This website uses cookies

Read our Privacy policy and Terms of use for more information.

If your brand has 100,000 followers on Instagram and the average post reaches 10% of them organically, you already know the deal. The platform turned your community into a paid channel. Every brand I talk to feels it. Nobody's really doing much about it.

Meta published their Q1 2026 earnings last week. Ad impressions up 19% globally. Average price per ad up 12%. In Europe, where we sell, price per ad climbed 19%. Nineteen percent.

More ads, higher prices, same feeds. Meta is prrrrinting money. Small brands are funding the printer.

The brands that survive this are the ones building direct relationships they own. Email, word of mouth, community. The stuff that doesn't scale overnight but compounds over years.

erlich textil will be one of these brands. So we made a call. Three weeks, zero euros to Meta. The whole budget goes into product and community. 500 of our bestselling bras, given away to the women who help spread the word. To get one, you join the waitlist and bring friends along. The 500 women who bring the most friends get a bra for the price of shipping. Everyone else gets a thank-you discount.

The campaign is called 500 BRAS. 1 MISSION .

I posted about it on LinkedIn this week. That was the short version. This is the full one, lets get right into it.

The stack. Aka this used to take an agency a few years ago

We are a tiny team, so the campaign architecture had to be highly automated without loosing the human element of our brand. A tool called Viral Loops runs the referral backend. You join the waitlist, get a unique link. Every friend who signs up: 10 points. Instagram story with @erlich_textil tag: 20 points. Reel or feed post: 50. After three weeks, the top 500 claim a bra for 4.95€ shipping. Everyone else gets a tiered discount.

Viral Loops does one thing well (referral tracking) and the UX around everything else is pretty rough. We hacked around it constantly. The native Klaviyo integration works ok-ish, but getting custom properties to sync for real-time point counting took way too much troubleshooting.

For social verification we're on ManyChat, hooked into our Instagram business account. Same thing, fancy frontend, rough backend. But we made it work. Story mentions get picked up automatically via the Mentions API. Reels and feed posts: participant sends the permalink as a DM, the bot reads the URL pattern and assigns points. I think i built 12 flows myself with Claude Code writing the logic and me testing every single path. Every bot message was written by hand and checked against how we actually sound. If you get a DM from our bot after tagging us, that's my writing. Please be nice.

On the email side, Klaviyo runs six campaign flows with more conditional splits than I want to count. Existing customers get a different sequence than prospects. People with an Instagram handle get different reminders than people without one. Existing customers can't win a bra but earn a 20€ kickback voucher per friend, up to 5 of them, so max 100€. Getting our loyal customer base to refer us to their friends is the seed, the kickoff to the whole dynamic. So a super critical step. Thats why each of those variations ships with its own copy and logic. We built all of it because nothing off the shelf did what we needed it to do, all of the templated flows you can get are just not on brand or custom enough so it would pass the quality gates of my brain. A blessing and a curse.

The landing page was 80% built with Claude Code and Claude Design, then put through round after round of human review until it felt right. Initally wanted to do shopify and fully templated but that fell apart after a few design interrations. The fonts, the spacing, the scroll behavior, the signup flow. None of that came out of a template. Exported it into a liquid theme for a shopify import and off we go. Solid, fast, doesn’t break.

The part that took the longest wasn't the tech.

It was the messaging. It always is. We spent hours and hours on one question: why would anyone actually share this? What do they get out of telling their friends? And how do we bring three very different narratives into one campaign that doesn't try to say everything at once?

The three narratives:

  • product conviction ("IN BRAS WE TRUST")

  • the Meta critique ("KEIN BH FÜR MARK. ABER 500 FÜR EUCH.")

  • the social angle ("ZURÜCK ZU SOCIAL IN SOCIAL MEDIA").

We wanted the campaign to resonate broadly and know that our different audiences react to different things depending on the channel we meet them in. So each narrative is designed to work on its own.

We designed shareable statement cards for participants: "SORRY MARK. WIR VERSCHENKEN UNSER KOMPLETTES AD BUDGET." and "WIR SPIELEN DAS SPIEL NICHT MIT. LIEBER SCHENKEN WIR DIR EINEN BH." Three variants, all served via email after signup and the manychat bot in scanrious where we feel someone needa a little helping hand. Participants don't have to create content from scratch. Pick a card, post it on their socials, done. Makes it easy for people who usually never publish things, another critical step in the campaign journey. M a k e t h e m p o s t.

We have all of our organic channels lined up for this. email, organic social, PR, onsite, we activated friends of your brand. On the paid side, we’re shifting the paid social budget partly to influencers we trust and sho share the same values as us. SEA stays on, no affiliates. Wanted to do digital out of home but decided to save the money.

The team worked weekends on this. I went from strategy deck to writing bot messages to reviewing Shopify code in the same afternoon. Twisted my brain several times but it’s also addictive. We sat in calls debating whether "Sorry Mark" lands as cheeky or bitter, whether first-time visitors understand the mechanic in under 10 seconds, whether the tone was rebellious enough without being aggressive. That's what a campaign this size looks like with seven people. Everyone touches everything. There's no agency in between.

Why i’m still nervous as hell

Three weeks without paid social is a real gap. If the referral mechanic doesn't work properly, May gets tight and June starts from behind. We've never run anything like this. The projections are as good as we could make them, but in the end just projections.

But hey. This team just created a campaign we are all proud of in a few weeks time. People get it when we show them the creative. We figure things out, always have.

Projekt 500 Bras launches Tuesday, May 12.

Support us, would make our day.

I'll come back with the reuslts. Real numbers, what broke, what surprised us.

Until next week,

Philipp

P.S. Tell me what I've missed. I want to know if you see blind spots in the mechanic I'm not seeing from the inside.

Login or Subscribe to participate

Share this newsletter with a friend using the link here below. Means the world to me.

As a thank you, i’ll send you my actual AI ops stack. The skills and workflows I actually use every day at erlich textil and for my advisory clients.

Have a great day,

Philipp