Most job searches are run on the wrong channel, which is why they feel like shouting into a void.
The data on this is old and keeps being confirmed. Granovetter's research found that most people land roles through personal contacts, and disproportionately through weak ties: the ex-colleagues and acquaintances outside your daily circle, not your close friends. A large share of roles are filled before they're ever advertised. The portal, where most people put ninety per cent of their effort, is where CVs go to queue.
Today's edition is the system for the channel that actually converts. It works mid-search, and it works better built before you need it, because the day you need it tends to arrive without notice.
Step one: build the witness list
One hour. Write down every person who has seen your work first-hand. Not friends, not connections: witnesses. People who watched you deliver, sat in your meetings, received your output, managed you, were managed by you. Most people find sixty to eighty names once they work through old teams, projects, clients and suppliers.
The witness distinction matters because of what these people can do that a stranger cannot: vouch. "Worth a look, I rate her" attached to your CV beats any rewrite of the CV itself, because it converts a cold document into a referred one, and referred documents get read differently.
Step two: work the list at five per week
Five real exchanges a week, using a message with a specific shape (day 19 gives you the full anatomy; the short version is: specific connection, one-line context, small ask). The ask that outperforms everything: "Who should I be talking to?" Not "do you have a job for me", which puts the other person in an awkward position and gets polite silence. A request for direction is small, flattering and answerable from a phone, and the name they give you arrives with an implied introduction.
Step three: separate the two problems
Losing a job, or hating one, is a logistics problem. Losing confidence is a different problem, and mixing them poisons both: desperation leaks into outreach, which fails, which erodes confidence further. The operational fix is structural, not motivational. Split the days: mornings for the search, afternoons for work that generates evidence of your own competence. A small freelance project, a certification, building something. The point is a daily, concrete reminder that your capability didn't change when your circumstances did. Confidence returns through evidence, not affirmations.
Step four: measure conversations, not applications
The metric drives the behaviour. Applications-sent rewards queue-joining. Conversations-live rewards channel-building. Target: five live conversations at any time, where live means there's a next step with a date on it. A later edition (day 31) shows this run as a full pipeline; today, the number to know is five.
Why this compounds
The witness list is not a job-search tool that you build in crisis. It's a career asset that happens to be most visible in crisis. Maintained loosely (a touch every few months for the top names, which day 16's protocol handles), it means market shocks arrive at someone with sixty warm channels instead of a login to a portal. The difference in outcome is usually measured in months of salary.
Your assignment
Build the list today. One hour, every witness, whether or not you're searching. File it where you'll find it. If you are searching: first five messages go out this week.
Talk soon, Yasar
P.S. The OVD Lab's network mapping tool was built for exactly this list, and members call it the single best thing in there: https://ovdlab.com/for
