Pinned Tweet
I'm looking for two eCom Klaviyo strategists to join my agency, remotely
You'll work directly with me to create the highest-converting emails for successful 7 and 8-figure brands.
Apply here:
emailengineers.notion.site/W…
love the schrödinger’s black friday sales:
the sale is both “ending in 3 hours”
and permanently running all November
as long as you don't refresh the page
my take on using AI for marketing materials (emails, ads, product shoots, etc) is simple:
A) either make it super realistic: 4 out of 5 people on your team shouldn't be able to tell it's AI
B) make it obviously AI: customers will forgive you if it's entertaining
the middle is the cringe zone, where you lose all trust
last week we’re taking new BFCM clients
if you want us to weaponize your Klaviyo list before Black Friday, DM me (Klaviyo says we squeeze up to 6.6x more revenue out of every subscriber than the average brand)
one rule:
deliverability score has to already be 50+
I’ve NEVER audited a spaghetti flow that converted well
12 branches
37 emails
claude has made it too easy to build very sophisticated ways of sending mediocre emails
would rather have a single-path flow with 10/10 emails
The Plan to Get 100 Brands Making More From Email Without Touching Klaviyo or Overpaying an Agency
I’ve spent the last six years running an email agency.
One thing I don’t like about the model: the less work we do for a retainer, the more money we keep.
It’s an incentive problem.
Sending more emails means more copy and more design, more human labor.
So we agree on a fixed number of emails, and anything extra costs more.
Meanwhile, most founders I talk to want three things from email:
- Get it done...
- Know it’s making money...
- Never open Klaviyo again, and spend that time on ads and product.
The agency model makes all three harder than they should be.
AI made me want to try a different model.
Could AI help brands send more (good) emails without paying an agency more for every one?
I’ve been working on that for the past few months.
Spent most of my free time building what I’m calling the AI Email Department.
The goal is to have 100 brands running their Klaviyo through it by the end of 2027.
It handles strategy, copy, design, self-optimization and reporting.
Brands manage it through Slack, without having to chase an agency for updates.
We have 7 brands in the beta and have generated about $100,000 for them so far (disgustingly low).
It’s early.
I want to get that number into the millions, but there’s plenty left to build and improve.
Here’s The Plan:
1. Focus on what we do well.
Klaviyo is building its own AI.
They have access to more data than I ever will, so I’m not trying to compete on that.
Our focus is direct response, emails that sell.
Long sales-page-like designed emails.
Text-based emails.
Storytelling emails.
Klaviyo has shareholders, lawyers and a public-company reputation to protect. Meaning they have every reason to play it safe.
We don’t.
We’ve worked with 50+ brands and generated over $52M through Klaviyo and developed our own strategies and tactics along the way.
I want every brand using AIED to benefit from what we’ve already figured out.
2. Give each brand a department with undivided attention.
When you hire an email agency, you share its team with every other brand on its roster.
Sometimes another 8 brands.
Sometimes 40.
Your copywriter owes five clients campaigns by Friday, so they’re not going to spend time researching your audience for the next copy batch.
You get the emails in your package. And you get improvements only when you ask for them.
Eventually, you realize you’re still managing email. You’re just paying someone else to make it.
I want every brand to get the attention an agency gives its biggest client, without having to be its biggest client.
3. Calibrate it for each brand.
Our senior team sets the initial voice, design and marketing direction, then reviews the first campaigns and corrects the AI’s decisions before it works on its own.
Even a tailor with all your measurements needs a first fitting.
4. Stop charging by the email.
Going from three campaigns a week to five shouldn’t mean renegotiating an agency retainer.
I want brands to be able to send daily when it makes sense, without buying a bigger package every time they want to send more.
5. Partner with ad agencies.
There are some seriously good Meta and Google ads killers on here.
I’d bet most have a few clients they’ve been telling to sort out their email for months.
I want to get to know those guys, get this working for a few of their clients and do a good enough job that they recommend it to the rest.
6. Partner with the ecom demons.
Simple really.
They’ve built an audience of founders who trust their judgment.
I want to get AIED in their hands. Have them try it, see the results, and recommend it if they think it’s good.
7. Charge for setup, then cover the running costs.
I’ll charge upfront for the setup and calibration, then a monthly fee to keep it running.
These emails are expensive to produce.
The monthly fee covers that, and it’s still 1/10th of what an agency normally charges.
8. Share the progress here.
I’ll post the numbers, what’s working and what we’re still fixing.
You can follow along as we go from 7 brands to 100.
Six years in email gave me a pretty clear list of shit I won’t send and clients I won’t take.
So a few rules:
1. No AI slop.
No “elevate your everyday.”
Don't need to explain.
2. No weekly discounts.
We’ll use promotions when there’s a reason to.
Not just to compensate for weak marketing.
3. Make Autopilot optional.
If one wants to review every email before it goes out, they get to approve it through Slack.
If they want it handled without waiting for approval they get to enable Autopilot.
They decide.
4. Work with killers
I’m building this with one guy in mind:
The Killer.
He scaled his brand to a few hundred grand a month and runs a lean team. Still makes the calls himself.
There’s no committee approving subject lines.
There’s him, with 20 other things to do.
He wants email done, wants to know it’s making money, and wants to get on with his day.
That’s the plan for the end of 2026 and 2027.
I’ll keep sharing updates as we build it.
If you’re the Killer, get three emails designed for your brand at AIEmailDepartment.com.
If you run an ads agency and have clients who need email sorted, DM me.
-Matias Perelli
I'm looking for two eCom Klaviyo strategists to join my agency, remotely
You'll work directly with me to create the highest-converting emails for successful 7 and 8-figure brands.
Apply here:
emailengineers.notion.site/W…
sometimes the algo drops me into saas twitter
guru with 400k followers celebrating $2k MRR on his new vibe-coded app
meanwhile ecom twitter has a 17-year-old doing $2k before lunch on a store he launched this morning
ecom marketers are on a different level
as a retention marketer, most "retention" activities are bullshit
so they bought your creatine gummies and now an email with "real people, real reviews 😁" is supposed to get them sold on your digestion formula pills?
???
fake and gay
it's a whole acquisition funnel all over again
unaware > problem > solution > product > most
yes you do have the benefit of trust, though
but man who trusts your boat still needs a reason to cross a river
just whipping out a spaghetti cross-sell flow in Klaviyo, expecting they're going to buy just because you sent a clever email is straight up arrogant
the best retention is acquisition
sad to see many 8-fig health brands make this mistake
what to do instead?
weaponize email
we create long direct response-like designed and text-based emails that walk them down through awareness stages
fancy?
not at all
works?
works
- Matt
"retention marketing" but it's just setting up a 7-split post-purchase flow with a 0.05% CR and blasting the "free credit" campaign every two weeks or so
creative strategists vs. creative tacticians
same happens in email marketing
swiping the "credit" email, the "us vs. them" angle, etc doesn't make someone an email marketer, but merely an email sender
If you’re looking for top CS talent, pay attention to swipers who call themselves creative strategists.
Swipers usually stalk all the most popular ad libraries
- Ryze
- Resilia
- Rosabella
- Koriderm
- Smooche
etc.
and they usually have impressive portfolios with good-looking ads but irl they have no clue what they are doing.
I recently interviewed a swiper who looked good on paper, with a strong portfolio
but when I asked him “How do you learn direct-response marketing, do you read any books and how do you study human behaviour”
he simply said
“I don’t read anything, I just learn by doing”
I wasn’t sure what to think about this because from my experience, in order to become a really good DR copywriter/creative strategist, you need to have studied DR copywriters, principles, frameworks, and what influences people to buy something.
So I gave him a trial and tbh this was the most terrible trial task I’ve seen.
The guy had no idea how to sell the product,
no clue how to influence the buying decision of the person he was targeting,
and had no idea what the core desire was that he needed to sell to.
He simply presented scripts that were 1:1 re-creation of the inspo ads he'd seen on the Ads Library.
Finding a good creative strategist is super hard and you need to make sure you have a proper hiring process before investing time & money into someone who will never succeed.
oct 29 update: launched 2 new ads
the first one is straightforward
to create the second one I followed one of my favorite principles from Rory Sutherland:
"the opposite of a good idea is often another good idea"
so I just uglified the first ad
which one you think will win?
respectfully, this is ass
With a single prompt covering the offer, brand style, and imagery, Kama used Composer built an on-brand email. A couple of quick follow-up requests to adjust the photos, and it was *ready-to-lunch.*
Try Composer: bit.ly/4sbk5DE



















