skip to content

HR's job adverts name your internal tools. What do you ask them to change, and what does it buy?

level: principalimportance: nice to knowfreq 30%

answer

  1. concede the floor before asking
  2. certificate logs cannot be withdrawn
  3. trim the increment, not the surface
  4. the gain is hours, not immunity
  5. spend on what survives a perfect message

basics

~20 s

Ask only for details with high lure value and low hiring value: internal system names and approval chains. It buys an operator an hour or two, not immunity, since filings and certificate transparency entries cannot be withdrawn.

solid answer

~50 s

Start by conceding that the surface cannot be emptied. Certificate transparency entries are a condition of holding trusted certificates, filings are compelled, and anything already published is archived. So the realistic ask is narrow: adverts can describe the capability rather than naming the internal system, and can avoid spelling out who approves what. That is a change HR can actually accept, because it costs little hiring signal. Be honest about the return: it adds research hours to an operator, and research hours are the cheapest thing they spend. The strategic call follows from that. If input reduction buys an hour and the surface has a hard floor, the durable budget goes to control classes that survive a perfectly credible request - phishing-resistant authentication such as FIDO2 or WebAuthn, out-of-band confirmation bound to the specific ask, and supplier payment-change processes that require a second path.

go deeper

for a junior

Recall that some published facts are compelled or permanent, so 'stop publishing it' is not a control you can simply apply to an organisation's public surface.

for a middle

Explain which specific details in an advert carry lure value - the internal system name and the approval chain - and what neutral phrasing preserves the hiring signal.

for a senior

Show you can state the return honestly in the adversary's currency: added research hours, against a surface with a hard floor and an operator who can spend another morning.

for a principal

Own the allocation. Concede the floor, scope an ask another function can accept, and argue the durable budget toward control classes that still hold when the request is perfectly credible.

### Why this is a judgment question, not a technical one The facts that make lures credible are published by people who do not report to you, for reasons that are good. Recruiters name systems because it attracts candidates who already know them. Marketing publishes bios because speakers sell. Filings are compelled. Certificates are logged because that is the price of being trusted. Every reduction you ask for is a cost to someone else's objective, and they can refuse. That is what makes this a decision about scope, evidence and where the money goes rather than a control to deploy. ### First: establish the floor Before asking for anything, be clear about what is not available at any price. - **Certificate transparency entries.** Publication to public append-only logs is a condition of issuance for publicly trusted certificates. Entries are not withdrawable, and the hostnames in already-issued certificates are permanently public. - **Statutory filings.** Officers, signatories and structure are public because the law requires it. - **Everything already published.** Adverts, bios and pages that have been indexed or archived do not become unpublished when you take the page down. So the surface has a hard floor, and any proposal that implies emptying it is not a proposal. ### Second: scope the ask to the increment What is left is a narrow band of details with high lure value and low business value: - Adverts that name the internal expense, ticketing, payroll or approval system by name. The hiring signal is preserved by describing the capability instead. - Adverts and bios that spell out the approval chain - who signs off on what, and at which threshold. Almost none of that helps a candidate decide to apply. - Bios that pin a named individual to a named authority, which is what turns an organisation-level lure into an attributable instruction. This is an ask a recruiting lead can say yes to, because it is specific, it is small, and it does not ask them to advertise less clearly. ### Third: be honest about the return The honest gain is that the operator has to spend more time, and possibly reach a less precise result. That is a real effect and a small one. Research hours are the cheapest input in the whole operation, and the crew has the option of simply spending another morning. Any business case that promises the end of tailored lures against the company is a case that will be checked and found wrong, and it spends credibility you will need for the larger ask. ### Fourth: the decision the leadership question is really about With a floor on the input and a cheap workaround for the adversary, the strategic conclusion is that the durable spend is on the outcome, not the input. The value of a control class here is exactly whether it still works when the request is perfectly credible: | Approach | Holds when the lure is perfect? | | --- | --- | | Reduce published facts | No - buys hours, has a floor | | Rely on the reader judging plausibility | No - priced against the cheap rung | | Authentication that cannot be relayed to an attacker-controlled site | Yes | | Out-of-band confirmation bound to the specific request | Yes | | Payment or supplier-detail change requiring a second path | Yes | That is the argument to make to a budget holder: the input side has a floor and an adversary who can re-earn any ground you take with two more hours of reading; the outcome side does not. ### Fifth: how to run the ask so it survives Take the partial win. Get the wording change into the advert template rather than negotiating advert by advert, so it becomes the default rather than a recurring request. Give the recruiting team a phrasing they can use instead of a prohibition, because a prohibition without a substitute is ignored under deadline. Do not chase historic content that is already archived. And bring the same honesty to the review afterwards: the measure of success is that the template changed, not a claim that anything was prevented. ### The wrong answer The weak version of this answer is an instruction to stop publishing, delivered as if security could issue it, priced as if it were free, and justified by a benefit it cannot deliver. The strong version concedes the floor, scopes the ask to what a hiring team can accept, states the modest return in the adversary's own currency of hours, and then spends the real budget where a credible message cannot be talked out of failing.

  • Which published facts can you not remove at any price?
    Hostnames recorded in public certificate transparency logs, because logging is a condition of issuance and the entries are append-only; statutory filings naming officers and signatories; and anything already indexed or archived. Taking a page down does not unpublish it. That floor is why input reduction is a rate limiter rather than a control.
  • How do you make the ask without telling the business to advertise less clearly?
    Scope it to a specific substitution rather than a prohibition: describe the capability instead of naming the internal system, and omit the approval chain, which no candidate needs. Put it in the advert template so it becomes the default, supply the replacement wording, and accept a partial win rather than negotiating each advert.
  • If you can fund only one of input reduction or outcome control, which and why?
    Outcome control. The published surface has a hard floor and the operator can buy back any ground you take with another morning of reading, so input reduction is cheap for them to defeat. Controls that hold when the request is perfectly credible - unrelayable authentication, out-of-band confirmation of the specific ask - cannot be re-earned with research hours.
  • How would you report on this so it does not overclaim?
    Report the change that actually happened - the template wording, the share of live adverts following it - and state the expected effect as added research hours for anyone preparing a lure. Do not report prevented incidents, because you cannot attribute them, and an overclaim here spends the credibility needed for the larger control-class argument.

saying these in an interview costs you the question

  • Proposes deleting the public surface as if that were available
  • Treats certificate transparency entries as suppressible
  • Prices the ask without naming the business cost to hiring
  • Assumes another team will comply because security asked
  • Claims the change will stop tailored lures against the company

context