How to Research a Company Before an Interview (Three-Layer Company Research)
How to research a company before an interview: the business problem, the role's challenge, and the competition, with a fifteen-minute version.
By Alexander Hoff · Founder
Research the same three things you researched before you applied: the business problem, what this particular role is there to fix, and who the company is up against. The difference is what you do with it. Before you apply, the research goes into the documents. Before an interview, it goes into how you answer and, more than anything, into the questions you ask at the end. You're looking for two or three questions nobody else in the process will ask, and for the one fact that changes how you talk about the whole thing.
Searching the phrase itself takes you somewhere else entirely, so a quick note before going further. "Company research" mostly means finance. The CFA Institute describes research analysts building models that end in a buy or sell recommendation, and Morningstar publishes that work as equity research for investors. Market research is different again: studying a whole category and its buyers, the kind of thing Gartner sells to companies. Neither one is what a candidate does the night before an interview.
What is three-layer company research before an interview?
The three layers are the business problem, the role's specific challenge, and the competitive landscape. They're the same layers we walk through in what to actually research about a company, which covers the application stage. Nothing about the layers changes when you get to the interview. What changes is that you now have to say what you found out loud, to someone who works there and knows more than you do. That raises the bar on accuracy a lot and makes volume almost worthless. You're better off being sure of one thing than half remembering six.
What are the three layers, and what changes when you're interviewing?
- What has changed about the business in the last two quarters, and what has stayed the same for years?
- What has the company said in public that it's worried about?
Read every press release from the last two quarters, oldest first, and write one sentence per quarter about where things are heading. If the company is public, pull up the latest earnings call and read only the analyst questions, which are far less rehearsed than the prepared remarks. If it's private, the most recent funding announcement usually says what the money is for. There's a deeper method in how to find a company's real problems.
- Why is this seat open right now? Is it a backfill or new headcount?
- What's the biggest problem this hire is expected to solve?
Open the careers page, filter to the same department, and read every other open role on that team. Three openings in one function usually means a team being built. One senior backfill next to two junior openings usually means someone left and the team is still absorbing it. Then ask the recruiter two questions before the first round: is this a backfill or a new seat, and what should the person in it have changed by the end of the first quarter? Recruiters can usually answer both.
- Who does the company actually lose deals to, and why?
- Who does it partner with? At big companies the same name often shows up on both lists.
Read two pages the company didn't write: a competitor's comparison page that names them, and that competitor's pricing page. Then search the company's own newsroom for partnership announcements from the same period. Bring what you find into the room as a question rather than a conclusion.
How long does this take, and what do I do if I only have fifteen minutes?
One HR technology operator we work with, interviewing for roles outside her own field, put her real budget plainly: "Sometimes it's like 15 minutes before the interview, depending on how much time I have." That's most people, most of the time. So treat the fifteen-minute version as the default and the long version as the exception.
Done by hand, a full pass on one company takes two to three hours. For comparison, across the kits we've analyzed, Telosi's research step reads a median of 184 sources from 99 different websites in about 16 minutes, cites roughly 48 of them in the brief it writes, and the finished cover letter ends up carrying a median of 7 specific facts about the company. That ratio is the useful part. You read a lot more than you keep, and you walk in with about seven things you could say without looking at notes. Telosi's free Research Brief is that whole pass delivered as one document.
| Layer | Fifteen minutes | Two hours |
|---|---|---|
| 1. Business problem | The newsroom, last two quarters, oldest first. One sentence on where things are heading. | Add the analyst questions from the earnings call, or the funding announcement. Write the direction up as a dated paragraph. |
| 2. Role's challenge | Reread the posting once, looking for what problem it implies, not for keywords. | Add the other open roles on the team, the background of the person you'd report to, and the two recruiter questions. |
| 3. Competitive landscape | One competitor comparison page that names them. | Add pricing pages, partnership announcements, and an honest note on where the company is behind. |
My own long version ends the same way every time. I upload the research brief to NotebookLM, generate the audio overview, and listen to it on the way in. Hearing it once does more for me under pressure than rereading it does.
What would I pull from a company page before a Databricks interview?
Telosi publishes a company page for Databricks, where layers one and three arrive already assembled and dated. If the company you're interviewing with doesn't have a page like that, the same half hour comes from the two most recent press releases, a recent interview with an executive, the latest partnership announcement, and the careers page.
Growing fast with the margin question still open
First-round interview in thirty minutes, nothing prepared. Every fact on the page is dated and linked to its source, so the half hour goes into reading instead of checking.
- Layer one. Databricks said in August 2026 that it had passed a $7 billion revenue run-rate, up more than 80% year over year, and closed a $5 billion round at a $190 billion valuation. In June the CEO said gross margin was heading lower, without giving a number. Growth going up while margin goes down is a business problem you can ask about without accusing anyone of anything.
- Layer three. Microsoft and Databricks announced an expanded partnership in July 2026 that both sides describe as running into the 2030s. At the same time, all three clouds that host Databricks also sell their own competing platforms: Microsoft Fabric, AWS SageMaker Lakehouse, and Google BigQuery. So Microsoft is on both lists, partner and competitor.
- Process, which isn't a layer. Databricks publishes its own hiring process: seven stages from finding the role to the offer. Knowing there are seven changes what you ask in the first one.
- Layer two isn't there. No company page knows why your particular seat is open. That still comes from the posting, the other open roles, and the recruiter.
What should I ask at the end of the interview?
This is where the research shows up, or doesn't. One candidate, heading into a third round, named the miss himself: "I did not ask enough questions at the end... I have to do a 30/60/90 impact review... I don't really know their KPIs that they're looking at, what's their onboarding process... I'm kind of taking a little bit of a swing in the dark." Writing down one question per layer before you walk in fixes this.
- "The last two quarters look like they've been about expanding into a new segment. Is that how it feels from inside, or is the real priority somewhere else?"
- "What is the team worried about for next year that wouldn't be obvious to someone reading the press releases?"
- "What would have to be true at the end of ninety days for you to say this hire worked?"
- "Which numbers does this role own outright, and which ones does it only influence?"
- "When you lose a deal, who do you usually lose it to, and what do they have that you don't have yet?"
You usually only need one real finding. One of our users took his research brief into a final round where the format was to give a sales presentation on any product and then talk about the business. The brief couldn't find any group purchasing organization relationships in the company's public record, which surprised him: "I was just shocked by that." He brought it up in the room. Afterward he said, "The most valuable part was what the research uncovered for me that I would not have uncovered myself. It looked at different facets of the business that I assumed they probably already had taken care of, but they didn't." The brief had flagged three gaps, and by his own account only that one was strong. He took the job.
What does company research not fix?
It doesn't rescue a bad thirty minutes. One candidate who had prepared and still came out flat said afterward, "There was a lot of situational questions, and it was about manufacturing... I did not do the way I wanted to... on another day I probably would have done well with that." Knowing the company inside out doesn't help you answer a situational question about a field you've never worked in.
It also doesn't decide the outcome. I once applied cold to a role in a space I hadn't worked in, as one of roughly 1,500 applicants, and made it through eight rounds to the final two, including an hour-long partner strategy presentation and a live pitch to two executives from a partner company. I got that far because I could talk about the situation the company was actually in. Then they hired nobody.
So keep the goal modest. Write one question per layer before the interview starts. Fifteen minutes is enough for that.
What other questions come up about interview research?
How do I research a private company with no earnings calls?
Piece it together from what private companies do publish: founder interviews on podcasts, customer case studies, the paragraph in a funding announcement that explains what the money is for, and job postings across different teams. In the United States, private companies that raise money usually file a Form D with the SEC, which is public on EDGAR and gives you a dated record of what was raised and when. What a company ships and who it hires tells you its priorities more reliably than what it says about itself.
Should I bring the research up, or wait until they ask?
Mostly wait. It belongs in two places: as the framing on a situational answer, and in the questions at the end. Briefing someone on their own business in the first five minutes comes across as a performance, and you risk correcting an insider before you know which parts of your picture are out of date.
Can I use an AI assistant to prepare?
Yes, and plenty of candidates already do. One job seeker described her routine between rounds this way: "I am always updating my perspective on a role, and I'm running everything I know about it through Claude for the next round, to be prepared as much as I can be." Assistants are good at organizing what you've gathered and rehearsing how you'd say it. The catch is that sooner or later one will hand you a confident fact that's wrong, and an interview is an expensive place to find that out. Check anything you plan to say out loud against the company's own site, a filing, or a dated news story.