I've been meaning to do this for a while. Every month we as a team make real calls, and 90 days later I can't always reconstruct the reasoning. That's a problem. Half of what I do in advisory is helping someone else reconstruct a call they made six months ago. If I can't do it for myself, I shouldn't be charging for it.

So the big calls are going in here from now on. Question, call, reasoning, confidence score at the time. If I'm wrong later, I'll come back and say so.

Here are April's three. One of them I'd bet real money on. One of them I'm nervous about.

Question: Do we replatform the entire store in the middle of our busiest stretch of the year, with a skeleton crew and one freelancer?

Call: Yes. Launch May 5th.

Reasoning: Four things, in order of weight. Agility comes first. A small team needs to ship fast, and our current stack punishes speed. Second is frontend performance. Store speed and a conversion rate which is not fully maxed out is a tax we've been paying quietly for years, especially for paid acquisition. Third is usability. One person on our team should be able to run 80% of day-to-day store work without opening a dev ticket or getting lost in weird menues which look like windows 95. Fourth is the Shopify app ecosytsem. It closes gaps we'd otherwise have to build ourselves or pay a freelancer, and building ourselves is the most expensive thing a 7-person team can do. Also, shopware just feels like SAP and nobody wants that.

The timing is absurd on paper. Spring Sale just wrapped. Our summer lingerie drop Dreamy Lace Wilde Orange (super excited about what the team has put together here!) launches three days before go-live, which means we're setting the collection up on both platforms, uuugh. Seven days after launch, we kick off something that will shift a meaningful slice of marketing budget into organic (more on that in decision 3). All of it pulled off with our own team plus one external freelancer, with a lot of AI-assisted workflows and very short decision paths.

Confidence at the time: 7/10. The minus three points are the timing. The seven points are the math on what we currently lose to store speed, dev backlog, and app gaps we can't close without an engineer we don't have. If you're a 3 to 10M brand on Shopware right now, I'd tell you to check your 12-month roadmap and ask yourself how much of it depends on speed and agility you don't currently have.

Question: Is it worth spending real money to bring back customers who bought from us years ago and only know our old products? And are therefore not up to speed with all of the quality improvements?

Call: Yes. Print first, with a tactile quality angle, good messaging and a story about the new collections and why it’s worth to give us another try. Personalised email flow as follow-up.

Reasoning: First, the product. If you don't fully trust what you're re-introducing, don't run a win-back. A win-back is about resetting perception, and if the perception you land them on is still wrong, you've wasted the spend. Second, the mechanic. Print still cuts through for a segment that hasn't bought in 24+ months, and email alone does not move them. Third, the math. Dormant cohorts are the single most underpriced list a brand has. We already paid to acquire them years ago. They just forgot we got better.

What could potentially kill such a thing is stoppable. Four things: off-brand messaging because you’re rushing it, wrong timing (we deliberately landed ours away from high-noise retail weeks), a sloppy address sync where you don’t do a second adress check, and a wrong Klaviyo audience definition that includes or excludes the wrong people. All four are preventable.

Early result: Close to CM2-positive on first order after 5 days. Will be positive after this weekend. I’m projecting this to go up to 3x the costs we put into this.

Confidence at the time: 10/10. That's rare for me. But this one always works.

Question: Do we deliberately cut into our largest new-customer channel right when we need volume most?

Call: Yes. Starting mid-May.

Reasoning: The number that convinced me is 65. That's what we currently pay Meta for a new customer. A bra costs us significantly less than that in COGS. You can do the math. The channel works if lifetime value earns it back and we’re currently managing to do that, but the trendline is not going into the right direction. Average CAC in DTC is roughly 60% higher than it was five years ago, and most small brands I talk to are pretending that's fine.

Here’s a little hint of what we’ll be doing: Word of Mouth is our second biggest acquisition channel behind email. Zero euro CAC (well, almost). Highest AOV of any channel. Less returns, higher CLV, higher fullprice share. People want to recommend us, they just need a reason and a story. So we're going to give them one, at scale, in May.

We're trading short-term top-line for the kind of cohort you can't buy on Meta. And for a message we stand behind and i believe will resonate with a lot of people out there.

Confidence at the time: 7/10. The minus three points are that cutting paid at peak risks a visible dip, and we know it. Also it’s a bold message, which it needs to be. I’m still nervous and taker this as a good sign. The seven points are that we trust the quality of the customer we're trying to attract, and we think the Meta trade has quietly become bad for brands like ours.

Next newsletter is the full reveal.

Some of these I'll be wrong about. Decision 3 especially could go wrong. I'll update you when we know. If you have strong views on any of the three, drop me a line, I mean it. Reader pushback is the kind of input i really value!

And if there's a decision on your own desk right now that's costing you sleep, the quickest useful thing I can offer is this. Write out four things and keep them separate: the question, the call, the top three reasons, and your honest confidence score. Most of the calls I see stuck are stuck because the person making them never separated those four. If you want a second pair of eyes on it after that, reply to this email. Roughly half my advisory calls start that way.

Until next week, 

Philipp

If you like what you’re reading, share this newsletter with a friend using the link below in that little box. 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.