
The five outreach mistakes that kill the most pipeline are calling stale lists, calling at the wrong hour, opening with a pitch, letting callbacks and follow-ups die, and running each channel on its own. We can see all five in the 1,000,000 French B2B cold calls we analyzed with Katana, and each one has a fix that costs nothing. Here they are, with the numbers.
A list exported in January and dialled in April is a list of people who have moved, changed priorities or already bought. The rep hears "not relevant" twenty times and concludes cold calling is dead.
The data says the list is dead, not the channel. Reps use about 20% of the signals already in their CRM. The other 80% (closed-lost deals turning one year old, contacts who changed jobs, website visitors from target accounts, dormant leads after a demo) is where half of our own pipeline at Sonik comes from.
The fix: no list older than two weeks, and every contact on it carries the reason you are calling today. If you cannot name the reason, do not dial. A rep who calls 40 contacts with a reason will have better conversations than one who calls 200 without, and the CRM already holds the reasons; it is a filter away.
Most teams call when the rep is free, which is 8am before meetings start or 2pm after lunch. Our benchmark puts 8am at a 9.8% connect rate; 4pm reaches 15.8% and 10am 14%, against an 11.3% average. Same rep, same list, 60% more conversations at 4pm than at 8am.
Persona matters as much as the hour. Founders pick up around 10am, marketing around 4pm, HR around 1pm. Sales and leadership personas connect above 15% overall.
The fix: calling blocks per persona, placed at that persona's slot, and the admin moved to the dead hours. Your calendar is a connect-rate lever.
65% of reps open by explaining their product. The prospect hears a pitch and the call is over in 20 seconds.
Trigger-based openers, which start with something that happened to the prospect ("I saw you posted for two SDRs last week"), reach a 7 to 10% meeting rate per conversation. Pitch-first openers reach 2 to 3%. Three times the meetings from the same conversations.
The same logic holds later in the call. Discovery questions that start with "how" lift conversion by 190%. Asking "What do you mean?" before answering an objection lifts it by 197%. The reps who book are the ones who talk less in the first minute.
The fix: one opener per list, written around the trigger, and the pitch removed from the first 30 seconds. Listen to ten openers this week and count how many start with "we".
Most pipeline is lost after a good conversation, not before. The prospect said "call me back Thursday", the rep was in a block on Thursday, and the callback happened the following Tuesday to a voicemail.
In our experience, most connects happen within the first three attempts. The follow-up is not optional; it is where the connect rate lives. And a callback, a prospect dialling your number, is the warmest contact in your CRM; in our experience it should be answered within the hour or it cools back into a cold call.
The fix: a 30-minute follow-up block each day, callbacks routed to the rep who spoke to the prospect, and every "call me back" logged as a task with a time, not a note.
The email goes out from the sequencer on Monday. The call happens from a different tool on Thursday with no idea whether the email was opened. The LinkedIn note arrives the week after from a third person. The prospect sees three strangers.
In Europe, cold email deliverability has dropped faster than mobile pickup rates, so the phone has to be the centre, and the other channels have to point to it.
The fix: one trigger per sequence, the email sent one hour before the call slot, the call referencing the email, the voicemail referencing both. A pickup ends the sequence; a conversation decides the next touch.
Sonik was built around the five fixes. Lists come from CRM signals and refresh themselves, so no rep dials a list older than the last signal. Sessions are scheduled per persona at the slot where that persona answers, and the parallel dialer gives reps about 5x more conversations per block.


AI call prep puts the trigger on screen as the line connects, so the opener is trigger-based by default. Callbacks ring the rep who had the conversation, and every "call me back" becomes a timed task. Call outcomes drive the next email, so the channels run as one cadence.
We measured the mistakes before we built the fixes, in our own outbound and in a million calls. That order matters.
Stale lists, wrong hours, pitch-first openers, dead callbacks and disconnected channels: five mistakes, five fixes, no budget required. This week, pick one. Our suggestion is the calendar: move one calling block to 4pm, keep the list and the script identical, and compare the conversation count after five days.