Digital PR & Earned Media Lead interview questionsEarned Media Round — The Coverage That Did Nothing round

Purpose

Explicitly licences a negative verdict on the interviewer's own team's work, because the second section requires exactly that and a candidate who suspects they are insulting their host will hedge into uselessness. The constraint about using existing assets is deliberate: it separates people who can find a story in what a company already holds from people whose plan begins with a research budget.

Purpose

The opening discriminator and it resolves on whether the candidate treats the count as data or as a claim. Forty-seven is not forty-seven of anything; it is two editorial pieces and a syndication tail. A candidate who opens two of the URLs inside the first three minutes and notices the text is identical has done this job.

Signals to score

• Opens several of the placements rather than reasoning about the list abstractly
• Notices that many carry identical text and names syndication as the mechanism
• Asks whether a wire service was used, and treats that as the explanation for the count
• Checks the links themselves: whether they exist at all, and what rel attributes they carry
• Separates a mention from a link from a followed link, and refuses to sum them
• Asks which placements the target audience actually reads, rather than which have the biggest vendor score
• Asks what the campaign was trying to achieve before judging whether it did
• Asks what "flat" means and over what window, and whether referral was ever a plausible outcome
• Treats two genuine editorial pieces as the real result and asks how those two came about
• Asks who wrote the report and what it counted, without being sneering about it

Follow-up questions

• Forty-seven is what the tool reported. Is the tool wrong?
• Which of those placements would you have been happy with, if any?
• Some of them are on sites with very high domain scores. Doesn't that count for something?
• Referral traffic was flat. Should it have moved?
• If I gave you the same budget again, what's the first thing you'd change?

Purpose

The question that separates this role's current version from its previous one. Links remain the durable mechanism and a candidate who abandons them for a fashionable story is wrong; but a mention with no link now has a route to influence that it did not have five years ago, and a candidate who cannot articulate that route is behind. The round wants both held at once.

Signals to score

• Values a followed editorial link on a relevant publication as the durable asset, and says why relevance beats reach
• Knows that nofollow and sponsored are treated as hints rather than absolute directives, and does not overstate what that means
• Says plainly that an unlinked mention passes no link signal, before discussing what else it might do
• Articulates the route by which unlinked coverage matters now: it is a source an assistant may retrieve and cite
• Names the specific surfaces that get cited for commercial questions — review aggregators, forums, journalism, video — rather than saying "AI"
• Does not abandon links, and resists the framing that link building is over
• Distinguishes coverage that states checkable facts about us from coverage that merely names us, and says the first is more useful to a retrieval system
• Raises consistency: coverage repeating wrong facts about us is actively harmful in a way it was not before
• Asks whether we know which sources actually get cited for our queries, rather than assuming
• Proposes going after fewer, more relevant, more factual placements rather than more of them

Purpose

Data-led earned media is the only reliable way to earn genuine editorial coverage at this budget, and it turns on a specific skill: finding the question a journalist would ask that our data happens to be able to answer. A candidate who proposes a survey has not read the constraint; one who proposes an analysis with no news value has not read the audience.

Signals to score

• Proposes an analysis of first-party data rather than a survey or a commissioned study
• Finds an angle with genuine news value — a trend, a regional difference, a seasonal effect, a counterintuitive finding
• Names the audience before the outlet: who cares about this, and which desk covers them
• Says what the headline would be, in a journalist's words rather than a marketer's
• Anticipates the methodology questions a journalist will ask — sample, period, how representative
• Raises privacy and aggregation before being asked, and says what could not be published
• Plans for the finding being boring, and says what happens then
• Proposes taking it to a small number of named outlets rather than a wire release
• Offers an exclusive or a first look as the mechanism for getting editorial rather than syndicated coverage
• Names what the company gets even if coverage is thin: a reusable asset, a citable page, a fact that lives on

Purpose

Earned media has the worst measurement problem of any function in marketing and the honest answers are the useful ones. A candidate who promises attribution is either naive or selling; one who has nothing is unemployable at this level. The separation between evidence and inference is the whole score.

Signals to score

• Separates directly observable outcomes from inferred ones, explicitly and without being asked twice
• Counts links and mentions as leading indicators rather than as results
• Names something checkable: referring domains gained, the specific pages that gained links, whether those pages moved
• Proposes measuring whether the cited sources changed what assistants say, and admits the causal chain is long
• Declines to attribute revenue to a placement, and says why the attribution would be fictional
• Offers brand search volume as a signal while naming its confounds
• Proposes a before-and-after on a specific, narrow claim rather than on the business
• Suggests a holdout or a staggered regional rollout where the campaign structure allows it, and says when it does not
• States what would make them call the campaign a failure, in advance
• Reports the two or three things they are confident about and does not pad the deck with the rest

Purpose

The first half finds out whether the candidate has a line before somebody tests it. The second is the scenario this role will actually face, the word accurate is doing the work, and the answers separate people who understand earned media from people who understand publicity.

Signals to score

• Names something specific and declines it on the merits — paid links presented as editorial, undisclosed sponsorship, fabricated survey data, review incentivisation, private blog networks
• Knows that paid links passing endorsement without disclosure violate search engines' published policies, and says what the enforcement risk actually is
• Defends the refusal in the CMO's currency — detection, enforcement, the cost of the recovery — as well as on principle
• Reads "accurate" in the second scenario and does not propose fighting the facts
• Distinguishes correcting an inaccuracy from disputing an unflattering truth, and treats them completely differently
• Escalates internally before responding externally, and names who needs to be in the room
• Considers that a response can amplify a story that would otherwise stay small
• Raises that the piece is now a durable source and may be retrieved and cited long after the news cycle ends
• Proposes addressing the underlying problem as the only durable remedy, and says so plainly
• Gives sales and support something accurate to say inside the first day

Purpose

Discloses that the flawed artifact is current and owned by people the candidate would have to work with, which is the honest version of the job and the thing most likely to be withheld. Naming the unclaimed opportunity is the strongest available closing argument at this band, and it costs nothing because a candidate who is not excited by it is not the hire.