Informational interview research fails for a reason that has nothing to do with the quality of the conversations. The conversations are usually excellent. Someone who has spent a decade in a field will, given twenty five minutes and a genuine question, tell you things that no job description contains. The failure happens afterwards, when nine such conversations have to be turned into one decision and there is no record capable of supporting the comparison.
What survives instead is an impression. The third person was encouraging. The seventh sounded tired. Those impressions are close to worthless as decision inputs, and they are what most career changes actually run on.
An informational interview is an informal conversation with someone working in a field you are considering, held to gather information rather than to ask for a job. UC Berkeley’s career service describes it as an informal conversation, usually lasting 20 to 30 minutes, and is explicit that it is not a job interview. The distinction matters practically as well as socially. Because nothing is being asked for, people answer candidly, and candour is the whole value of the material.
It also means the person is doing you a favour on their own time. Every handling decision that follows should be read against that fact.
This is the part that gets skipped, and it is the whole method.
A single informational interview produces one person’s account of a field, shaped by their employer, their seniority, their tolerance for the work, and whatever happened to them that morning. Taken alone it is an anecdote. Someone says the hours are manageable and someone else says they are brutal, and both are describing the same industry accurately from where they sit.
The information only becomes usable at eight or ten conversations, and only if the same questions were asked each time. No single answer carries the weight. The distribution does. Seven people out of nine naming the same unglamorous task as the bulk of the job is a finding, where one person naming it is a remark.
So the unit of analysis turns out to be the question rather than the interview, read across every conversation at once. Almost nobody has notes structured that way.
Two reasons, and they compound.
Writing during the conversation degrades the conversation. This is a live conversation with a stranger who is being generous with their time, and looking down at a notebook to capture a phrase means missing the next three sentences and signalling inattention while doing it. Most people resolve this correctly, by staying present and writing afterwards.
Writing afterwards degrades the notes. Ten minutes after a twenty five minute conversation, what remains is the shape of it plus two or three vivid fragments. The specifics have already gone: the exact figure quoted for a salary band, the name of the certification that apparently matters, the phrase used to describe why people leave. Those specifics are the comparable data. The shape is not.
So the choice appears to be between a good conversation with bad notes and a bad conversation with good notes, and neither produces a decision.
There is a third option, and it comes with an obligation attached.
Ask. Before the conversation begins, say that you would like to record it so that you can pay attention rather than take notes, that the recording is only for your own reference, and that you will not share it. Then wait for an actual answer.
Some people will say no. That is a complete answer and it ends the matter. Do not record, take notes afterwards as usual, and treat that conversation as a lower resolution data point. Recording someone who has declined, or who was never asked, is not a grey area. It converts a favour into a breach, and in a field small enough that you are interviewing nine people in it, that is also a practical catastrophe.
Most people say yes, particularly when the reason is framed as wanting to listen properly. Some will ask what happens to the file afterwards, which is a reasonable question and deserves a specific answer rather than a reassuring one.
Assume six of nine conversations were recorded. What now exists is roughly three hours of audio in six files, and the naive approach is to transcribe each one and read six transcripts.
That approach reproduces the original problem in text form. Reading six transcripts one after another still leaves the reader holding six separate accounts in working memory and trying to compare them, and holding six accounts in working memory is the task that failed the first time.
What is needed is the ability to ask one question of the whole set. What did each person say about the first two years. Which of them mentioned certifications, and did any of them say the certification was unnecessary. Where did the accounts diverge most.
This is the one job in this workflow that a transcript alone does not do. Vomo voice memos keeps each conversation as its own record while making the set searchable and questionable across files, with speaker labels so that your questions are separable from their answers, and summaries per conversation for the pass where you are reminding yourself which was which. Output can be taken as text, DOCX, PDF or Markdown when the comparison needs to leave the tool.
The mechanical work of comparison stops being the bottleneck. The judgement remains yours, which is as it should be, since the judgement is the part you are actually qualified to make.
The comparison only works if the inputs are comparable, so the question set has to be fixed before the first conversation rather than evolving across them.
Six or seven questions is the practical ceiling for a twenty five minute conversation. They should be answerable concretely: what does a typical week actually contain, what proportion of the work is the part people complain about, what did the path in look like, what would make someone leave, what is genuinely learnable on the job and what has to be brought.
Leave room at the end for whatever the person wants to raise. That unstructured segment is frequently the most useful minute of the conversation, and it is also the one that varies most across interviews, which makes it worth having in a searchable form.
Resist improving the questions midway. A better question asked to interviews six through nine gives you four data points and destroys the comparison with the first five. Write the improved question down and use it next time.
The question about what happens to the recording deserves a policy rather than an improvised answer, and having one ready is part of asking properly.
A reasonable policy has four elements. The recording is for your own reference only. It is not shared with anyone, including a partner or a coach or a group chat. It is not quoted publicly with the person’s name attached, and if something from it is worth repeating, it is repeated unattributed or with permission asked again. And it is deleted once the decision has been made, because keeping a personal archive of nine people’s candid career opinions indefinitely is not something any of them agreed to.
Say the deletion part out loud when you ask. It changes the reception of the request noticeably, because it converts an open ended permission into a bounded one, and bounded permissions are much easier to grant.
If the person is currently employed at a company you might apply to, be more careful still. Their candour about their employer is being extended to you personally, and it becomes a liability for them the moment it exists anywhere they did not intend. In that situation the conservative choice is to take notes afterwards and not record at all, even if permission would probably have been given.
At the end of nine conversations you should be holding a small number of statements you can support, rather than a feeling about the field.
Seven of nine described the same task as occupying most of the week. Four raised the same structural complaint without prompting. Two independently named the same qualification as the thing that opened doors, and one specifically said it did not. Nobody mentioned the aspect of the work that had, until then, been the main reason for considering the change at all.
That last kind of finding is the one worth the whole exercise, and it is only visible when the answers sit side by side. It is also, reliably, the finding that impressions and memory never surface, because a thing that nobody mentioned leaves no trace in recollection.
A career decision made this way can still turn out badly. But it will have been made on what people actually said, which is a different position from having made it on what you remember them having meant.
P.S. Before you zip off to your next Internet pit stop, check out these 2 game changers below - that could dramatically upscale your life.
1. Check Out My Book On Enjoying A Well-Lived Life: It’s called "Your To Die For Life: How to Maximize Joy and Minimize Regret Before Your Time Runs Out." Think of it as your life’s manual to cranking up the volume on joy, meaning, and connection. Learn more here.
2. Life Review Therapy - What if you could get a clear picture of where you are versus where you want to be, and find out exactly why you’re not there yet? That’s what Life Review Therapy is all about.. If you’re serious about transforming your life, let’s talk. Learn more HERE.
Think about subscribing for free weekly tools here.
No SPAM, ever! Read the Privacy Policy for more information.
