40 applications in, and you still can't answer the obvious question: what's actually not working?

Is it your resume? Your interviewing? The companies you're targeting? Or are you simply not generating enough conversations to find out?

Most people try to answer these things from memory, but once you're juggling a dozen applications, you stop remembering which recruiter promised a follow up, which rejection mentioned "communication," which company replied fast. You remember how last week felt, not what happened.

A job search should be a system, not a string of events you react to. In 2025 I tracked mine for 5 months to understand where my process was breaking.

The bottleneck wasn't where I expected

I assumed interviewing was my weak point, but the data said otherwise. The funnel looked like this:

• 40 applications -> 11 conversations with a human (28% pass rate)
• 11 conversations -> 6 interview loops, with 5 withdrawn by me (100%)
• 6 interview loops -> 4 team matches (67%)
• 4 team matches -> 2 offers (50%)

Once I was actually talking to someone, I converted well. The "leak" was earlier, in getting a human to respond. That number changed where I spent my time: I focused less on interview prep and put some effort into outreach.

Find the bottleneck before you optimise. Measure where your process breaks, then spend your effort fixing that step instead of improving what already works.

Different leaks, different fixes

Nobody responding -> visibility problem (unclear positioning, weak network, or not enough outreach)

Recruiter screens going nowhere -> narrative problem (your experience doesn't connect into a clear career story)

Technical or behavioural interviews failing -> interview readiness problem (technical depth, experience stories, or communication)

Interviews going fine but offers not following -> fit problem (level, experience, or culture mismatch)

Without the funnel, every one of these issues gets the same treatment: just try harder. With it, you know which lever to pull.

This system's job is to compress decisions

I used a spreadsheet, but the tool was incidental. This was Q1 2025, ancient times before LLM assisted workflows were common. The important part was the system: what I tracked and what I did with the data.

What it let me do was compress a lot of scattered information, salary, reporting line, team size, interviewer backgrounds, into one fast decision: is this somewhere I'd want to spend the next 3 to 5 years?

The system worked, but it wasn't perfect. Looking back, there are a couple of things I'd change if I ran it again.

I'd rate every role and company out of 5 starting from the first conversation. I did this eventually, and it made comparing very different opportunities trivial.

I'd log red and green flags the moment I noticed them, not from memory after. A disorganized recruiter. A hiring manager talking over me. A team member mentioning "firefighting" twice. Those are the bits that are you won't find in the job description.

One pattern in the data changed my direction. Initially I deliberately targeted startups, to push myself outside my comfort zone. Cold applications to smaller companies converted very badly. Larger, more established companies converted better, mostly because my experience carried more weight there and I had a network that I was able to use to get my foot in the door. It was a little embarrassing to admit at the time, but the data forced that correction faster than my ego would have.

Confidence, not optimism

There's one very underrated benefit I want to talk about, beyond optimising stuff.

Walking into an interview thinking "I really need this to work" changes how you present yourself, and not for the better. Knowing exactly where you stand doesn't make rejection hurt less, but it keeps your head clear.

"I haven't heard back" stops meaning "something's wrong" immediately. Some replies took more than 3 weeks to arrive, so a month became my real threshold for getting ghosted.

The same notes changed the questions I asked. Reviewing what I'd learned about a team before a final round meant I showed up on the call already understanding the role, which meant better questions, which is itself a signal to whoever's deciding whether to hire you.

Closing insight

Treat your job search like a system producing feedback, not a sequence of events you absorb and react to one at a time. Every application is a data point, and every rejection tells you something specific if you're willing to look for it.

What’s next

The hardest part of a job search is that you are optimising from incomplete information. You know what you sent, but you rarely know what made someone say yes or no.

Next, I’ll look at the other side of the table: how hiring managers assess risk, scope, and trust signals when deciding who gets moved forward.