Finn

Building AI products.
Documenting what works—and what absolutely doesn’t—in agents, automation and acquisition.

[Finn @finnsdesk]

I started SEO two months after launching my SaaS. Then I spent weeks publishing polished articles nobody was searching for. I made almost every beginner mistake. Here’s what I got wrong, what I did right, and how I’m rebuilding it🧵:

[Finn @finnsdesk]

@alepbuilds It’s true … always need market validation!

[Finn @finnsdesk]

Je partage mille fois tes conseils. On a toujours envie de passer rapidement en conversion … mais c’est con tant que la data suit pas ! Je descends aussi par palier de 50 ! On part de la landing, au signup, puis la fin de l’onboarding … pour enfin aller à la conversion. Et mille fois d’accord : 1 seul adset, 1 seule créa … ou alors il faut un autre budget ! En général, en revanche, au tout début, avec un mini budget, 2/3 créas en concurrence pour voir celle que l’algo retient ! Question : toi aussi tu penses qu’il vaut mieux mettre beaucoup d’un coup que d’étaler pour justement passer ce seuil des 50 pour le pixel … ou pas ?

[Finn @finnsdesk]

Paradoxalement … je te dirais de ne pas autant l’étaler. L’enjeu, le vrai, c’est au début de chauffer le pixel. Tu risque de cramer beaucoup au début … si tu ne mets pas assez ! Il vaut mieux mettre vraiment beaucoup au début pour passer la barre des 50 événements butoir d’Insta pour que le pixel « chauffe ». Vérifie bien que tout ton setup est parfaitement propre btw à ces niveau là d’investissement. Mets le CAPI aussi pour être sûr que tout remonte bien !

[Finn @finnsdesk]

@RosoAI Je comprends … le seul truc chiant, si j’ai bien tout compris et qui pour le coup est entièrement géré par Lemonsqueezy en tant que MoR, c’est la gestion de la TVA en tant que service numérique si tu vends à l’international !

[Finn @finnsdesk]

@sebuzdugan Exactly! It’s a good compass, not proof. In the end, real users decide whether the product actually solves something worth paying for.

[Finn @finnsdesk]

When I feel lost about where to take my product, I always come back to one very simple question: would I actually use it myself?

Not because I think I represent the whole market. I don’t.

But if I’m building something I wouldn’t use, don’t enjoy using, or that doesn’t solve a real pain for me, it becomes very hard to believe I’m creating something genuinely useful for others.

On the other hand, when I catch myself using my own product because it saves me time, removes something tedious, or gives me a result I genuinely need, that’s a strong signal.

If the pain is real for me, there’s a good chance I’m not the only one experiencing it.

Of course, that doesn’t prove product-market fit. You still need users, feedback, retention and people willing to pay.

But when I’m lost, it gives me a compass:

Would I use this if I hadn’t built it?

If the answer is no, I probably need to rethink something. If the answer is yes, then I need to find the other people who have the same pain.

[Finn @finnsdesk]

@sebuzdugan Absolutely!

[Finn @finnsdesk]

I started SEO two months after launching my SaaS. Then I spent weeks publishing polished articles nobody was searching for. I made almost every beginner mistake. Here’s what I got wrong, what I did right, and how I’m rebuilding it🧵:

[Finn @finnsdesk]

Not really 😂 @posthog is a product analytics tool (with a really generous free trial … I still haven’t pay since I started using itl!). It tracks everything users do in your product: clicks, events, funnels, drop-offs … and, most importantly for me : session replays, which let you see where you users get lost. Game changer for me!

[Finn @finnsdesk]

@Henrylabss Exactly! “Would I use it?” is a much better filter than “would this be cool to build?”

[Finn @finnsdesk]

@firasouesl4141 Exactly! That’s probably the simplest way to put it: first build something you genuinely need, then find the people who need it too.

[Finn @finnsdesk]

100 followers in 4 days. This place is honestly insane!

In less than a week, I’ve had more genuinely interesting conversations here than on most platforms in months.

Big thanks to @synopsi for taking the time to share how he’s using Pinterest to acquire users for @vaultaire_app

The fact that you can casually exchange ideas with people who have a lot of experience and deep technical expertise is still kind of insane to me.

What @ThibautMelen is building with Nika (a crazy portable agentic workflow system) and what @AlexDerocq is building with Casanea for people looking for apartments are both amazing.

You just post something, someone smart replies, and 10 minutes later you’re discussing distribution, agents, product, SEO, or whatever you’re both obsessed with.

Definitely loving X!

[Finn @finnsdesk]

@firasouesl4141 Thanks a lot! Exactly! It’s easy to spend weeks optimizing onboarding when the real problem is simply that not enough people are entering the funnel. Distribution has to come first—or you’re optimizing with almost no signal.

[Finn @finnsdesk]

One thing I’ve learned building a SaaS: progress is rarely about “fixing the whole funnel.” It’s about improving one step at a time.

First, ship something.

Then build a funnel to bring people to it.

Not enough traffic? Improve SEO, social media, distribution, and targeting.

Traffic but no signups? Improve the landing page, positioning, and offer.

Signups but people don’t finish onboarding? Fix onboarding.

Onboarding completed but nobody actually uses the product? Fix activation.

People use it once but don’t come back? Now you have a retention problem.

And only then does it make sense to seriously optimize conversion to paid.

The important part is knowing exactly where people drop.

That’s why analytics became essential for me. I use PostHog everywhere now—and I love it. Every important action is tracked, so instead of thinking “something isn’t working,” I can see exactly which step isn’t working.

Ship → Traffic → Signup → Onboarding → Activation → Retention → Revenue.

You don’t need to solve everything at once. Find the weakest step, improve it, then move to the next one.

Building a product feels much less chaotic once you start seeing it as a sequence of small problems to solve.

[Finn @finnsdesk]

@gaurangbuilds Agreed … but need at least frugal landing page!

[Finn @finnsdesk]

@otherelias @alexii_9 Good strategy for sure!

[Finn @finnsdesk]

@alexii_9 Thx for this discussion because it’s a real issue! You have unfortunately not a lot ways to find them … get traffic (SEO/GEO), mailing (Agentmail, warmup, …), Reddit (though to automated …), here on X … and after : ads, UGC, ..

[Finn @finnsdesk]

@senodev_ It’s better 😉

[Finn @finnsdesk]

@bowora_com For founders I don’t know … but for products, for sure! Some product need to be on instagram to be sell. You need to be where your customers are!

[Finn @finnsdesk]

@BuildWithDeepu It depends … most of the time … no!

[Finn @finnsdesk]

@thunkoid I’ve smiled!