Jake Abrams

I help 7-9 figure DTC brands with revenue & CAC | I write The Process, a newsletter about modern growth marketing | Sold 100K+ personal cooling devices

Consumer Marketing & Growth Leader.  

Brands I've worked with:

Caraway
Jones Road Beauty
Skylight
Canopy
Insurify
Naked Wines
Rugiet
Sundays For Dogs
Momofuku
Sunday Lawn Care

100+ others

Book a discovery call in the featured section below 👇

Founder

The lean DTC growth marketing partner that I wish existed when I was an in-house marketing lead. 

At Odyssey, we help 7 & 8-figure DTC brands increase revenue & ROAS by finding the right combination of:

- Messaging
- Audience targeting
- Media channel allocation
- Creative

https://www.growodyssey.com/

Comfrt Clothing, Frost Buddy, and Underbrush are all using the same strategy to consistently go viral on TikTok while building massive creative pipeline.

Some marketers call it "nichemaxxing." I've created a 19-page playbook to help you deploy the exact same strategy for your brand.

Here's what's inside the Brand Nichemaxxing playbook:

- 4-tier content matrix from AI to hero content
- PDA framework for building micropersonas
- Full tech stack: Higgsfield, Genviral, Octo, Snowball, Discord, etc...
- Creator outreach templates that actually get replies
- Cost structures for every creator tier
- Winner selection framework for scaling organic to paid

The idea is simple: own the interest graph first, then amplify with paid.

Most brands do it backwards. They spend on ads before they understand what resonates.

This playbook walks you through 6 phases: niche definition, account setup, content architecture, creator recruitment, production rhythm, and measurement.

Get the 19-page playbook for free. No email gate. Just take it and use it.

1. Follow me
2. Like this post
3. Comment "playbook" and I'll send you the link.

Nichemaxxing across paid & organic social for 8-9 figure consumer brands | Writing about AI & advertising | Sold 100k+ personal cooling devices

This AI agent analyzed 12,500+ customer reviews and support tickets for an 8-figure DTC brand, created its own documentation for overcoming objections across hundreds of SKUS, then just started doing it...directly on the PDP.

Using AI like this can mean the difference between a 2.8% CVR and a 1.7% CVR. 

Essentially an agent that lives on your site, is trained on your catalog and every meaningful customer touchpoint, that can answer every question in 20 seconds or less.

These agents can handle conversion killers like:

Product fit: "Will this luggage fit in Delta's overhead bin?"
Comparisons: "Hard shell vs soft-shell for business travel?"
Technical specs: "Does it have 360° spinning wheels?"

I'm giving away the an 8-minute Loom video showing use cases for this type of agent, plus a guide to setting this up on your Shopify site in 15 minutes.

To get the video and guide:

1. Connect with & follow me
2. Comment "agent" below
3. I'll DM you the link

#DialogPartner

undefined undefined

The smartest people at Meta will tell you to build ads around motivators, not just micropersonas. 

I like this shift to motivators, because it forces you to start the creative process from a better spot.

Personas make you think about demographics. You end up with a bulleted list about a fictional human named Sarah who likes yoga and shops at Whole Foods. Nothing about that tells you what makes Sarah pull our her credit card and spend $150. 

Motivators do, though. 

When I do creative strategy now, I start by mapping out the actual motivators behind someone would buy or be interested in a brand. 

So for Sarah, maybe it's "I need an outfit that works for the December holiday party."

Then for each motivator, you map one core benefit that delivers the solution. That way, you build the creative around that feeling, not around "Sarah."

Then I do it for another motivator. And another....

By the time you've worked through 4-5 of them, you've got a full portfolio of creative assets, each built around a different reason to buy. Different themes, different visuals, different emotional angles. All for the same product.

That's the part people miss, I think. You're not trying to reach different types of people. You're trying to speak to different reasons to take action. A single product can have five completely different ads that all make total sense.

I feel like that's how you scale creative without running out of ideas or making yourself crazy with format multiplication.

Turn 35 today. Sheesh.

Here are 35 takeaways I have from the last 6 years as a "non-business" person building a business with 30+ people:

1. Expect daily disappointment
2. It's ok to be demanding if that's the clear standard
3. The ideas in your head are smart
4. Respect the audience's attention
5. Show undeniable proof
6. Work with challengers
7. Build systems for 50x output
8. Stay on the forefront of technology
9. Extreme bias for action
10. Creative generalists win now
11. Get over fear of speaking to your phone
12. AI + bad taste = epic fail
13. Rip structures, use original thoughts
14. Incentives drive everything
15. Pretty much everyone is winging it
16. Overcome initial content creation friction
17. Grow your own accounts first
18. Authenticity on Zoom is a superpower
19. Daily follow-ups with blockers, Mr Beast style
20. Work should feel like video game
21. Non-business people build differently
22. Ideas die without proper execution
23. Attention is the most sacred currency
24. Charge more than feels comfortable
25. Comfortable teams ship mediocre work
26. Systems unlock exponential output
27. Tech resistors always lose
28. Speed + direction = unfair advantage
29. Generalists are more important than specialists
30. Loom videos beat meetings
31. Tools don't fix shitty taste
32. You can only get so far ripping other's ideas
33. It's ok to start ugly
34. Practice what you preach, or else you're a fraud
35. Build leverage before you need it

---

Feel free to follow along as I document the path to building a $10M+ agency in 2026 by helping 50+ brands manifest the next era of paid social.

https://lnkd.in/epcpxd4V

I built a 250-video visual script library application in a day with Replit, Airtable, and Gemini. 

So I recorded an 8-minute Loom video outlining the entire process. 

I'm not an engineer. I did run marketing for a coding bootcamp, and maybe that experience helped me overcome the fear of "being technical."

Here's what I cover in the Loom video:

- Why my Airtable database was useless until I made this
- How I used Replit to put a visual frontend over raw data
- The hover-to-play feature that made this actually usable 
- How it auto-analyzes any TikTok or Dropbox link in ~60 seconds
- The prompt I use to extract hooks, transcripts, and full scripts
- How we use it to rewrite winning ads for new brands
- The N8N nightmare I replaced with this

I've been in marketing since 2013 and run a creative agency working with 50+ brands. This is the internal tool I wish I had years ago. 

Want the full 8-minute Loom walkthrough?

Follow or connect with me
Like this post
Comment "Build" 

And I'll send it over.

---

Repost for priority access.

I know an AI agent that sat on a DTC brand's PDP for 2 weeks, captured every question from shoppers, then showed the brand exactly why visitors weren't converting.

Using AI like this can mean the difference between a 2.8% CVR and a 1.7% CVR.

Essentially, this is an agent that lives on your product pages, answers questions in real time, and logs every conversation as a data point. After 2 weeks, you have a clear picture of what's blocking purchases within an easy-to-understand dashboard. 

These agents can handle conversion killers like:

- Compatibility: "Is this compatible with my device?"
- Confusion: "What's the difference between X and Y?"
- Fit: "Will this actually work for me?"

85% of this feedback normally dies in chat logs. But agents like this can turn that data into an actionable CRO roadmap.

I'm giving away a step-by-step Notion document that explains exactly how to set up this system on your PDP’s… 

So you have AI handling conversations, running an analysis on why people didn't buy, and delivering that to you in a dashboard with visualizations.

Works for any Shopify brand doing $1M+. Free to install, takes ~15 minutes.

To get the Notion doc:

1. Connect with and follow me
2. Comment "insights" below
3. I'll DM you the link

#DialogPartner

One of our creators just hit 50k views on a single TikTok.

She has 6,500 followers. She's paid ~$500/month for ~25 posted TikToks plus bonuses based on views her content drives.

This creator is an actual customer. She had a real transformation with this product. The proof is undeniable in the video she posted, and that's why it worked.

You can't script something like this. You can't hire a generic UGC creator to fake it. You can't recreate it in Higgsfield.

She's in our Discord community, where we're sharing tips for going viral and building a community around helping people earn by sharing their authentic stories.

It's all real. No generic content or fake testimonials.

This is the nichemaxxing strategy for 2026.

Massive communities of real people within a specific niche who you coach to go viral on TikTok. Everybody wins.

You have to be doing this as a brand in 2026.

The creative bottleneck doesn’t exist in 2026.

Learn:

Claude
Octo
Wispr Flow
Snowball

Combine with:

Exceptional creative taste
Operational psycho-ness
Next-level communication skills

Might sound harsh, but you gotta figure it out or you'll get left behind.

It takes you way too long to write scripts, and it's not your fault. 

It's because you keep having to re-feed the context to your creative strategy copilot every single time, and that's extremely annoying to do. 

Instead, run use this 3-part architecture that I custom built out for my agency:

First, work backwards from a format you already know worked. Have Gemini analyze the entire video. 

Second, you pull that context into Claude, which already has project files about everything you need to know about the brand you're writing for.

Third, your Claude needs a set of overarching skills that sit within your Claude organization that tell it how to write like your agency or you. 

When you do this, the output is actually good. It's not a shitty first draft. 

Because Claude already has the direction...what made the original format work, how to write like the brand, and how to be good at creative strategy. 

That's the architecture.

I recorded a quick video walking through exactly how I set this up. Watch it if you're tired of re-explaining your Claude how to write like your brand or write like a real creative strategist.

Here's the embarrassing truth that a lot of "AI is replacing everything" bros haven't talked about honestly...

Making good AI video content at scale is actually kinda hard.

Getting prompts right is a skill. It takes time to develop. You have 8-12 seconds per scene. Iteration is slow. A lot of the output you get early on is going to be unusable and you're going to waste a lot of time. 

Compare that to certain types of content that are just genuinely faster to shoot on your phone.

For example: I've been making 7 talking head "yap" videos per week for my own TikTok that I've grown to 27k+ followers. I just pick up my phone and speak for 60-90 seconds. And then it's done.

Why would I spend the time to clone myself and try to recreate something with AI tools that will always feel slightly off, when it doesn't save me any time or cost? Plus people will hate me for AI-ing myself. 

How about if you're Comfrt Clothing, and your top performing TikTok affiliate videos are just Gen Z-ers in their bedrooms taking selfie videos? They just show off their hoodie, show off a few features, and add a caption. That's a 15-minute shoot on an iPhone.

There is no AI workflow that beats that for speed, cost, and authenticity.

So before you burn 20 hours crafting an AI workflow for videos that are easier to just shoot with your phone... ask yourself if it's actually worth it.

A Matthew McConaughey & Timothée Chalamet interview recently went viral, in which McConaughey said that the moral plea against AI is not strong enough and will lose out to the economic incentives.

No lies told there. 

But here's what might actually slow it down:

Everybody hates "AI passed off as real" content.

No major brand I know would unleash an army of AI influencers right now who are pretending to be real. That nuance is everything. The brand risk is enormous.

Maybe that changes in 10-15 years when Gen Z and Gen Alpha are the primary consumers and the cultural relationship with AI-generated content has shifted.

Impossible to actually know.

But in the next year? 

AI takes over certain categories, but humans retain control of most categories.

Because authenticity and emotional connection is still what builds brands. 

Faking it has massive consequences.

DTC brands, I would STOP making $300 UGC videos and immediately firing them into your ad account.

This used to be a logical strategy. But in 2026, you're missing the bigger opportunity.

I've worked with 100+ consumer brands, and so many are spending 100% of their time "creating content for paid", and 0% of their time deploying it organically first.

If your ad account is full of unspent creatives, do this instead:

- Scrape your customer reviews for content ideas using Motion (Creative Analytics) or Claude
- Scrape organic TikTok for winning angles using Snowball
- Talk to your ad account data using Octo
- Identify good creators with Meta creator marketplace
- Set up a Discord to manage and motivate them

Then:
- Get creators on performance deals where they post to their own accounts
- Build your own network of slideshow & faceless accounts
- Deploy every creative test organically first
- Prioritize putting paid spend behind content that earned organic views

Test this for 30-60 days.

You'll see massive improvements in creative volume & winning creatives while reducing wasted ad testing spend. Plus you'll build a massive halo that lifts your entire business. 

Let that drive which content deserves your paid budget.

You're stuck at $250k/month in Meta spend because you don't understand exactly what's happening inside the system you're feeding money into. 

Let me break down the 4-model architecture running your ads:

ANDROMEDA
This is the retrieval engine that scans tens of millions of ads to find candidates for each impression.

GEM
This is the ranking superbrain that takes Andromeda's pool and decides which ad wins the impression.

LATTICE
This is the consolidation layer that connects multiple ranking models and shares learnings across Feed, Reels, and objectives.

SEQUENCE LEARNING
This helps model user behavior over time. Not just one impression, the full sequence of actions before and after.

How they work together:
→ Andromeda retrieves 
→ GEM + Lattice rank 
→ Sequence learning personalizes.

What this means if you spend $250k+ / month as an 8-9 figure brands:

- Creative diversity is how you deliver ad candidates
- More variants = more signals = better matching
- The AI handles distribution, you handle inputs

You need operational willingness to hit quality + quantity + differentiation b/c that's what the entire system runs on.

I built my first real vibe coding app with Airtable, Replit, and Gemini. 

Process was:
- I built out this big Airtable databse
- Gemini analyzes any video and put the analysis in Airtable
- I created a front end for this using Replit

You can hover over the videos, and they auto play. You can open them up, and get the script. .You can even copy a prompt that you can preload into Claude
to do a rewrite of that video for your specific brand.

The prompt auto checks the project files in Claude that tell it how to write like that brand, and then it does a rewrite and gives you all this extra helpful context.

This was a massive game changer for my process of:

- Working backwards from a format
- Adapting it for the brands that I'm working with
- Applying my creative taste to edit the content in a human way afterward

"Content creators are toast, AI is going to replace all of you."

Here's why I believe this take is fundamentally wrong...

I run an AI-native creative agency. I use AI every single day. And I genuinely believe AI is one of the most important shifts happening in our industry.

AI will infiltrate certain categories and never touch others.

Any animation that is not Pixar-level, top 5%...is probably gone. Why pay $5,000 when you can prompt and tweak something that is relatively comparable?

Same with word walls, cinematic story ads, static creative, memes, and product renders.

But then there's content where the entire f*****g point is that it's real.

Pull up Comfrt Clothing on Kalodata. They're doing $10M+ per month on TikTok Shop. Look at their top 10 revenue-driving videos. None of them are AI. Almost all are real micro and mid-level influencers.

Because the content that works on TikTok is "I am a real person who bought this thing and here's what I think."

The second that's AI, the whole thing falls apart. 

At the end of the day, authenticity has always been the driver of building the best brands.

That didn't change just because it got cheaper to fake it.

Brands that have hired ex-Odyssey team members:

Gruns
Hexclad
AG1
Trip

I don't believe there is a better place to learn modern creative strategy, as long as you're hellbent on joining the top 1% of humans at this skill. 

We're hiring again for a critical position to help us deliver on our goal to elevate 8 & 9-figure consumer brands into the top 1% of their category for creative diversification & distribution. 

Apply here: https://lnkd.in/eYex6PPW

To apply: send a Loom video showing proof of two things:

1. A TikTok Shop you've managed...walk us through the results & what you did personally to grow it.

2. A Discord creator community you've managed...show us the server, your activity, and how you kept it engaged and growing.

"Do you get brands from posting on TikTok?"

I don't even know what to say back, cause I don't think of things that way. 

All I think about is:

- The thoughts in your head about marketing are actually smart
- If you pair them with the right hook, people will pay attention
- Just keep putting that out there, into the world
- They will find the right people
- When they're ready, they will just reach out
- I have a better chance than someone they never saw in their feed

That's literally it.

Let's say you need to build 10+ affiliate programs from scratch.

The goal: onboard 30+ micro-influencers per brand within 15 days.

Here's the setup:

Target: creators with 5k-50k followers already posting in each brand's niche (not random people)

Half the brands sell on TikTok Shop. Half don't due to business model constraints.

Compensation structures vary by creator:
• Commission-only + free product
• Retainers for high volume (30 posts/month) + view bonuses
• Retainers + sales commissions

Requirements:
• Full content buyout for ad rights across all platforms
• Dead simple process for sparking and partnership ads
• Ability to manage multiple programs simultaneously

Nice to haves:
• Payment management
• Dynamic creator payments based on CPM model

What's the 10/10 tech stack for this for < $2,500 / month?

Is it:
• Just Upfluence
• Just Aspire
• Just SARAL - The Influencer OS
• Refunnel + a CRM like Streak 
• Social Snowball ☃️ + a CRM like Streak 
• Superfiliate + a CRM like Streak 
• Some other combination
• Something else entirely

You need operational efficiency here. 

What have you used at true scale that actually works?

Upfluence

Aspire

SARAL - The Influencer OS

Refunnel

Streak

Social Snowball ☃️

Superfiliate

Jack Dorsey fired 40% of Block staff "because of AI."

Here's exactly what to do next if you're afraid of being next...

18 random thoughts from growing my TikTok from zero to 26k+ followers in a few months.  

- I kinda think you're a fraud if you make content for a living but are unable to grow your own social accounts
- Every thought-out LinkedIn post is a script for TikTok
- Unscripted TikToks become more LinkedIn Content
- Just fire off that "hot take"

- Audio quality matters more than camera quality
- Just buy the RØDE mic, it's awesome
- Making your own videos is best way to become a better CS
- Every CS should try to grow a TikTok, even it's a faceless account

- The hook is everything
- The hook should be oddly specific: use keywords and numbers
- TikTok "categorizes" you, and it becomes easiest to get views in that category
- The random thought in your head is good enough for TikTok, as long as you have a specific hook

- It's easier to get views on TikTok than LI w/o resorting to "lead magnets"
- It's possible to work out and make 3 TikToks during my kid's naptime
- People will watch and comment even if it's just a talking head video
- Your audience is on TikTok b/c everybody is on TikTok

- TikTok drives real business opportunities
- I had 3 leads comes through last week who discovered us on TikTok

Anyway, those are my thoughts.

taplio