Why your contract CV gets no callbacks

Most contract CVs get no callbacks for a boring reason: the claims fall over at the first follow-up question. Not in an interview – at the desk, during the fast read, where the question gets asked silently and the CV gets no chance to answer it.

I've spent the years since 2018 reading contract CVs at a specialist IT desk in Brisbane. Those years aren't the whole credential – the desk around me held decades more, and I spent mine absorbing the questions its best readers asked. This playbook is the interrogation I now run on every claim in a paid rewrite, written down in full. Five probes, each applied to a real bullet, each ending with the rewrite. Run them on one role this afternoon and you'll finish with rewritten lines, not notes.

One promise, stated carefully. The probes don't promise interviews. They take the CV off the suspect list, so whatever happens next is information.

The part you control

A contract search runs on variables nobody hands you. The organisation's process. Whether the recruiter can reach the hiring manager. Timing. Who else applied that week. Whether the role survives to Friday. None of it sits in your hands, and most CV advice goes wrong by pretending otherwise.

Two things do sit in your hands: what you share, and the effort it costs a stranger to understand it. Call that second one translation cost. It's the whole game.

Here's why. Your CV is read most often by people who were never in the room with you. The fast read at the agency. The retelling in a client's office. The panel's pre-read the night before. Every hop taxes translation cost, and the tax compounds – what survives the first read still has to survive being repeated by someone who read it once.

And one mechanism from inside the process, stated plainly because it rarely gets said out loud: for the thirty seconds your CV sits in a client's inbox, it's the sender's work product. A recruiter forwards a CV they're proud of faster, and with a better cover note. Lower the sender's risk and your submittal odds rise. It moves outcomes and it never appears in any process document.

Why the fast read exists

A desk runs roles at every phase at once – some at search, some at submittal, some at interview – and spread like that, the load is manageable. The crunch comes when several arrive together, because search is the heavy phase: each one opens a three-to-five-day window to get a shortlist in, each ad pulls in its hundreds, and the reading stacks into the same week. Most contract roles go out to three agencies at once; some go to ten, and federal government panels run wider again. So every desk is racing every other desk to have candidates covered first. Speed is the job. The fast read is what speed looks like.

The ad is also only one door. Plenty of shortlists get built from the other direction: a keyword search over the agency database, a LinkedIn search, a candidate half-remembered from a role six weeks back. And the signals that surface you can be stale – an availability flag from your last check-in, a tool you touched ten years ago still matching a search today. Your CV competes in retrieval as much as review. When the call comes off a stale signal, the page is what corrects the record.

The fast read isn't laziness – it's arithmetic. Reading for signals is a skill, and the good readers extract a career from a page faster than feels fair. Your job is to make the extraction cheap, and the memory sticky.

Why these five questions

The word coming up on the client side more often than ever is unicorn. What it actually means: matching went multi-axis. The question stopped being "have you delivered this" some time ago. It's now delivered this, in that kind of stakeholder environment, at that phase, leaving that state behind. Vague CVs fail the test by default, because there's nothing on the page to match against.

The five probes are that matching function, reverse-engineered. They're the questions the desk already asks – written down for your side of it.

Probe	The axis it feeds
1 · Why did the work exist?	Phase and driver
2 · What were you brought in to do?	The mandate
3 · What did you actually do?	The deliverable
4 · Who did you have to move?	Stakeholder terrain
5 · What did it leave behind?	State at exit

One guard before we start, because it's the difference between this method and embellishment: the probes can't manufacture contexts you didn't have. They stop you hiding the ones you did.

One note on order. The desk's questions run in a fixed sequence: where has this person delivered this – and only when the answer is yes, what was the context, and was it comparable. The probes deliberately run differently. A claim gets built from its foundations, force first, but a page gets read match-first. So the bullets carry what the scan hunts, and the context sits directly above them, where the eye goes on the yes.

Every bullet below is a composite of real CVs with details changed. The patterns are exactly as they arrive at the desk.

One note on format. This playbook works at the claim level, and claims live anywhere on a CV: a bullet, a context line above the bullets, plain prose. Where each answer belongs on the page, and how hard you tailor once you know what you're hunting, is a later playbook's job. Here, interrogate whatever you've got.

Probe 1 · Why did the work exist?

The read. This is a job description, and the job description was never in question. Every BA in the country holds these duties, so the fast read prices them at zero and goes looking for the thing the page withheld: the reason. Somebody approved a day rate for this seat. Something forced that. A regulator, a deadline, a platform going end-of-support, an audit finding, a merger. The force gives the work its phase, its stakes and its shape.

This is why most duties lists fail probe one. Duties describe the role, and the role was never the reason anyone paid for it.

What changed: the force showed up, and the duties collapsed into verbs with objects. The reader now knows the phase, the stakes and what the seat was for – an underpayment review needs no further explanation.

Probe 2 · What were you brought in to do?

The read. The program is doing all the work in this sentence, and the program isn't applying for the role. Contractors get engaged for something specific: a stream to run, a capability to stand up, a mess to remediate, a gap to hold. That mandate is the first thing the desk tries to extract, because the match is made against it – the PD describes a mandate, and the fast read hunts your bullets for its twin. A $30m program seats forty people. Sponsor, program director, stream leads, PMO, coordinators, analysts. "Part of the team" covers all of them equally, so it identifies none of them, and the fast read defaults to the smallest mandate the sentence supports. Its cousin is the "we delivered" bullet: we names a crowd, and the shortlist wants one person with one mandate. The mandate also rarely belongs inside a bullet: its natural home is a context line above them, which is where this rewrite puts it.

Inflation gets priced the same way. A claimed mandate that out-runs the evidence gets discounted harder than a modest mandate stated precisely.

What changed: the mandate moved to its home – a context line above the bullet, phase included – and the bullet went back to carrying evidence. The ceiling line stays: "go/no-go authority sat with the program director" costs nothing and buys trust.

Probe 3 · What did you actually do?

The read. "Involved in" spans everything from ran-it to attended-the-standup, so the reader supplies the smallest meaning the words will hold. That's how sceptical reading works: vague claims get priced at their floor. The fix is verbs with objects. Planned what. Ran what. Wrote what. Sequenced what around what.

One more thing hides in this probe: dialect. Your organisation called it "the TOM uplift"; the market calls it operating model redesign. The PD's words are the search terms of the fast read, so internal dialect reads as a miss even when the work matches. Translate the work into the words the market uses for it – the PD will tell you which ones.

What changed: three verbs with objects replaced one preposition. The wave-one admission stays in – a survivable setback, stated plainly, reads as evidence the rest is true.

Probe 4 · Who did you have to move?

The read. This is an adjective standing where evidence should be. Every CV on the desk claims it, so it prices at zero on arrival. Meanwhile it's the axis clients probe hardest. Nearly every panel question I prep candidates for is a version of the same thing: someone pushed back – what did you do? The CV that answers before the panel asks is the one that gets retold in the client's office.

The discipline: describe the resistance structurally. Name the situation that made holding out rational, then show the change that made it stop being rational. The people stay acquitted; the mechanism does the explaining.

What changed: the resistance has a shape (a tool doubling as a register), the move has a mechanism, and the problem stays with a tool, never a person. That reads as accuracy, and accuracy retells.

Probe 5 · What did it leave behind?

The read. "Significant" is a number that refused to commit, and "initiatives" is work that refused to be named. But the contract-specific question sits deeper. You roll off. Everyone prices that in. So the desk's real question about any contractor claim is: what was still standing two quarters later? A saving that lapsed at rollout was a visit. A capability that survived the exit is the thing worth a day rate.

What changed: the number committed, and the tense does the heavy lifting. "Now chair" and "runs without me" are survivals stated in the present. Probe 5 is mostly a tense – what still is, versus what once was.

The retell test

Your CV gets discussed in rooms you're never in. The sound of a CV working is a hiring manager saying "who's that candidate who rebuilt the asset register – the one who did the payroll remediation before that?" Retold, from memory, by someone who read it once.

That's the final gate, and it's stricter than the probes: could a stranger retell your claim in one sentence after one read? Concrete nouns survive retelling. Adjectives die in the first hop. It's also how the database door opens – the half-remembered candidate who gets the call is the one whose nouns stuck. If a claim can't be retold, it can't be matched. Run it back through the probes until it can.

The checklist

Run every claim through six gates.

What this can't do

These are questions, not rules, and the judgement stays yours. The probes fix the paper; they can't manufacture experience, and a CV that survives them can still meet silence. If it does, that's information. The suspect is fit, aim or process, and The Bench exists to tell those apart.

---

Playbook 002 takes the same evidence into the room: the interview is the bullet, spoken. One email a month when a playbook drops. Take a seat →