes its own hook.
Speakers also pushed research beyond reviews. Become a professional doomscroller. Build separate social accounts that behave like your ideal customer, so the feed shows you what they see. Scroll YouTube and TikTok to find the formats, hooks and creator styles your buyer already responds to. Research is no longer a quarterly deck. It is a daily habit.
A shared language for ideas
Teams that ship creative at volume need a common way to describe ideas, so strategists, creators and editors can move fast without long meetings. One framework from the day was simple enough to adopt immediately.
Every ad combines a message, a concept and a hook. Messages fall into three types: core, story and trend. Concepts fall into four: transformation, credible explainer, testimonial and tutorial. Hooks fall into four: visual, pope in the pool (an unexpected image that stops the scroll), familiar face and callout.
On top of that sits format. Speakers listed formats that are working now: green screen, 1 shot, 10 shot, carousel, yap, vlog, podcast, hyperedit, aura edit and fully produced. Multiplying messages by concepts by hooks by formats gives a team hundreds of distinct, nameable ideas from a single product. That is how creative diversity happens on purpose instead of by accident.
One speaker framed this as a pyramid. Hook strategy sits at the top and draws on customer research and emotional motivators. Concepts sit in the middle and drive planning and resourcing. Individual ads, with their formats and messaging, form the wide base.
AI is in every step, with humans signing off
Every team that presented its workflow showed AI inside it. The tools varied. The structure did not.
One agency mapped its creative workflow into four stages and named the AI tools behind each. Strategy uses performance analysis through Motion, a social listening agent, and Google and Slack connected through MCP. Ideation uses a custom copywriting agent and Figma Weave. Production uses ElevenLabs for voiceover and music, TwelveLabs for video indexing and Airtable for tracking. Post production uses DataClay with After Effects automation for versioning. Claude, with custom skills, ran across all four stages.
The idea of skills came up again and again. A skill is a reusable set of instructions an AI system follows for a specific task, such as finding personas in reviews, writing hooks or scoring an ad against a brief. One strategist showed her own library: persona finder, review audit, hook writer, messaging angles, customer language, scriptwriter, voice guide and creative analysis. Another showed a video editing skill set covering A roll processing, B roll processing and selection, captions, framing, headlines and post previews. One desktop usage report showed AI prompts taking 62% of a strategist's working time.
The results were real. One team described AI creative development taking 2 hours instead of 2 weeks. Another showed polished testimonial style variations produced across two brands from a single locked template and spreadsheet.
What kept this from becoming a slop factory was governance. The strongest teams used one brief structure for both human creators and AI tools. Brand and visual guidelines were built into the skills so ads did not look like everyone else's. A human signed off at the brief and again at the output, whether a person or a model made it. And every fix fed back into the skills, so each round improved on the last. Speakers were honest that early AI output is often poor. The system, not the tool, is what makes it usable.
Creators sit at the center of the brand
The last major theme widened the lens beyond the ad account. Anna Tutckaia, VP of Growth Marketing at ManyChat, argued that brands should put creators in the middle of their universe. In her view, paid, creator and organic are no longer three channels. They are one.
Her team laid out the difference between old social and new social. Old social had three rings: owned, organic and paid. New social adds a fourth ring between owned and organic, called enabled. That is content made by employees, customers and creators the brand equips and encourages. It is often the fastest growing source of reach a brand has.
Recruiting for that ring happens three ways: internal people with incentives, customers who already show outward signal, and creators sourced through platforms like Trybe, Sideshift and Muse. Speakers suggested brands map their full cast of characters, including founders, employees, experts, creators, influencers, customers and celebrities, alongside the sets they can film in and the company, culture and industry events across each quarter.
One more slide stayed with us. It plotted channels on two axes, sales against awareness and normie against fringe. Familiar channels like Meta Ads, Google Ads and email sat on the normie side. Affiliate, live selling, SMS, AppLovin, seeding, multi account strategies, clipping, social shows and streaming sat on the fringe side. The point was not that fringe channels replace core ones. It was that the brands growing fastest are experimenting at the edges while running the basics well.
Reporting is solved. Insight is not.
A final point came up in passing but may matter most for agencies. Several speakers described the current moment as the real time analysis era. Social feedback, ad results, campaign results and customer analysis now flow into an LLM that answers questions on demand.
One operator put it plainly. In house teams are already savvy at media buying. Reporting is real time and largely done. What they need from partners is creative. That was a sharp message for any agency whose main deliverable is a weekly performance report.
Part two: what to expect across DTC marketing in 2027
Based on what we heard, and on what we see across the beauty and skincare accounts we manage, here is where we believe DTC marketing is heading next year.
Media optimization becomes table stakes
Bidding, budget allocation and audience building will keep moving into platform automation. Brands will still need sound account structure, clean conversion tracking and disciplined budget management. But those skills will stop commanding a premium on their own. Brands will expect them as the baseline and judge partners on what comes next.
The creative strategist becomes the most valuable marketer on the team
One speaker traced the star of each marketing era. Print had the copywriter, who wrote hooks. TV had the creative director, who made ads people wanted to watch. Digital had the performance marketer, who lived by the numbers. The feed era belongs to the creative strategist, who combines all three: customer insight, hooks, formats and testing discipline.
We expect brands to hire for this role aggressively in 2027 and to evaluate agencies on whether they can supply it.
Creative supply chains replace creative calendars
A monthly content calendar cannot support 16 new ads a week. The brands that grow in 2027 will run creative like a supply chain, with weekly output targets, defined inputs, quality checks and logistics more like inventory management than campaign planning. Speakers described this as a high velocity supply chain, and the comparison fits. Running out of fresh creative will hurt like running out of stock.
Research becomes continuous
Persona research will shift from an annual exercise to a constant one. Review mining, social listening and feed research will run every week, often with AI doing the first pass. Brands that understand which customer moment each ad serves will get more from their ad spend, because the platform will route each ad to the right buyer.
UGC stays the core, but the mix gets smarter
One speaker shared a creative mix with UGC at 65%, statics at 15%, and influencer partnerships, founder led ads, an offshore creative team and other formats at 5% each. We expect UGC to remain the backbone of DTC creative in 2027. What changes is how it gets made. Creator networks, customer content programs and AI assisted editing will all be needed to hit the volume the data calls for.
AI creative grows, and beauty needs guardrails
AI generated video, avatars and voiceover will become standard tools. For many categories they will produce a large share of test volume. Beauty and skincare need more care. An AI avatar describing results it never experienced is a synthetic testimonial, and that is a line we will not cross. We expect the best beauty brands to use AI for demos, explainers, editing and versioning while keeping customer results tied to real customers.
Volume also multiplies compliance work. Sixteen ads a week means sixteen sets of claims to check. Brands that build claims review into the creative process will move faster than brands that treat legal as a final gate.
Paid, creator and organic merge into one plan
Separate budgets and separate teams for paid social, influencer and organic will look dated by the end of 2027. The winning content from any one of them should feed the others. A creator video that performs organically becomes a paid ad. A winning paid hook becomes a creator brief. Brands that plan these together will compound faster.
The agency model splits
Agencies that sell account management alone will face pressure on price and relevance. Agencies that combine performance discipline with creative strategy and production capacity will become more valuable. We expect consolidation, new partnerships between media and creative shops, and a clear divide between the two models before the year is out.
Part three: how Pennock is responding
We have managed paid media for beauty and skincare brands for years, and we are proud of that work. We also heard the message at Chief Marketer SF clearly. Here is what we are changing.
We are launching the Pennock Creative Lab
The Creative Lab is a dedicated team inside Pennock that runs the full creative flywheel for client brands: research, ideation, briefing, production, QA, launch and reporting. It operates in 2 week sprints so every cycle produces new ads, new learnings and better inputs for the next round.
We are starting with a pilot of three beauty brands. We will learn in public with those clients, refine the system, and then extend it across our book.
Our targets are specific
Vague goals produce vague results. For each pilot brand, the Creative Lab targets top quartile creative volume for that brand's spend tier and twice the monthly winners the brand finds today. For a brand in the $50K to $200K tier, that means about 16 new creatives a week and 2 winners a month. We will measure creative volume and winners in Motion, using its winner definition, and measure revenue impact in Triple Whale.
We will also track the cost and hours behind each asset. The point is not volume for its own sake. It is the cost of finding each winner.
Our media buyers become creative strategists
The biggest change is to our own team. For years, much of an agency media buyer's day went to monitoring accounts. Platforms now do much of that work. So we are redesigning the role around what platforms cannot do: customer research, persona work, hook development, testing design and turning performance data into the next brief.
Our team already knows which ads drive results for beauty brands. The Creative Lab puts that knowledge to work earlier, at the start of the creative process instead of the end.
We are building on Claude skills
Pennock already runs a creative brief skill that turns monthly performance data into a creative matrix and brief deck. The Creative Lab extends it. We are adding a creative roadmap input, trend research and full brand guidance for every design element. We are building new skills for persona finding, review audits, hook writing and QA scoring.
Every brief follows one structure, whether a creator, an editor or an AI tool works from it. Every brief and every asset gets human sign off. Every correction feeds back into the skills.
Research becomes a daily habit
Our strategists will run ideal customer research accounts for each pilot brand, mine reviews continuously and track formats and hooks across the feed. Ideas, formats and winning concepts will live in one shared library so every brand benefits from what we learn on the others, without ever sharing one client's data with another.
Claims review lives inside QA
Beauty and skincare brands carry real regulatory responsibility. Our QA rubric scores every concept on three questions before production: is it on brand, is it direct response enough, and is the hook strong enough to win. Claims checks against each brand's approved language sit inside that rubric, not after it. AI is used for demos, explainers, editing and versioning. Customer results stay with real customers.
Paid, creator and organic in one plan
The Creative Lab will plan with creators at the center. We will help pilot brands map their characters, sets and moments for the year, recruit creators and customers into the content engine, and route winning content across paid and organic. Our affiliate team already works with creators every day. The Lab connects that work to the ad account.
The bottom line
Chief Marketer SF did not tell us anything we could not see in our own accounts. It made the pattern impossible to ignore. Platforms have absorbed much of the work media buyers used to own. The work they cannot absorb is understanding customers and turning that understanding into a steady stream of ads worth testing.
The brands that win in 2027 will not have the cleverest campaign structure. They will have the best system for finding their next winning ad, and the discipline to keep finding them every week. That is what we are building at Pennock.
Source: Motion, Creative Benchmarks 2026. Other figures as presented by speakers at Chief Marketer SF, November 2026.