Maybe It's a Signal

Back in Paris, back at the desk.

And this Sunday is the right time to move forward. I'm rested after 2 months by the sea, I have a clear head about the projects in progress and the directions to take. Codex limits reset tonight, so practically a full day without worrying about token consumption.

Six months in the void

For the past 2 days, I've gotten very seriously back into LinkFeed Pro and XFeed Pro, two products I launched ~6 months ago that never really took off.

Broadly speaking, they very noticeably improve the UI of LinkedIn and Twitter discussion feeds, by removing all the blocks that aren't content and allowing you to increase the font size and feed width. For fifty-somethings who are constantly zooming in on their screen, it's a must-have...

I kept maintaining them mainly because I use them every day. I had practically stopped expecting anything from them.

Initially I thought that having natively designed the product (site and extension) in multiple languages would automatically position me in the Extension Stores and drive traffic. I'll give you one guess: zero results (and I mean truly zero!).

I did a few (bad) LinkedIn posts and some ads (wrong target audience) on my Udemy coupon site (Comidoc) with no more success.

A few attempts at guides to improve the sites' SEO with the help of Fable 5 as well, but again it was all pretty negligible.

$29 falling from the sky

Then, out of nowhere, one evening on August 15th, an email notification in my building lobby: a lifetime sale.

At first I couldn't believe it. I didn't even remember the notification email template. I read it 3 times.

Another email from Stripe arrived, confirming the sale.

No traffic. No marketing. A Store presence but no reviews. And someone still paid $29.

30 minutes later an email from the buyer because the product wasn't working the way he'd hoped (my first user feedback!).

I open Codex and off we go. Indeed, my use case (reading my feed) didn't match my buyer's (mainly searching). What followed was about half a dozen submissions of new extension versions over a few days; the product made a significant leap in features and robustness.

Copyable but not maintainable

I've come to think that yes, the product is copyable, for sure, especially in the age of AI. It's essentially just CSS rules to apply in real time on certain pages, at specific moments.

But the complexity involved in managing the particularities of each language, changes to LinkedIn's structure as it evolves, the rules engine that doesn't require republishing the extension when LinkedIn makes changes... all of that still creates a solid barrier to entry.

You can ask Claude to code the same thing. It'll get close to working for 7 days (it will always be missing months of trial/error, optimizations, and polish) then it'll break at the first LinkedIn format change, and so on.

When all that's left is marketing

The advantage of this type of single-function product compared to the complex SaaS products I've also attempted is that once you've refined the technical foundation and execution to the maximum, there's nothing left to do in terms of code.

Whereas with complex SaaS I always have a feature list as long as my arm, and it only keeps growing. Here you quickly reach the optimum and I no longer have any valid excuses to avoid the part I usually steer clear of at all costs: marketing.

So here's where we stand.

Put better screenshots on the stores, polish the descriptions, micro-optimize the landing page, communicate on social networks (this post is part of that!), maybe ads if the traction holds up.

It's time to see if this sale was a stroke of luck… or a real signal.