Common Coffee Chat Mistakes in IB Recruiting (and How to Fix Each One)

I once emailed more than 100 employees at PJT in a single week. I got zero responses.

Not a handful. Not one. Zero.

What stung most was sitting right next to me. A friend was going after the same firm with an objectively worse résumé, and he got through to a handful of analysts. The only thing he did differently was pace himself: two emails a week to PJT's New York office, patiently, over time. I carpet-bombed the entire firm in seven days. I sent my follow-ups too. Still nothing. There are other reasons a firm might go quiet, but when the weaker candidate gets the chats and you get silence, the lesson is hard to dodge. I hadn't just failed to stand out. I'd rubbed a group of people at PJT the wrong way, and no email was going to undo it.

That's what makes coffee chat mistakes so dangerous: almost none of them announce themselves. Nobody writes back to tell you that you came across as entitled, or that you sounded like you were reading off a script. You just don't hear anything, and you're left to guess. Over my own recruiting I did more than 300 coffee chats, I've since listened to over 150 of my students' chats, and I've been on the other side of the table, coffee-chatted by students, more than 50 times. The same avoidable errors come up again and again. This guide is a catalog of them, walked through one at a time, each paired with its fix.

First, one piece of context, because none of these mistakes land without it. A coffee chat exists to earn you a referral, and that referral is your ticket to the first-round interview. But you almost never find out whether you "got" the referral until invitations go out. The entire game is played in the dark, which is exactly why small, controllable mistakes decide so much. The students who win aren't the ones chasing a magic line that converts a chat into a referral. There isn't one. They're the ones who are consistent, who avoid the unforced errors below in every single chat.

One more number, to set the stakes. Take 100 finance students. About 80 will say they want to do investment banking. Around 50 will actually start networking. Of those, maybe 30 sustain it across two semesters, roughly eight months. And only about 15 track their outreach methodically enough to know who to follow up with and when. That last group is who proceeds to first rounds. Most students don't lose to better candidates. They lose to avoidable mistakes, and then they quit.

Let's make sure you're not one of them.

The Meta-Mistake: Playing Recruitment Reactively

Almost every other mistake in this guide is downstream of this one, so start here.

Most students play recruitment reactively. They see a posting on a job board, tweak their cover letter, fire off a few outreach messages to people on the team, and call it networking. Plenty only do the first two. By the time that outreach lands, it's one of hundreds of near-identical messages the recipient has gotten since the posting went live. You look average, or your message gets lost, and either way, reaching out a few days before a deadline is a near-guaranteed way to not get an interview.

There's an old line from Sun Tzu: "Every battle is won before it is fought." The battle here is for the offer, and you won't win it if you don't stand out. Reacting to a posting is the opposite of standing out, because by definition you're doing the same thing, at the same time, as everyone else.

The fix is a mindset shift. Stop treating recruitment as a series of deadlines to respond to, and start treating it as relationships you build before the deadline exists. The student who reaches out months early, with no posting in sight, doesn't look like one of the desperate hundreds. They look like an asset: organized, prepared, and more serious about their career than 99% of their peers. Everything below is just the specific, concrete ways students fail to make this shift.

Outreach Mistake #1: Reaching Out Too Late

Ask a student why they can't network eight months before an interview process, and the honest answer is usually some version of: "because that's too early." It's a wonderful non-explanation. "I can't network early because it's too early."

The real reason it feels too early is the email you're picturing. You're imagining a message that's overtly built around a hiring cycle, the kind that mentions the SA'26 process, says you're "interested in learning about potential internship opportunities," and that you'd "love to learn about the firm." Think of the classic version, the sort sent during penultimate-summer networking.

It leads with the sender's own summer seat, announces they're "very interested in pursuing" a specific internship cycle at the recipient's firm, and asks for a call about "potential internship opportunities" – every line pointing back at the sender's agenda. I take that exact email apart line by line in how to cold email a banker.

Sent well in advance of the cycle, that email won't land. All networking is transactional, but this is so overtly transactional that it works against you. Of course it feels wrong to send eight months early. It's about a job that doesn't have a posting yet.

So don't send that email. Make it about them instead. Build your cold outreach around the recipient, their background, and the specific things you genuinely find interesting about both. Do that and two things happen. The recipient is pleasantly surprised to get a student email with no mention of a job, so you stand out. And they're far more likely to actually take the call (assuming your résumé clears the bar for that firm). Here's the same outreach, reframed:

Hey Bob,

Hope your week's going well!

My name is… pursuing a career in Investment Banking.

I noticed that you were part of the deal team for Cisco's acquisition of Splunk and on the only Venture Capital club at your college.

After looking through the resources you authored for your club I immediately shared all of them with my team. They're stellar and better than anything else I've found online.

As someone who's worked at an incubator before and is deeply interested in M&A, I thought you'd be a great person to reach out to in order to learn more about tech M&A and what it was like to diligence VC investments through that club.

Would you be open to a quick phone call sometime this week or next?

No cycle. No "internship opportunities." Just a real reason you're reaching out to this specific person. It takes more than a couple of seconds to write, which is exactly why most students won't do it, and exactly why it separates you.

Reaching out early does something else, too. It buys you time to nurture the relationship. Have the chat, then send a short nurture email every three months or so just to stay on their radar. Six months later, having never once asked for a job-related favor, you've banked enough goodwill that this person becomes an advocate for your first-round interview. Compare that to emailing them the day the posting drops.

And the cost of waiting compounds. If two months ago 50 students were networking this firm, now there are 200, next month 400, and by the time the posting hits there are thousands. The later you start, the less you stand out, until you don't at all. Worse, the late networker looks like one of two unflattering things: a desperate student angling for a résumé push, or someone with no real interest in the role who just wants an informational chat. If your résumé is average, both roads lead to roughly zero shot at a competitive first round.

Outreach Mistake #2: Blasting an Entire Firm in One Week

My PJT story above is the version that cost me nothing but silence. Friends of mine got the version with a paper trail. One emailed four BDT-MSD analysts in a single week in November of 2023, for the SA'25 cycle. Two days later an analyst wrote back to tell him people in the office had talked and were displeased to see a student emailing four of them at once with near-identical messages. He didn't get a first round. Another friend emailed five Ducera Partners employees in one week during the SA'23 cycle and quickly got a very angry email from a Director telling him he wasn't "allowed" to do that.

The fix is a rule of thumb: aim to speak with two to five people at a given firm, and get there slowly. Better still, get there organically. The ideal isn't five cold emails to five strangers in the same week. It's one good chat with the first person, who refers you to the second, who refers you to the third, and so on. That first person only does that if they think you're a winning horse, which is why everything earlier in this guide matters. Spaced out and referral-driven beats simultaneous and cold, every time. My friend who beat me to PJT proved it with a worse résumé and two emails a week.

Targeting Mistake: Starting at the Top Instead of With Analysts

When students do mass-email a firm, they often aim too high while they're at it: straight to the MDs, the seniors, sometimes HR. The logic is seductive. "Who says no to an MD? HR? Bah." I've done it too. It's still a mistake.

For the larger firms most readers are chasing, start at the very bottom: analysts, or even incoming and past interns. There are concrete reasons.

Analysts are closest to you in age, which makes them the easiest conversations you'll have, and the most generous. To a senior banker, an undergraduate is nowhere near an asset, so they tend to answer curtly and try relatively little in a chat. To an analyst, you might be an asset, so they're more forthcoming with information and more forgiving of gaps in your understanding. They also recruited successfully for that exact firm very recently, which makes their interview answers a goldmine: copy what worked. And they hold the kind of insider information you can't get from the outside, like which analyst runs recruiting, who in HR is running point, or which MD is the gatekeeper for first rounds.

Analysts are also, on their own, usually enough for a referral. Most firms require the backing of an analyst or above, and an analyst's word typically clears that bar. They're also far more likely to reply to a student's cold email than anyone senior. So if you're short on time, a single analyst chat is your best possible use of it.

None of this means you never talk to seniors. Their word carries more weight than any analyst's. It means you talk to them after the juniors, and there's a tactical payoff for doing it in that order. When you finally email a senior, you can name the people you've already spoken to, which does two things. It justifies why you're contacting them specifically, "for a different perspective," which makes it harder for them to simply forward you back down the hierarchy. And it's a quiet display of work ethic. Picture an MD with three student emails: one mentions no prior contacts, one mentions a single chat, one mentions five. Even setting résumés aside, you know which one gets the reply. Here's what that email looks like:

The fix is an email that name-checks the four juniors you've already spoken with before asking the MD for anything – it reads as work ethic instead of flattery. The template, line by line, lives in the who-to-network playbook.

HR deserves the same patience. Some firms lean on it heavily. At Bank of America, for instance, the first-round interviews are group coffee chats run by an HR employee who evaluates you from start to finish. That's a stark contrast to a firm like Moelis, whose first rounds are always carried out by analysts and associates. It's easy to get a chat with HR, but impressing them to the point of a referral is another story, so treat them like the seniors: reach out only after you've spoken to a handful of analysts or associates first.

The Introduction Mistake: Winging Your "Tell Me About Yourself"

The TMAY exchange happens in the first two to five minutes of every coffee chat, and it's your single best opportunity to prove you're a colleague. Wing it, and you can eliminate your shot at a referral before the real conversation even starts.

Here's why the stakes are so lopsided on this one. Your introduction is the lens through which the other person views you for the entire rest of the chat. There are only two stretches of a coffee chat where you have free rein to say what you want: the beginning and the end. The end comes after your referral is effectively decided, so really only the beginning counts, and the beginning is your introduction. Deliver it poorly and you put a sour taste in their mouth: "This student can't even introduce himself. This is going to be a long thirty minutes."

Because your TMAY is a constant across every chat you do, it has to be perfect. Other students will deliver theirs flawlessly, some with worse résumés than yours, so a clumsy intro looks even worse by comparison. From the banker's side, the fact that you can't nail the one thing you repeat in every chat signals either that you aren't doing enough chats or that you aren't taking this seriously. Neither makes them want to work with you.

The Cardinal In-Chat Mistake: Running a Q&A, Not a Conversation

You've heard "have a conversation, not a Q&A" before, and you've probably found it impossible to actually act on. I never got a straight definition of the difference either, so here's a simple one. Measure the talking time after the small talk and TMAY are done. If they're talking more than 80% of the time, you're running a Q&A. If the split is closer to 60–70% them and 30–40% you, you're having a conversation. Record your next chat and time it. You won't fix the problem until you come face to face with it, and I'd wager a lot of you are running Q&As, consistently.

Here's the move. Instead of jumping straight to your next question, first demonstrate that you understood their point. If they talked for two minutes, summarize it back in under 25% of that, around 30 seconds. Then do one of two things: relate it to a past experience of yours and use that as the segue into your next question, or name the specific point you want to clarify and walk them through your thinking before you ask. Those are the two flows of a strong coffee chat response:

Their (2-minute) Answer => Your (~30 second) Summary => Your Past Experience => Your Next Question

Their (2-minute) Answer => Your (~30 second) Summary => Point to Clarify => Your Thought Process / What You Expect Them To Say => Your Next Question

Let me show you the difference with a real exchange. One of my students was chatting with an associate on Bank of America's NYC investment banking team, who opened with this:

"…have a background in piano performance and was professionally trained as a concert pianist. I'm a 1st-year associate and did my 2 years as an analyst… I'm part of the Technology team here in NYC. … when I started as an intern I was exposed to the culture at BofA, it's a great culture that provided a great learning experience, and then as a full-timer I came in and had a ton more responsibility… The more reps I got on deals the better I got… In a nutshell, it's been great, the people are amazing and very generous with their time… So yeah, I'll pause there in case you have any questions."

Here's what the student did with it.

He summarized her introduction back to her, brought up his own childhood piano training – Certificate of Merit, advanced level – and only then asked a question, and even that one was about her pianist years rather than the job. I walk through the full exchange in my breakdown of how to stand out in a coffee chat.

See the structure. He summarized her introduction so she felt heard, related a genuine experience of his own, and then segued into a question. The moment he paused to double-click on the piano detail, that associate knew she wasn't talking to an average student. Now compare it to the version most students would give:

"Got it, thank you for the background. What was it like to be a professional pianist?"

It's not disqualifying. But by skipping the summary and omitting his own piano experience, that student falls behind anyone who included them. Same question, a fraction of the rapport.

The hard part isn't any one of these steps. It's doing all of them at once: listening, building your response, finding a relevant past experience, and organizing your thoughts in real time. One tip that helps enormously: the instant a question pops into your head, write it down in your notes. As someone talks there's a whirlwind in your brain, and a good question will vanish into it if you don't capture it. Don't stop to judge whether it's any good. Do that later, as they're wrapping up their answer.

Don't ask yes/no questions

Nothing turns a chat into a Q&A faster than a yes/no question. It breaks the flow, it positions you as someone who didn't do their research, and it often just irritates the other person, because most yes/no questions are things you could have answered yourself online or with a few seconds in ChatGPT. If you can find it online, why are you asking a banker?

The fix is to make every question seek an opinion. The genuinely useful questions are the ones where a piece of information is missing from what's online: their unique perspective, an applied judgment, the "what it's actually like." When you ask one of those, show your work. Walk them through the research you did or what you expect their answer to be, then ask. If you don't show the research, they'll assume you did none. The only time a yes/no question is fine is as a quick framing step before a much larger, more thoughtful question.

Follow a structure, but don't read from a script

Two opposite mistakes live here. The first is having no structure at all, hopping between unrelated questions, which makes you look disorganized and makes the other person feel unheard. The second is the overcorrection: treating your prepared question list as a bible and marching through it regardless of what they say. That's a fast track right back to Q&A-land, because in about 90% of chats the conversation goes somewhere you didn't predict, and a student welded to a script can't follow.

The structure to follow is dynamic, and it's hiding in plain sight: it's the chronology of their own introduction. When they introduce themselves, they walk a timeline, with chapters like pre-college, college, first analyst years, and current firm. Note those chapters and use them to bucket your questions, moving roughly in their order. It's subtle enough that they won't consciously notice you're following a structure. What they will notice is the absence of confusion, the sense that this student came in organized and is genuinely tracking what I say. Your prepared questions exist to take the load off forming questions on the spot, freeing your brain to craft better responses. They're an outline, not a cage.

Tone Mistakes: Arrogance, Bad Timing, and Being "Too Polite"

A handful of tone errors quietly erode the rapport you're working to build.

Speaking in absolutes

The fastest way to sound like an arrogant know-it-all is to describe the other person's world back to them in absolutes, as if you understand their job better than they do. It usually looks like this:

"I saw you moved from Equity Research over to Investment Banking, and I'm sure that was the plan all along since ER is really just a stepping stone for most people, and you must be relieved to finally be working on live deals instead of just writing about them. Did you ever think about going to a Hedge Fund instead?"

The problem is that it leaves no room for any alternative, and frames any alternative as abnormal. The tone is prescriptive ("you should have done this") where it should be inquisitive ("you did this, here's why I think you did, but I'd love your perspective"). Stay humble. The people you're talking to are ahead of you. Treat them that way.

The right question at the wrong moment

You also kill rapport by asking a question that's transparently designed to show off your knowledge. Trust me, it's always obvious. Here's a deliberately exaggerated version to make the point:

"Given you're in the Energy group at Bank of Zimbabwe, what do you think of the recent spike in Crude Oil prices motivated by the regulatory changes recently enforced by the 5 major African oil producers, especially seeing as OPEC has expressed intentions to halt exports to the Western hemisphere after Bloomberg published that piece on the ever-worsening US-China trade war, have you seen that piece?"

No one reading this is showing off quite that egregiously, but you can probably hear traces of your own convoluted questions in it. The interesting thing is that the question itself isn't the problem. A genuinely sophisticated question about an industry or market event is exactly what signals how much of a colleague you are. The issue is timing. Think of the restaurant cliché, "location, location, location." In a coffee chat, the location of a question is when you ask it, specifically how related it is to what they just said. A great question dropped in unrelated to the current thread reads as showing off. The same question, asked when it flows naturally from their last point, reads as insight. You'll only develop this sense after 10 to 20 chats, so pay down that "ignorance debt" early.

Being too polite

Our instinct as students talking to professionals we've never met is to be overly formal, because we're terrified of saying something rude. Fair enough, but remember you're trying to position yourself as a colleague, and you can't do that while thanking them after every single answer. Thanking someone after each thing they say is, tellingly, a behavior that only exists in Q&As. It sounds disingenuous and it makes the whole chat stiffer than it needs to be.

The right balance: say "thank you" at the start and end of the chat, in reference to the time they're giving you, and thank them in the moment only when they share something genuinely insightful. Offering to call them when they accept is just common courtesy. They're normal people. You're not walking on eggshells.

Repeating the same acknowledgement

This one is low-hanging fruit, but it's so easy to miss and so noticeable when you do it. The common acknowledgements are "got it," "that makes sense," "thank you for that," "okay," "sounds good." Lean on any one of them too often and you sound like you've checked out, even when you're fully engaged and just forgot to vary the word. The fix costs almost nothing: keep about five different acknowledgements written on your screen during the chat, and when you sense yourself repeating, glance up and grab another.

The Follow-Up Mistakes: Going Silent, or Going Overboard

Following up has two failure modes, and they pull in opposite directions.

The first is not following up at all. Most students don't, which is precisely why doing it sets you apart. When your email goes unanswered, your gut says they judged you unworthy, but far more often they're simply slammed, or you slipped off their radar entirely. These are front-office professionals who get a staggering number of emails a day, and a student chat sits at the bottom of their priorities. A follow-up is just a polite reminder. And because tracking your outreach is the only way to know who to remind and when, and only about 15 of every 100 students do that, every single follow-up you send separates you from the pack.

The second failure mode is pestering. The line isn't about how many times you write, it's about how you write. Never imply you're owed a response. The fix is to follow up while stripping out any blame for not replying and any obligation to have replied, and to shrink the ask. Phrases that do this well:

By minimizing the time commitment, you make it easier for them to say yes. Done this way, your persistence reads as diligence, which is what got me five emails deep with that Moelis ED, rather than as the entitlement that got my friend an angry email from a Director.

The Mindset Mistake: Reading Your Odds Off Their Disposition

The last mistake is internal, and it'll quietly wreck your confidence if you let it: trying to decode your odds of a referral from how the other person seemed on the call.

It can't be done, and chasing it is corrosive. I know because I tried. While recruiting, I got interview invitations from firms whose employees were dry and genuinely hard to talk to, and I got ghosted by firms whose employees had all but told me I was in pole position. My read on those people was completely wrong and totally unproductive. There are countless reasons someone behaves the way they do on a call, and putting on a warm face for a student is nowhere near the top of their list. They might be having a bad day. They might be slammed. They might have forgotten to do the thing they said they would.

This is also why I'd steer you away from making "the referral" your stated aim, since you usually can't confirm one until invites go out. That's holding yourself to an invisible, unachievable standard. If you want a rough read, aim instead for a referral to another employee, because connecting you onward is something nearly anyone can do, and their willingness to do it loosely approximates their impression of you. But even there, don't over-read a non-referral. Maybe they were just busy.

Putting It Together

Notice what every mistake in this guide has in common. Not one of them is about being smarter, having a better résumé, or knowing more finance than the next student. They're about the controllables: starting early instead of reactively, pacing your outreach, leading with analysts, nailing your introduction, talking less than they do, staying humble, varying your acknowledgements, following up like a professional, and refusing to read tea leaves. You can lose for reasons outside your control. Don't lose for these.

And if you're sitting here convinced you're hopeless at this, I promise you can learn it. I'm deeply introverted, I spend most of my time alone, and coffee chats genuinely scared me. My early ones had 15-second awkward silences. I got better the only way anyone does, by doing them over and over. It's now my full-time job helping students make exactly that transformation. If I could get there, so can you.

A last practical note. Technicals almost never come up in a coffee chat, and if they do, take it as a good sign, because firms only bother evaluating candidates who already look like they deserve a first round. To be ready anyway, keep your behavioral answers, your scripted deal walkthroughs, your résumé, and a searchable technicals reference on hand. I used ibvine.io (set to "All," in List view, for easy searching); a combined PDF of the BIWS guides works too. That's insurance, not the main event. The main event is everything above.