The problem with how most analysts process primary research
Most analysts read an expert call transcript the same way they read a news article: sequentially, once, and with a highlighter. They finish with a set of underlines, a vague sense that the call was useful or not useful, and then they move on to the next item in their queue.
This is the least effective way to extract value from primary research. The transcript is raw material, not a finished product. What you do with it in the 90 minutes after reading determines whether it becomes a 3x position or a footnote in a post-mortem.
This piece describes a workflow that converts a transcript — or a set of transcripts — into a research deliverable that is actually useful at the IC table. It is based on how the best-performing research teams we work with operate, abstracted into a repeatable process.
Step 1: Before you read — write down your three questions
Before opening the transcript, write down the three questions you most need answered. Not five. Not ten. Three.
This constraint matters because it forces you to be explicit about your thesis and its vulnerabilities before you encounter the expert's framing. If you go in with no specific questions, you will adopt the expert's framing by default. That's not necessarily wrong, but it means you're letting the source define the research agenda rather than using the source to stress-test your own.
Your three questions might be: Does channel pricing support the gross margin expansion assumption in the model? Is the new product getting real traction with enterprise buyers, or is it a sales conversation piece? Has the competitive dynamic with the number two player shifted in the last 12 months?
Write them down. Keep them visible while you read.
Step 2: Read with structure, not with a highlighter
The first read should be fast — 30 minutes for a typical 60-minute transcript. You're not trying to absorb everything. You're trying to locate the passages most relevant to your three questions, and to get a gestalt sense of the expert's credibility, level of access, and potential biases.
On the second read — which should be slower and more deliberate — annotate with three categories: Confirms (evidence that supports your existing thesis), Challenges (evidence that cuts against it), and New (information you didn't have before that is directionally significant). Don't try to synthesise as you go. Just categorise.
Step 3: The 10-minute gap
After the second read, close the transcript and wait 10 minutes before you write anything. Do something else. This sounds trivial. It is not.
The gap allows your working memory to consolidate what it absorbed and surface the two or three most significant insights without the noise of everything else. The things that come to mind first in that 10-minute gap are almost always the things that matter most. Write them down immediately when the gap ends.
Step 4: Synthesis before summary
The most common mistake in translating primary research into a deliverable is writing a summary when you should be writing a synthesis.
A summary answers: what did the expert say? A synthesis answers: what does what the expert said mean for my thesis?
The IC memo doesn't need a bullet-pointed recap of the transcript. It needs three to five sentences that explicitly connect the expert's perspective to the investment decision. What did you learn? How does it change your conviction level? What's the specific implication for how you're sizing the position?
Write those sentences first. Then add the supporting evidence from the transcript if it's warranted.
Step 5: Cite properly
When using transcript-derived insights in an IC memo or research note, use the format: Expert call, [Sector], via Transcript-IQ, [Date]. This provides sufficient attribution for audit trails while maintaining expert anonymity.
More importantly, it is honest. It tells the reader exactly what kind of evidence you're citing — a single practitioner's perspective, not a published study or a management statement. That context matters for how the reader should weight the evidence.
The multiplier: working across multiple transcripts
The workflow above is for a single transcript. The most powerful version of primary research synthesis involves working across three to five transcripts on the same question — looking for where independent accounts converge and where they diverge.
Convergence builds conviction. A pricing pressure observation made independently by a former VP of Sales, a former channel partner executive, and a former customer success lead carries an entirely different epistemic weight than the same observation made by a single source.
Divergence is even more interesting. When two credible sources give conflicting accounts of the same reality, the explanation for the divergence is usually where the real insight lives. Different tenure, different function, different geography, different competitive context — the reason two people remember the same company differently often tells you more about how the business actually works than either account alone.
Build this multi-source synthesis discipline into your research workflow and your IC presentations will be qualitatively different. Not because you have more data, but because you've processed the data you have with more rigour.