How Many Articles Can One Expert Interview Produce? (1–4)

One well-run, hour-long expert interview can produce one to four publishable articles, and the honest median is two. The number is not set by transcript length — it is set by the smallest of four counts: distinct reader tasks the conversation actually served, distinct evidence for each task, distinct artifacts you can extract, and quotes you have permission to use in each new framing. A shallow interview supports one article no matter how long it ran; an exceptional one occasionally supports four or five, and inflating past what the transcript defends produces the thin near-duplicates that Google's spam policies describe as scaled content abuse.

This page is the capacity-estimation step: it gives you a defensible article count for a specific transcript before you commission anything. The decision of which angles become articles — and how to keep them distinct — is a separate job, covered by the interview-to-article splitting map.

Who this is for

This is for content leads, editors, and founders budgeting interview-led content: you are deciding how many briefs to open, what to promise a stakeholder, or whether one conversation can feed a month of the calendar. The interview was expensive to get, so the pressure is always toward a bigger number. The worksheet below exists to push back with counts you can point at.

It assumes a normal practitioner interview of roughly 45 to 75 minutes. Panel recordings, multi-hour oral histories, and five-minute hallway chats sit outside the ranges used here, and the worksheet says so rather than pretending one curve fits everything.

What decides how many articles a transcript can support?

Four counts decide capacity, and the smallest one wins:

Factor	What to count	Why it caps yield
Task-distinct angles	Jobs the conversation helps a reader finish ("do X," "avoid Y," "choose between A and B")	Each article must own a distinct reader task, or the pages answer the same query twice
Unique evidence per angle	Stories, numbers, named tradeoffs that belong to one angle only	Google's helpful-content guidance asks whether a page adds substantial value compared with other pages — including your own
Extractable artifacts	Checklists, decision tables, templates the transcript can honestly support	An angle without an artifact is a section, not an article
Permission coverage	Quotes and examples cleared for use in each new framing	Uncleared material cannot carry a page, and clearance for one article does not automatically cover a second

Transcript minutes matter only as a ceiling: a rough working assumption in this worksheet is that one publishable, task-distinct article needs at least 12 to 15 minutes of on-topic, specific conversation behind it. That figure is our own planning assumption, not a measured statistic — it exists so that a 50-minute transcript claiming six articles fails the arithmetic immediately.

The interview yield estimation worksheet

Copy this table, fill the count column from your transcript, and take the minimum. That minimum is your estimate; everything above it is a section or a cut.

#	Worksheet row	How to count it	Your count
1	Usable minutes	Minutes of specific, on-topic conversation (exclude setup, bios, tangents), divided by 13 and rounded down	—
2	Anchor stories	Complete incidents with a sequence: what happened, what failed, what they decided	—
3	Task-distinct angles	Reader jobs served with real substance, phrased as "helps someone do/avoid/choose"	—
4	Artifacts extractable	Checklists, tables, templates, or decision rules the material honestly supports	—
5	Permission-covered angles	Angles whose load-bearing quotes and examples are cleared, or clearly clearable	—
—	Estimated yield	The minimum of rows 1–5	—

Two reading rules:

• A minimum of zero is a real answer. If no angle has cleared, unique evidence, the transcript supports zero standalone articles today — its material strengthens existing pages instead.
• A minimum above four deserves suspicion. Re-check row 3 for synonym-splitting: "expert interview tips" and "how to interview experts" are one task, not two. When several of your pages chase one query, search engines pick a winner among the near-duplicates, which is the failure the estimate exists to prevent.

How do you count task-distinct angles honestly?

Count angles by finishing this sentence for each candidate: "after reading, someone can ___." If two candidates complete the sentence with the same verb phrase aimed at the same decision, they are one angle. Tag transcript passages with the job they serve rather than the topic they mention — topic tags ("pricing," "onboarding") overlap; task tags ("estimate a migration budget," "recover from a failed cutover") separate cleanly.

Then attach evidence to each angle exclusively. A story may anchor only one article; if two candidate articles both need the same story to stand up, merge them. The one tolerable overlap is a single bridging quote reused with different surrounding analysis.

Worked example: a one-hour interview, estimated

Assumption: the numbers below are a labeled hypothetical, not a real interview. An editor reviews a 62-minute recording with a platform engineer about a payments migration:

Worksheet row	Count	Notes
Usable minutes	48 → 3	14 minutes were bio, tooling small talk, and an off-topic hiring tangent
Anchor stories	2	The failed first cutover; the vendor-selection saga
Task-distinct angles	3	Run the migration; choose between vendors; avoid the rollback trap
Artifacts extractable	2	A cutover checklist; a vendor decision table (the rollback material is thin)
Permission-covered angles	2	Vendor criticisms need comms sign-off that may not come
Estimated yield	2	The rollback angle becomes a section inside the migration article

The plan this estimate defends: two commissioned articles, one section, and a follow-up question list for the expert. The plan it kills: the five-post "migration week" series someone proposed in the meeting.

When should you deliberately produce fewer articles than the estimate?

Take the estimate as a ceiling, not a quota. Publish fewer when the second article would land in a cluster you already own (merge the material into the stronger page), when the expert's review bandwidth is the real constraint — every derived article needs its own clearance pass under the quote permissions workflow — or when your maintenance calendar is already behind, because every additional URL carries a permanent review cost. One dense, well-evidenced page that owns its query outperforms two adequate siblings that split it.

Claim ledger

Claim	Source	Status	Review rule
Google's spam policies describe scaled content abuse as many pages that add little value, regardless of how they were created.	Google Search Essentials spam policies	verified on 2026-09-14	Recheck within 90 days or if guidance changes
Google's helpful-content guidance asks whether content provides substantial additional value compared with other pages in search results.	Google Search Central helpful content guidance	verified on 2026-09-14	Recheck within 90 days or if guidance changes
For duplicate or near-duplicate URLs, Google selects a canonical version rather than ranking them all.	Google canonicalization documentation	verified on 2026-09-14	Recheck within 180 days
The one-to-four range, the two-article median, and the 13-minutes-per-article assumption are editorial planning rules of thumb.	Our own, dated 2026-09-14	disclosed	Present as planning judgment, not measured data
The yield estimation worksheet and worked example are our own artifacts with a labeled hypothetical.	Our own, dated 2026-09-14	disclosed	Present as workflow advice only

FAQ

How many blog posts can one expert interview realistically support?

One to four, with two as the honest median for a good hour-long conversation. The count is capped by the smallest of: task-distinct angles, unique evidence per angle, extractable artifacts, and permission coverage — not by transcript length.

Does a longer interview produce more articles?

Only weakly. Minutes are a ceiling, not a driver: ninety minutes of generalities supports fewer articles than forty minutes anchored on two concrete incidents. If you want more yield, improve the interview's specificity with a run sheet rather than extending the call.

What should I do with angles that fail the worksheet?

Downgrade them to sections inside the articles that passed, hold them as follow-up questions for a second conversation, or use them as quotes inside existing pages. Material that fails the worksheet is rarely worthless — it is just not a URL.

Is publishing several articles from one interview bad for SEO?

Not when each article owns a distinct reader task with its own evidence and artifact. It becomes a problem when the pages are reshuffled versions of one answer — the pattern Google's helpful-content and spam guidance warns about, and the reason this estimate takes a minimum rather than a sum.

How is this different from the interview-to-article splitting map?

This worksheet answers "how many?" before you commit budget; the splitting map answers "which ones, exactly?" with per-article verdicts on task, query, evidence, and artifact. Run the estimate first, then the map for the angles that survived.

Where this fits in the interview pipeline

Yield estimation sits between the call and the commissioning decision. Upstream, the expert interview content brief and the interview run sheet determine whether the transcript has anything to count, and the SME question bank supplies questions that raise the yield before the call happens. Downstream, the splitting map assigns each surviving angle its task, query, evidence, and artifact, and the quote permissions workflow clears each derived article separately.

Sources and last-reviewed notes

Last reviewed: 2026-09-14. Sources checked: Google Search Central guidance on helpful, people-first content, Google Search Essentials spam policies (scaled content abuse), and Google's canonicalization documentation on duplicate URLs. The yield estimation worksheet, the one-to-four range, the two-article median, the 13-minutes-per-article assumption, and the worked example are our own editorial artifacts; the example uses labeled hypothetical numbers and no real client, expert, or traffic outcome is referenced.