GTM engineer behavioral interview questions

• 01 Tell me about a time you caught a bug in a GTM system. Whether they expect their systems to break and plan for it

  Follow up with

  • How long had it been running like that before you caught it?
  • What did you implement so you would be notified next time?

  Strong answer

  A strong candidate has one ready and can say how long it had been wrong before anybody noticed, which is usually longer than they are comfortable admitting. Expect them to name what it cost, in records or in a rep's time, and to say what they put in afterwards so the next one surfaces on its own. Some good candidates cannot think of a bug offhand and instead describe the logging and alerting they built so a bad run would announce itself, which is the same answer arriving from the other side.

  Weak answer

  A weak candidate cannot name one at all, which usually means they have not run a system long enough to watch it break. Or they name something that should have been guarded against from the start and stop at the fix, with nothing about how they found out or what they changed so the next one would find them.

• 02 A funnel you own is leaking between acquisition and activation. What are the first three things you do? Whether the first reflex is diagnosis or tooling

  Follow up with

  • What would change your mind about where the leak is?
  • When would you conclude the targeting is wrong rather than the funnel?

  Strong answer

  A strong candidate treats the stated leak as somebody's hypothesis and starts by asking what it is based on. The first move is questions rather than actions: which segment is leaking, whether it has always leaked or started recently, and whether the activation event is instrumented well enough to say. They usually get to whether the people arriving were the right people at all, which moves the problem out of the funnel and into targeting.

  Weak answer

  A weak candidate accepts that the leak sits between acquisition and activation because the question said so, and answers with tools they would add: a session recorder, a warehouse, a dashboard. Nothing in the three moves would tell them the premise was wrong.

• 03 You are handed a cold-email domain that also needs to run the team's day-to-day mail. How do you set it up? Whether deliverability knowledge is operational or repeated

  Follow up with

  • What does the warmup ramp look like week by week?
  • At what daily volume per inbox does placement start to tilt?

  Strong answer

  A strong candidate stops on the premise, because a domain's MX records point at one provider and the setup as described does not exist. They separate the sending domain from the corporate one, usually a lookalike that redirects, and only then talk about authentication and the ramp. Ask for the ramp week by week and they give you weeks and numbers rather than a principle.

  Weak answer

  A weak candidate takes the premise at face value and starts configuring inboxes and warmup on the corporate domain. The conflict is sitting in the question, and reciting a sound warmup ramp on top of it does not make the domain safe to send from.

• 04 Pick the GTM system you are proudest of. Describe it in two sentences, then tell me what was different for the business because it existed. Whether they can state an outcome or only narrate a build

  Follow up with

  • What was the number before, and what was it after?
  • How much of that delta would you defend as yours?

  Strong answer

  A strong candidate gets through the build in roughly the two sentences you asked for and spends the rest of the answer on what changed: pipeline created, reply rate, hours handed back to reps. Ask what the number was before and after and they have both, or they say plainly which one they never had. Expect them to volunteer how much of the change they will claim, because other things were moving at the same time and they know it.

  Weak answer

  A weak candidate is still describing the architecture two minutes in, and the tools and the clever join get more time than anything the business felt. The question asked what was different because the system existed, and the answer never reaches a number that moved.

• 05 What has a sales or marketing leader asked you to automate that you argued against building? Whether they have judgment about what should not be automated

  Follow up with

  • What did you propose instead?
  • How did that conversation end, and who made the call?

  Strong answer

  A strong candidate names the actual request and who made it: a scraped list, a personalisation trick, a routing rule that would have covered up a headcount problem. The argument they made is in business terms, about what it would cost the brand or the pipeline, rather than about the work being unpleasant to build. Expect them to say how it ended, including the times they lost the argument and built it anyway.

  Weak answer

  A weak candidate has never pushed back on anything, which at this level is itself the answer. Or every refusal they describe was a technical impossibility, so none of it was a judgment call they had to make and then defend in front of the person who asked.

• 06 Take one thing you shipped into the revenue funnel. What told you it was working in week one, and what told you in month three? Whether success was defined before the build or after it

  Follow up with

  • What did you instrument before launch, and what did you wish you had?
  • What number would have made you turn it off?

  Strong answer

  A strong candidate gives two different measures, because the question asks about two horizons and the same number cannot serve both: engagement or throughput in week one, pipeline or retention by month three. They can say why the early one could not settle the question on its own, usually that it moves before anything has had time to close. Expect them to know what they instrumented before launch, and to name what they wish they had instrumented and did not.

  Weak answer

  A weak candidate uses the same number for both horizons, or stops at opens and meetings booked with no account of what became of them. Asked what would have made them switch it off, they have nothing, because no such number was set before it shipped.

• 07 What GTM automation have you switched off, and how did you decide it should go? Whether they maintain a stack or only add to it

  Follow up with

  • Who noticed, and how long did it take them?
  • What did that teach you about how it got built in the first place?

  Strong answer

  A strong candidate names the thing they killed and what keeping it was costing, in licence spend, in maintenance, or in the confusion it caused downstream. They checked what depended on it before pulling it, and they can say who noticed and how long that took, which is often nobody and never. Expect them to draw something out of it about how it got built, usually that it was made for a campaign that ended and nobody owned retiring it.

  Weak answer

  A weak candidate has only ever added. Every system they have built is apparently still running and still earning its keep, and the question about how they decided it should go gets answered as a hypothetical, because it has not happened.

• 08 Take a GTM system you built and handed to an ops person or a rep. What happened to it after you stopped touching it? Whether they build for a non-technical operator or for themselves

  Follow up with

  • What did the handoff actually consist of?
  • What do you build differently now, knowing it lands with someone who will not read the code?

  Strong answer

  A strong candidate knows what happened after they let go, and it is usually specific: a credential expired, a vendor renamed a field, nobody could follow the branching logic. They can say what the handoff actually consisted of, and it is more than a call and a document. Expect them to name something they build differently now because of it, aimed squarely at the person who will not read the code.

  Weak answer

  A weak candidate has never gone back to look, so the question has no answer beyond an assumption that it is probably still fine. Or they know it broke and explain it as the fault of whoever inherited it, which leaves the handoff they ran unexamined.

• 09 Describe a GTM system of yours that ended up handling a lot more volume than you built it for. What happened? Whether they have met real scale limits or only imagined them

  Follow up with

  • What did that failure look like from the go-to-market side rather than the infrastructure side?
  • What did you rebuild, and what did you just prop up?

  Strong answer

  A strong candidate names the first thing that broke and attaches a volume to it: a vendor rate limit, a per-record cost that stopped being trivial, a sync that could no longer finish inside its window. Ask what it looked like from the go-to-market side and they have that too, usually a rep working a stale list or a number that arrived after the meeting. Expect them to separate what they rebuilt from what they propped up, and to be relaxed about admitting the second.

  Weak answer

  A weak candidate answers in general scaling vocabulary with no first failure and no volume attached to anything. The question asks what happened, and what comes back is what tends to happen to systems like that.

• 10 Tell me about a time you had to cut go-to-market tooling spend. What went, and what did losing it actually cost you? Whether they know the price and the value of their own stack

  Follow up with

  • What did you replace with something you built yourself?
  • What came back six months later?

  Strong answer

  A strong candidate talks in real per-seat or per-credit numbers, because they were the one who went through the invoices. They answer the second half of the question honestly and name what got worse after the cut, rather than presenting it as clean saving. Expect them to separate the tools that were load-bearing from the ones that were habits, and to say which cancelled tool came back six months later.

  Weak answer

  A weak candidate cut whatever was easiest to cancel, and reports that nothing was lost, which is rarely true and was never checked. The question asks what losing it cost, and no cost gets named.

• 11 Tell me about a time you built exactly the GTM system that was asked for and the number it was meant to move did not move. Whether they can see past a brief, and whether they say so

  Follow up with

  • What was that number before, what did it do after, and how long was it before anybody said so out loud?
  • When did you first suspect the brief was aimed at the wrong thing?
  • What stopped you from saying so at the time?

  Strong answer

  A strong candidate gives the before and after figures without being pushed, and says how long it took before anybody admitted out loud that nothing had moved. They name the real problem behind the stated one, which is usually who was being targeted or how a term had been defined rather than anything about the mechanism they built. Expect an honest answer to what stopped them saying so at the time, whether the person who wrote the brief outranked them or the work was already half done.

  Weak answer

  A weak candidate treats the brief as the specification, so the outcome belongs to whoever wrote it. The question hands them a build that did what it was asked and a number that did not move, and the two never get connected.

• 12 Tell me about a time something wrong went out to prospects in your name — a broken merge field, the wrong list, the wrong number. What did you stop, and who did you tell? Whether stopping the send and telling someone are decisions they make, or decisions they wait for

  Follow up with

  • Who made the call to pause the sending, or not to — you, or somebody above you?
  • How many had already gone out before you knew, and what went to those people afterwards?

  Strong answer

  A strong candidate answers the two halves separately, because stopping the send and telling somebody are different decisions with different costs. They know roughly how many had already gone out before they found out, and they can say what those people got afterwards, including the times the answer was nothing. Expect them to name who made each call and how fast, and to have something they would decide differently now.

  Weak answer

  A weak candidate describes the technical repair and stops there. Asked who they told, they waited for a manager to decide both things, or fixed it quietly and kept it inside the team, which leaves the prospects who received it out of the story entirely.

• 13 Tell me about a time your numbers contradicted what a revenue leader believed about their own funnel. Whether they can carry an unwelcome finding without caving or grandstanding

  Follow up with

  • How did you check that you were the one who was right?
  • What changed as a result, if anything?

  Strong answer

  A strong candidate re-verified their own work before taking it anywhere, and can say what that check consisted of. They brought the disagreement with the definitions attached, so the conversation was about what counts as a qualified lead rather than about who was wrong. Expect them to say what the leader did next, and to stay even-handed about it when the answer is that nothing changed.

  Weak answer

  A weak candidate has never had the conversation, at a level where having it is most of the job. Or they tell it as a story about having been right, with nothing about how they checked themselves and nothing about what happened afterwards.

• 14 Tell me about a GTM stack you inherited. What did you find running that nobody had mentioned? How they orient inside someone else's undocumented system

  Follow up with

  • How many things were running that nobody owned, and what happened to each one — killed, adopted, or left alone?
  • How did you work out what depended on what?
  • Who did you have to go and ask, and what did they not know either?

  Strong answer

  A strong candidate can count what they found and say what became of each one, so the answer has killed, adopted and left alone in it rather than a general impression of mess. The discovery method is a real one: following a single record end to end, reading the audit log, watching a week of runs before touching anything. Expect restraint about what they changed first, and a story about asking the supposed owner of something and finding they did not know either.

  Weak answer

  A weak candidate started rebuilding straight away, or writes the previous setup off as bad and moves on. The question asks what they found that nobody had mentioned, and nothing specific comes back, which usually means they never went looking.

• 15 Someone asks you for signals. What did you actually build the last time, and how did you decide what counted as one? Whether signal is something they can operationalise or a word they repeat

  Follow up with

  • Which candidate signals did you test and then throw away?
  • How did you check that a signal predicted anything, rather than just correlating with company size?

  Strong answer

  A strong candidate names the trigger they shipped, what it was supposed to predict, and what happened when they acted on it against a group they did not. The follow-up about what they tested and dropped is worth pressing, because the discarded ones are the interesting half of the answer. Ask how they checked it predicted anything and they take the question seriously, because most candidate signals turn out to track company size.

  Weak answer

  A weak candidate lists signal categories: job changes, funding rounds, hiring. None of them is attached to something they built, and the question asked what they shipped last time. Pressed on whether any of it predicted anything, the answer is that it seemed reasonable.

• 16 Tell me about something you built for reps to use. How many of them were still using it a month later? Whether they measure adoption or assume it

  Follow up with

  • What did the ones who stopped do instead?
  • What did you change once you found out, and did it work?

  Strong answer

  A strong candidate has the real number, and it is usually lower than the launch made it look. They went and watched people work to find out what the ones who stopped were doing instead, which is normally a spreadsheet or a colleague who does it for them. Expect them to say what they changed and whether it worked, including the case where the honest move was to retire the tool.

  Weak answer

  A weak candidate treats launch as the finish line and does not know the number a month later. Non-adoption gets explained as a training problem, so the reps who stopped become the thing that needs fixing rather than the tool.

• 17 Tell me about a platform migration or sunset you ran on a live GTM stack. What did you find depending on the old system that you had not expected? Whether they can change infrastructure under a team that cannot stop selling

  Follow up with

  • How did you find the dependencies nobody had documented?
  • What did you run in parallel, and for how long?

  Strong answer

  A strong candidate names what they found that nobody had written down: an inbound webhook from a vendor, a spreadsheet a rep refreshed every Monday, a report the board had been reading for a year. The discovery pass is described as work they did rather than as diligence they believe in. Expect a cutover with a way back, and a parallel run measured in weeks with something that told them the old system had gone quiet.

  Weak answer

  A weak candidate describes the new platform's advantages and treats the cutover as an announcement with a date on it. The question asks what they had not expected, and no surprises come back, which usually means the surprises arrived after the switch.

• 18 Walk me through the worst failure in a pipeline you owned that fed a go-to-market system — the CRM, a sequencer, a dashboard. Start from how you found out. Whether they own the blast radius or only the job that failed

  Follow up with

  • What had already moved downstream by then — a CRM record, an email, a number somebody had acted on?
  • What guardrail exists now that did not exist before?

  Strong answer

  A strong candidate starts where you asked, with how they found out, and it tells you a lot whether that was an alert or a rep on Slack. They stopped the damage spreading before going after the cause, then repaired the records that had already moved, and they can say what those were: a CRM field overwritten, emails sent on bad data, a number somebody had already acted on. By the time they finish, the guardrail follow-up is usually already answered.

  Weak answer

  A weak candidate fixes the job, re-runs it, and the story ends there. The question asks what had already moved downstream, and the bad rows that reached the CRM and the sequencer never get accounted for.

• 19 In the last cold-email program you ran, how many sends a day did each inbox carry, and how did you arrive at that number? Whether their deliverability numbers come from operating a program or from reading about one

  Follow up with

  • How long was the ramp before you got there?
  • What told you when you had pushed an inbox too far?

  Strong answer

  A strong candidate gives a per-inbox daily figure in the low tens and can say how they arrived at it, which is a ramp measured in weeks rather than a number a vendor recommended. Ask what told them an inbox had been pushed too far and they name a placement or engagement measure they watched, not open rate on its own. Most people who have run a program long enough have backed a number down at some stage, so it is worth asking what made them do it.

  Weak answer

  A weak candidate quotes hundreds a day per inbox, or cannot say what number they actually ran at, which answers the question by not answering it. The second half asks how they arrived at it, and what comes back is advice they read rather than a ramp they ran.

• 20 Walk me through the DNS records on the last sending domain you set up yourself. What is in each one, and why is it there? Whether they have configured sending authentication or only inherited it

  Follow up with

  • Which of SPF or DKIM was doing the aligning when DMARC passed?
  • What did you set the DMARC policy to on day one, and when did you tighten it?

  Strong answer

  A strong candidate accounts for each record and why it is there: an SPF record listing the senders that actually send, a DKIM key they generated and can rotate, DMARC started at none and tightened once the reports were clean. Ask the alignment follow-up and they know DMARC passes on alignment with the visible From domain rather than on SPF or DKIM passing somewhere. Most people who have configured these by hand got one of them wrong at some stage, and a candidate who volunteers which one is telling you they did the work themselves.

  Weak answer

  A weak candidate names the three acronyms and cannot say what any one of them asserts. Or the records were pasted in from what a sending vendor produced, so the question of why each one is there has no answer they own.

• 21 What did one enriched record cost you in your last setup, and how do you think about the trade-off between what you pay and how much you get back? Whether they treat data spend as an engineering constraint

  Follow up with

  • Where were you paying for data nobody ever read?
  • What did you waterfall, and what did you stop buying outright?

  Strong answer

  A strong candidate knows the unit cost and can point at the step that consumed most of it. They buy detail deliberately, running the cheap checks first and spending the expensive credits only on what survives them. Where they were paying for fields nobody read gets a straight answer, with a method behind it, usually looking at which fields the downstream systems actually queried.

  Weak answer

  A weak candidate discusses vendor quality with no number attached to any of it. Credits are somebody else's budget, so the trade-off the question asks about is not one they have ever had to make.

• 22 Tell me about an LLM step you put into a GTM workflow that runs without you watching it — scoring, research, routing, copy. How do you know its output is still good? Whether they operate LLM steps or only ship them

  Follow up with

  • What would tell you it had started producing something different?
  • How do you know a prompt change has not broken something that used to work?

  Strong answer

  A strong candidate describes checks that already exist rather than ones they would build: a held-out set re-scored on a schedule, an alert when the output distribution shifts, a regression run before any prompt change ships. Ask about prompt changes and they have a set of cases they run first, which is the habit a quietly broken improvement tends to leave behind. The thing to listen for at the end is how often they actually look, and whether those checks have ever come back with anything.

  Weak answer

  A weak candidate finds out when somebody downstream complains, or reads a handful of outputs when they remember to. The question asks how they know the output is still good, and nothing they describe would surface a change before a person happened to notice it.

• 23 Your CRM says a campaign produced 40 opportunities. The attribution report says 12. Where do you look? Whether they reach for definitions before dashboards

  Follow up with

  • Which of the two would you take to the board, and why that one?
  • What would you write down so this stops recurring every quarter?

  Strong answer

  A strong candidate goes at the definitions before either system, because both numbers can be right under their own rules: which object counts as an opportunity, what window, first touch or multi-touch, how duplicates were merged. They will ask what the number is for before choosing which one to take to the board. A good answer ends with the rules written down where both teams read them, since this argument returns every quarter.

  Weak answer

  A weak candidate assumes one of the two systems is broken and starts checking the integration. A gap between 40 and 12 is wide enough to be a definition rather than a sync failure, and that possibility never gets raised.

• 24 A vendor returned garbage for yesterday's enrichment run and you need to re-run it. How do you do that without duplicating rows or paying for the same records twice? Whether re-running a job is routine for them or something to avoid

  Follow up with

  • What happens if the re-run dies halfway and you start it again?
  • How do you prove afterwards that the day is now correct?

  Strong answer

  A strong candidate replaces the affected day rather than appending to it, and makes the write idempotent on a stable key so running it twice lands the same rows. Ask what happens when the re-run dies halfway and the answer is that they start it again, which is the point of having built it that way. Expect a concrete check that the day is now right, usually a count plus a sample compared against what the vendor returned.

  Weak answer

  A weak candidate deletes rows by hand, or re-runs the job and plans to clean up the duplicates afterwards. The question names both costs it is protecting against, and paying the vendor a second time for records already bought does not come up.

• 25 A sequence that was replying at four percent drops below one in a week. Nothing about the copy changed. Walk me through it. Whether they reach for infrastructure or for rewriting

  Follow up with

  • What do you check before you touch the copy at all?
  • How would you tell a placement problem from a list problem?

  Strong answer

  A strong candidate uses the sentence about the copy, and treats copy as the last suspect precisely because it did not change. They go at delivery first: authentication, domain reputation, a seed test, and whatever else moved that week in volume, provider or list source. Expect them to hold the answer open, since an unchanged-copy collapse can be placement, a worse list, or a shift in who is being mailed, and a week is not long enough to have separated those.

  Weak answer

  A weak candidate starts rewriting subject lines, which the question rules out in advance by saying nothing about the copy changed. Or they blame list quality without first establishing whether the mail is arriving at all.

• 26 Inbound demo requests are reaching a rep 30 hours after they come in. The SLA is five minutes. Where do you look first? Whether they can decompose a routing path in order

  Follow up with

  • How do you tell a routing failure from an ownership failure?
  • What would you put in place so you knew the next time it slipped?

  Strong answer

  A strong candidate walks the path in order and says it out loud: form submission, record creation, assignment rule, queue, notification, whether a rep was awake. Thirty hours against a five minute target is a stop somewhere rather than slowness everywhere, and they are hunting for the step where the clock parks. What they want afterwards is a timestamp at each hop, so the next slip shows up as a number rather than as a complaint.

  Weak answer

  A weak candidate proposes a new routing tool, or decides the reps are ignoring their alerts, before knowing whether the lead reached one. The gap in the question is wide enough to locate precisely, and nothing in the answer would locate it.

• 27 You are four days into a test of a new outbound sequence, the variant is replying 30 percent better, and a VP wants it rolled out to the whole list on Monday. What do you say? Whether they can hold a line on evidence without stonewalling

  Follow up with

  • What had you agreed before the test started?
  • What would make you happy to call it early?

  Strong answer

  A strong candidate explains in plain language why a four-day read moves around, without leaning on the word significance, and goes to what was agreed before the test started, including the case where the answer is nothing. Then they give the VP a route: what they would need to see to call it on Monday, or a partial rollout that keeps the test alive. Expect them to say what would make them happy to call it early, because sometimes the gap is wide enough and the downside small enough.

  Weak answer

  A weak candidate concedes on the spot because a VP asked, or lectures about sample size and offers nothing the VP can act on. Either way the person who has to decide on Monday leaves with the options they walked in with.

• 28 A prospect receives four emails in one day from four different systems. How do you stop that happening again? Whether they fix the instance or the architecture

  Follow up with

  • Which system should own the decision to send, and why that one?
  • How would you make the next tool somebody buys inherit the rule?

  Strong answer

  A strong candidate reads four systems as the problem rather than four emails, and puts the decision to send behind one arbiter holding suppression and frequency for all of them. They will name which system that should be and why, usually the one every tool already syncs through. The harder half is the second follow-up, and a good answer says how the next tool somebody buys inherits the rule instead of routing around it.

  Weak answer

  A weak candidate turns off the offending campaign, or proposes a standing meeting between the teams that own the four systems. The question asks how it stops happening again, and a fifth system could be bought next month under either answer.

• 29 Tell me about a time you took an LLM out of a workflow you had put it into. Whether models go in deliberately or by default

  Follow up with

  • What was it doing that a rule could do?
  • What did you give up by taking it out?

  Strong answer

  A strong candidate names the step and says why the model was the wrong tool in that spot: too slow for where it sat, too expensive per record at the volume it ran, or less reliable than the rule that replaced it. They answer the second follow-up honestly and say what got worse, because taking it out usually costs coverage on the awkward cases. Expect the judgment to be about that step rather than about models in general.

  Weak answer

  A weak candidate has only ever added AI to workflows and has nothing to take out. Or the story is entirely a cost decision, with nothing about whether the output was any good, so what they gave up never gets answered.

• 30 Tell me about something you built where a model does part of the work and a person still checks it. Who checks, and what are they looking for? Whether the human step is designed or just assumed

  Follow up with

  • What happens to the ones they reject?
  • How much of it do they actually get through in a week?

  Strong answer

  A strong candidate names the person or the role, what they are looking at, and roughly how many they get through in a week, which is the number that tells you whether the review is real. Rejections go somewhere: back to the model, into a rule, onto a list somebody works. Expect them to say what they did when the queue grew faster than the reviewer could clear it, because it usually did.

  Weak answer

  A weak candidate says a human reviews everything and cannot say what that person looks for. The follow-up about weekly throughput is where it comes apart, because the review they have described would take longer than the reviewer's week.

• 31 Tell me about a GTM system you built without writing code. What made you decide it did not need any? Whether they fit the tool to the problem or to their comfort

  Follow up with

  • Who had to be able to maintain it after you?
  • When would you have reached for code instead?

  Strong answer

  A strong candidate names the constraint that made it the right call: who owned it after them, how fast it had to exist, what volume it was ever going to see. They can say where that call would have flipped, and the flip is usually a volume or a branching condition rather than a preference. What separates a choice from a limit is whether they build both ways, and that usually surfaces without being asked.

  Weak answer

  A weak candidate builds one way and this happened to be it, so the question of what made them decide has no decision in it. Or they treat not writing code as a compromise needing an excuse, and spend the answer on what they would have built with a free hand.

No question carries that tag.