Salesoo Blog
How we delivered 78% lead relevancy on a first pilot — and what we learned
When Dataddo came to us, they needed to find the right companies in the CZ/SK market fast. This is what our process looked like, where we got it right, and what the second pilot will do differently.
How we delivered 78% lead relevancy on a first pilot — and what we learned
Every methodology sounds convincing on paper. The real test is the first live project — when your assumptions meet reality and you find out which ones were right.
Our first paid pilot was with Dataddo, a data integration platform operating across the CZ/SK market. They came to us with a clear need: identify companies that were genuinely ready to explore a data integration solution, not just firms that looked good on a spreadsheet.
Here's what happened.
The brief
Dataddo's sales team was experienced. They didn't need more contacts — they needed better ones. The ask was specific: a shortlist of companies showing active signals of readiness, with a verified decision-maker and enough context to make the first conversation meaningful.
No cold lists. No guesswork. Intelligence.
How we approached it
We started by mapping Dataddo's real ICP — not the aspirational version, but the behavioral one. What did their existing best clients have in common six months before they signed? What were they hiring for? What had their leadership said publicly?
From there, we ran our two-layer scoring process. Layer one covered direct evidence — job listings signaling data infrastructure investment, leadership statements about digital transformation, recent funding rounds, and technology migrations. Layer two covered qualitative judgment — timing window, contact quality, information availability, and absence of vendor conflict.
Every company on the final shortlist had to clear both layers. If the signal was there but the timing was wrong, it didn't make the cut.
The result
Of the leads we delivered, 78% were assessed by Dataddo's team as genuinely relevant — companies they would actively pursue. For a first pilot, with no prior feedback loop to calibrate against, that number validated the core approach.
But 78% also means 22% weren't right. And that gap is where the real learning was.
What the 22% taught us
Some misses came from incomplete public data — companies where the signal was real but the context was missing. Others came from timing — firms that would have been a strong fit three months earlier or later. A small number were simply wrong calls, where our interpretation of a signal didn't match the ground-level reality Dataddo's team knew from experience.
Every one of those misses went back into the algorithm as feedback. Not as a failure, but as calibration.
What the second pilot will look like
The scoring model now carries Dataddo-specific weighting — patterns learned from what their team confirmed and rejected. The second shortlist will be built on a foundation that didn't exist when we started.
That's the point of the concierge phase. Not to be perfect from day one, but to get smarter faster than any static database ever could.
78% was a strong start. The next number will be higher.
Changelog