How to Research a Podcast Guest in Under 90 Minutes
You can research a guest well in 90 minutes if you spend it in the right order: 40 minutes on their own work, 25 on two of their past interviews, 15 on recent social activity, and 10 writing a one-page brief. Read what they made before you read what others said about them. That sequence is the whole trick.
Most hosts get this backward. They skim a Wikipedia page, scan a LinkedIn, watch one viral clip, and walk in with the same five questions every other host already asked. The guest answers on autopilot, the conversation never lifts off, and nothing gets clipped. Below is a time-boxed workflow that fixes that, a fixed source order plus a brief template that turns your reading into questions you can actually ask on mic.
How do you research a podcast guest quickly?
Work in four tiers, in this order: their own work (books, posts, talks, episodes they hosted), then two of their past interviews, then their recent social and news, then a 10-minute synthesis into a one-page brief. The order matters more than the hours. Start with what they made, because that is where the questions nobody else asks live.
The reason to time-box it is that guest research expands to fill whatever you give it. Without a cap, you read a fifth article and a tenth tweet and end up with trivia, not questions. Ninety minutes is enough to sound informed and find two or three threads no other host has pulled. Here is how that time breaks down.
Why the source order is the whole game
The order beats the hours because each tier has a different job, and doing them out of order wastes the early, fresh part of your attention on the lowest-yield material. Their own work gives you original questions. Past interviews tell you what to skip. Social tells you what's live right now. The brief turns all of it into something you can read at a glance while recording.
This matters more than it used to because the interview isn't just for the people listening live, it's for the clip. 57% of listeners now rely on social media for podcast recommendations, the first time it has passed friends and family (InsideRadio, 2025). The moments that travel are the ones where a guest says something they haven't said on ten other shows, and you only surface those by reading what they actually made, not by recycling the press-tour questions. The audience is also watching more than before: 53% of new US weekly listeners now prefer to watch a podcast, up from 30% in April 2022 (Backlinko), so a host who's visibly fishing for a question on camera is obvious in a way they never were on audio.
If your booking is already on the calendar and you want the conversational skills to go with the research, asking open questions instead of leading or yes/no ones is the natural next read.
The four research tiers, ranked by question-yield
Not every source produces questions at the same rate. Below is the order I'd spend time in even if I only had 30 minutes, based on which sources reliably turn into questions a guest hasn't been asked before. The ranking is editorial, my read from coaching hosts and guests, not a survey, but the gaps between tiers are real.
Tier 1: Their own work (40 minutes)
Start with what the guest actually made, their book's introduction and one core chapter, their three most-shared posts or essays, a recent talk, or episodes of their own show. This is the highest-yield 40 minutes because it surfaces their real arguments and the lines they care about, which is where original questions come from. You're hunting for a claim they make that you can push on, a story they tell that you can ask for the version they don't usually share, and the one belief they hold that most people in their field don't.
Don't try to consume everything. Read the intro and one chapter, not the whole book. Watch the first ten minutes of a talk and the Q&A, where the unscripted thinking lives. Note exact phrases, quoting a guest's own words back to them ("you wrote that hiring is a hypothesis, what happens when you're wrong?") is the fastest way to signal you did the work and to get a real, unrehearsed answer.
Tier 2: Two past interviews (25 minutes)
Pull up two interviews they've already done, ideally a recent one and one on a show like yours. You're listening at 1.5x for two things: the questions they get asked constantly (so you can skip them) and the moments where they started to say something interesting and the host moved on (so you can go back there). The second category is gold. A half-finished thought from another interview is a question that's already pre-warmed.
This is also where you calibrate. If they've told the same origin story on five shows, asking for it a sixth time is dead air. Hosts who skip this step are the reason guests give canned answers, the guest is on autopilot because the host gave them the same prompt as everyone else. Listening well in the room starts with active listening as a host, and that's far easier when you already know which threads are worth following.
Tier 3: Social and recent activity (15 minutes)
Spend 15 minutes on what they've posted, shared, or been quoted on in the last month or two. This is the tier that makes a conversation feel current rather than archival. A recent post they're clearly fired up about, a project they just launched, a take they're testing in public, these give you a timely opening that no evergreen prep can.
Keep it light and keep it kind. The goal is one or two current threads, not a dossier on their personal life. Scanning recent activity also flags anything sensitive you should know going in, a public dispute, a loss, a topic they've said they're done discussing, which is the start of respecting a guest's boundaries before you record.
Tier 4: Write the one-page brief (10 minutes)
The last 10 minutes turn everything into a single page you can glance at while recording. Without this step, your research stays in 14 browser tabs and none of it reaches the conversation. The brief is the deliverable, not the reading, the page.
Here's the template I use. One page, five rows, no more.
| Section | What goes here | Example |
|---|---|---|
| The one-line frame | Who they are and the angle for this episode | "Hiring expert, but our angle is what she got wrong scaling her own team" |
| 3–5 anchor questions | Open questions built from Tier 1, in rough order | "You call hiring a hypothesis, when has yours been most wrong?" |
| 2 threads to reopen | Half-finished moments from past interviews | "On [other show] you started to say culture decks are theater, finish that" |
| 1 current thread | Something live from the last month | "You launched X last week, what surprised you in the first 48 hours?" |
| Do-not-ask / sensitive | Questions they're sick of, plus any off-limits topics | "Skip the origin story. Don't raise the [public dispute] unless they do." |
The brief is bullets, never a script. You're not reading it line by line, you're using the anchor questions as a spine and letting the conversation breathe between them. A good interview is a guided improvisation, and this page is the guide. If you want the page emailed to you pre-formatted, the kit below has it.
Common mistakes that waste your 90 minutes
Starting with Wikipedia. Third-party summaries give you facts, not questions. Use them at the end to confirm a name's spelling or a date, never as the foundation. Building an interview on a profile page is how you arrive with the questions everyone else asked.
Researching to impress, not to question. Memorizing trivia so you can show the guest you did homework is the wrong goal. Name-dropping their dog from a 2019 episode reads as trying too hard. The research is for the questions, not for proving you read it.
Skipping the past-interview step. This is the most-skipped tier and it's the one that prevents canned answers. If you don't know what they've been asked a hundred times, you'll ask it again, and they'll give you the autopilot version. Twenty-five minutes here saves the whole conversation.
Over-researching and writing a script. More tabs is not more prepared. Past a point, extra reading produces trivia and anxiety, not better questions. Cap it at 90 minutes, write the page, and trust it. A script makes you stop listening, which is the opposite of what good host etiquette on mic asks of you.
Not noting the sensitive stuff. If you don't flag what's off-limits going in, you'll find out live, on camera, when you accidentally raise it. The do-not-ask row exists for exactly this.
Frequently asked questions
How long should you spend researching a podcast guest? About 90 minutes for most guests: 40 on their own work, 25 on two past interviews, 15 on recent social and news, and 10 writing a one-page brief. Spend more for a high-profile guest with a deep body of work, less for a peer you already know. The order matters more than the total.
What should I research about a podcast guest first? Their own work, book, essays, talks, or episodes they've made, before anything written about them. Their own material is where the claims, stories, and beliefs that produce original questions live. Save Wikipedia and LinkedIn for the end, to confirm facts and spelling, not to build the interview on.
How do I avoid asking the same questions every other host asks? Listen to two of their past interviews and note the questions they clearly get constantly, then strike those from your list. Replace them with questions built from their own work and from half-finished thoughts you heard them start in other interviews. That's how you get answers they haven't rehearsed.
Should I send the guest my questions in advance? Share the general topics and themes if they ask, most guests appreciate it. Don't send a word-for-word question list, because it nudges them to script answers and stiffens the conversation. Themes plus your own anchor questions on a private brief is the right balance of structure and room to improvise.
What if I can't find much about the guest? A lighter footprint changes the research, not the method. Lean harder on their own work, even if it's just a handful of posts, and use a pre-interview chat to fill gaps. With a less-public guest you also have more genuinely new ground to cover, which is an advantage, start by building rapport before you record so they open up on mic.