My own agency became our first customer
I used to run a LinkedIn outbound agency for several years. At some point, we stopped asking people if they're looking for our services. Instead, we waited until they told the internet they are.
For example, a founder comments on LinkedIn: "Anyone know a good SDR tool?"
A VP of Sales likes three posts about outbound. A RevOps lead starts following every AI outbound company on LinkedIn.
Within 24h, I message them. I use a tool that we built in-house for our agency to identify those signals of interest.
Cold outreach is hard. Starting a conversation with someone who's already thinking about the problem is a lot easier. The conversion when doing outbound for our agency multiplied by ~5x.
That eventually gave us the confidence to fully concentrate on that internal tool. We're now running this as a standalone SaaS solution (IbexAI).
___
PS: Follow me on X along my building in public journey.
The strongest part of this is that the agency forced a real feedback loop before the SaaS existed. I would keep that loop visible in the product: show which intent signal led to each lead, when it happened, and why it is fresh enough to contact. That helps users avoid turning “high-intent” into just another scraped list, which is where outreach quality usually collapses.
nicely quoted
The 5x conversion result is valuable, but I would break it down by signal type before treating all engagement as intent. A direct problem post, a competitor comment, a follow, and three likes have very different precision. Compare each cohort against a matched cold baseline, track time-to-contact, positive reply, meeting, and eventual revenue, then show the source event beside every recommendation. That evidence also helps users avoid turning a useful timing signal into another high-volume list.
The shift from asking people if they need something to watching for the moment they say it themselves out loud is the real insight here, more than the tool built around it. It changes the entire cold outreach math because timing does more work than the pitch. Genuinely curious what percentage of those signals turn out to be noise once you scale past a small number of tracked accounts, since intent at small volume is easy to spot and gets messier at scale.
Nice
That's one of the best ways to validate a product. By becoming your own first customer, you experience every pain point firsthand and can improve the product based on real-world use rather than assumptions. Building for yourself often leads to a more practical, reliable solution that resonates with future customers. Great insight!