Skills by Matteo
3 Skills
Matteo Fois
Co-Founder at Kinetyca
Kill Or Keep
Judge every live outbound campaign against thresholds set before anyone looks at the numbers, and return one verdict per campaign: fix deliverability first, kill, scale, keep going, or too small to judge. One metric decides it, positive replies per send; opens and raw reply counts decide nothing. Reads campaign totals, the per-step funnel and per-sender stats through Clay, flags senders breaching the bounce guardrail, shows which sequence step produces the replies, and, when you supply per-variant results, calls A/B tests only when the sample can actually support a winner. It only reads and suggests; nothing changes unless you approve one specific action for one specific campaign at the end. Needs campaign data to read: Clay campaigns, a connected sending tool, or a campaign export file. Use whenever someone asks: should I kill this campaign, which campaigns are working, is this A/B test done, which variant won, my reply rate dropped what do I do, audit my live campaigns, which sequence step gets the replies, is this sender hurting deliverability. Do NOT use it to write or rewrite copy, to build or clean a list, to pause, launch or delete a campaign, to change senders or sending limits, or to diagnose DNS and authentication.
Matteo Fois
Co-Founder at Kinetyca
List Clearance
Check a lead list before it goes anywhere near a sequencer, and return one verdict for the whole list: go, fix, or no-go. Seven gates, each pass or fail on its own: buyer title, firm fit, verified email, no block-list or suppression hit, not already in a live campaign, required fields present, and contacts per company under the cap. The worst gate decides, never the average. It also cuts the usable subset out of any list, go or not, and hands back every rejected row with its reason. Reads a CSV export or a Clay table; spends no credits; uploads, edits and deletes nothing. Use whenever someone asks: is this list ready to send, check my list before I upload it, how much of this list is actually usable, why did my last list bounce, are the right people on this list, is anyone here already in a sequence, clear this list. Do NOT use it to find or verify emails (find-work-email, verify-email-deliverability, clean-email-list), to merge duplicate records (dedupe-contacts), to build the list in the first place (build-prospect-list), or to judge a campaign that has already sent (kill-or-keep).
Matteo Fois
Co-Founder at Kinetyca
Swap Test
Measure how much of a "personalised" cold email is template, then rewrite the template sentences from sourced facts. Every sentence gets one label: slot (would still read the same with another prospect's name and industry swapped in), a fact about the prospect, a fact about the sender's work, or exempt (greeting, sign-off, opt-out line). More than 30% slot sentences fails the email. Each slot sentence is then replaced by one that spends a fact with a source, or deleted; a number that appears in no source row never enters the rewrite. Run on a batch across several prospects and it also catches any rewritten sentence that came out identical for two prospects, which is template by definition. Returns the rewritten email, a before-and-after sentence table with the fact behind every change, and the density line ("2 of 9 swappable"). Use whenever someone asks: is this email actually personalised, how much of this is template, make this email specific, why does this read like a sequence, check my first line, review my ABM email, does this pass the swap test. Do NOT use it to build a prospect brief (an account research skill does that), to pick which case study to use, to check deliverability words, or to write a sequence from a blank page (a sequence-writing skill does that; this one takes its output).