How to research a company before applying, not after

It's Friday afternoon. You get the email. A recruiter wants to schedule a screen for the Senior Engineering Manager role you applied to three weeks ago. You read the email twice. Then you go back to the job posting and try to remember which version of "we're hybrid-flexible with collaboration days" this one was. You open Glassdoor. You open the company's careers page. You start the research now, the night before a 30-minute call, because that's when most people do it.

You already paid the application cost. The 45 minutes you spent tailoring the resume, the cover letter, the screening questions. That cost is sunk. You spent it on a company you hadn't decided you wanted to work for.

That's the move worth fixing. Not the research itself. The order.

A different way to think about the research

If you've read the pillar piece on building a job search that compounds, you know the four artifacts that should hold the real knowledge of your search. Company dossiers are the one that breaks the pattern. The others compound across applications. A company dossier compounds inside a single application, four times: when you decide whether to apply, when you tailor the resume, when you take the interview, and when you negotiate.

That asymmetry is the whole game. Research that happens after an interview invite has already lost two of the four payouts. Research that happens before the application captures all four.

What changes is not how much you research. What changes is when, and what you're trying to decide.

What you are actually trying to decide

Three things, in this order:

Does the role exist as advertised? The posting is the first artifact. Most red flags are already in it. Vague scope ("you'll work on exciting problems"), shifting seniority ("Senior/Staff/Principal welcome"), and team size that contradicts the engineering org chart on LinkedIn are all the company telling you something. So is a posting that has been open for nine months.

Does the company match the posting? Look at recent news, funding state, layoffs, and the careers page. A company that just cut 20% of staff and is now "aggressively hiring" is telling you something. So is a company whose careers page describes a culture that contradicts everything in the role description.

For remote or hybrid roles, is the work model real? "Remote" is the most abused word in tech hiring. The same label covers fully distributed companies and city-first offices that allow Wednesdays from home. This is where most jobseekers get burned, and it's the part most generic research guides skip entirely.

Each of these questions has an answer the company has already published somewhere. Your job is to find it before you spend 45 minutes writing them a cover letter.

The numbers worth holding in your head

Two anchors that change how you read everything else.

The first is from Ashby's 2025 Talent Trends Report, based on first-party hiring data from 1,200+ venture-backed startups: 44% of startups have almost all remote positions, and 35% have no remote options. The market is bifurcating; only about a fifth sit anywhere in between. When a posting says "remote-friendly," the odds are the company is in one of those two camps and the posting is hiding which.

The second is from MIT Sloan's 2022 research on attrition predictors: toxic culture is 10.4x more powerful than compensation in predicting whether employees leave. Read that twice. Salary is not what holds people. Culture is. Which means the research questions that feel softest, what is it like to work here, how do people get along, what does leadership actually value, are the ones with the highest hit rate on whether you'll still be there in eighteen months.

Research catches both. The advertised work model versus the actual one. The advertised culture versus the actual one. The application form does not.

Where to look

Most general guides will give you a list of seven places. The list is roughly right. What matters more is the filter you apply to each.

• The job posting itself. Read it three times. Underline the words it repeats. The repeats are the signal.
• The company's careers page and recent news. Funding round, layoffs, leadership changes, product direction. The dates matter. A "we're hiring aggressively" page next to a "we restructured for efficiency" press release is the company telling you two things at once.
• Glassdoor, Blind, Reddit. Look for repeated patterns, not individual angry reviews. Six people describing the same failure mode in different words is evidence. One is noise. The most useful Glassdoor reviews are not the one-stars; they are the three-stars from people who clearly wanted to like the place.
• LinkedIn. Where employees in the relevant function actually live (city distribution), how long people in similar roles stay (median tenure), and whether leadership is concentrated in one office (the structural answer to whether remote is real).
• The GitLab All-Remote Handbook. Public, opinionated, and useful as a checklist for evaluating any remote role specifically.

The repeated-patterns filter is what makes review sites usable. Most jobseekers reading Glassdoor either dismiss the whole site as cranks, or they spiral on one terrifying review. Both reads miss the signal. The right read is: out of the last fifty reviews, how many describe the same problem in different words? That number is the signal.

The remote question deserves its own pass

If you are applying for remote or hybrid roles, the work-model verification step is not optional. Most generic research guides skip it because they were written for any role at any company. Yours is not any role.

A useful filter, the five-tier work model framework Remotivated maintains, makes the distinction explicit: Fully Remote, Remote-First, Flexible Hybrid, Structured Hybrid, Onsite. "Remote-friendly" is not on that list because it does not mean anything in practice.

Before you apply, you should know which of those five the company actually is, not which of those five the posting wants you to think it is. The fastest way to figure this out: check whether the company's executive team and leadership chain operates the same way as the role you'd take. If the engineering VPs are all in San Francisco and the posting says "remote anywhere," the answer is probably "Flexible Hybrid pretending to be Fully Remote."

This is also the bridge to the company directory at remotivated.com/companies, where the work model and REMOTE Score for each profile are already classified. The point is not that you outsource the research. The point is that the structural question, what is the actual work model, is settled before you spend energy on the qualitative ones.

That dossier carries through. When the screen lands, it becomes the structure your interview questions come from, and it sets the salary anchor when the offer call arrives.

Weighing the evidence: a three-verdict frame

The signals you've collected do not tally mechanically. A single hard one, like a return-to-office mandate buried in a Glassdoor review from last month for a posting that still says "remote anywhere," outweighs a pile of small concerns. Severity matters more than count.

Force the dossier into one of three verdicts before you decide whether to spend the 45 minutes:

The middle row is where most of the work lives. "Proceed with caution" is not a soft yes. It is a conditional yes, with a list of questions the company has to be able to answer in the first call, or the dossier flips to skip.

Questions to ask if you proceed

These are the questions the dossier earns you. Good companies can answer them directly. Weak answers are data.

• "What problem made this role necessary now?"
• "What would success look like after six months?"
• "How is the team distributed today?"
• "Which decisions happen async, and which require meetings?"
• "What changed after the last reorg, funding round, or layoff?"
• "How do remote employees stay visible for promotion and high-impact work?"
• "What is the hardest part of working here right now?"

If the recruiter or hiring manager handles three of these well and ducks four, that's the dossier doing its job: surfacing the gap between the advertised role and the lived one before you commit further.

What this looks like inside the toolkit

The company-research skill in the open-source job-hunt-skills repo runs this loop as a pre-application pass. It pulls the posting, the careers page, the latest news, and a Glassdoor pattern read, then produces a one-page dossier you reuse through tailoring, the interview, and the offer. The reason it's split into its own skill, separate from resume-tailor, is that the decision to apply should be made on its own, before you start writing copy that argues for the application.

The smaller game and the larger one

Most company-research advice treats the dossier as interview prep. A list of three smart questions to ask. A few facts to drop in the small-talk window. A way to seem prepared.

That is the smaller game. The larger game is using research as the cheapest decision tool in the search. Two minutes on the posting, ten on the careers page and recent news, ten on Glassdoor patterns, five on LinkedIn structure. Twenty-seven minutes total. Spent before the 45-minute application.

Most jobseekers spend the 45 minutes first, then do the 27 minutes the night before the interview. Reverse the order. Almost nothing else changes about your search and the unit economics tilt back in your favor.

Research once. Use it through offer day. Decide first.

The full thinking behind why research compounds inside a single application rather than across the search: Nothing earned is lost: the case for a job search that compounds. The open-source toolkit that runs this loop with whatever LLM you already pay for: github.com/Remotivated/job-hunt-skills.

Sources

• Ashby Talent Trends Report 2025: first-party hiring data across 1,200+ venture-backed startups, 32K hires, and 11M applications; remote-vs-onsite bifurcation (44% / 35%). ashbyhq.com
• MIT Sloan Management Review (Sull, Sull, Zweig): "Toxic Culture Is Driving the Great Resignation," Jan 2022; toxic culture is 10.4x more powerful than compensation in predicting attrition. sloanreview.mit.edu
• GitLab All-Remote Handbook: public documentation on async-first practices, including a checklist of questions for evaluating any remote role. handbook.gitlab.com