• LinkedIn
• Twitter
• Instagram
• YouTube
• RSS Feed

Apple •Spotify• Pocket Casts •Youtube •Overcast •RSS

What’s up everyone, today we have the pleasure of sitting down with Kelsie Dube, Director of Marketing Operations and Automation at Unanet.

Summary: Kelsie breaks down the 30-day audit she runs before changing a single tool, how she sorts 30-plus problems into big rocks, quick wins, and parked FY27 headaches, and why she’d rather build a “digital twin” to argue with than buy another platform. She gets into the tiger-team playbook that makes tool replacements actually stick, the measurement fight every team should have out loud, and what happens to junior marketers when AI eats the entry-level rung. It’s a full field guide to walking into a new company and earning trust before you change anything, plus the interview question she uses to know if you’ll thrive on her team.

In this Episode…

• Why New Marketing Ops Leaders Should Audit Before They Act
• How to Run a Marketing Ops Audit When You Join a Company
• How to Balance a Full Audit Against Your Boss’s Priorities
• How to Spot a Process Problem Disguised as a Tool Problem
• How to Surface Shadow Processes and Duplicate Tools in an Audit
• How to Choose a New Martech Tool With a Tiger Team of Power Users
• How to Turn a Marketing Ops Audit Into a Prioritized Roadmap
• How to Align Your Team on a Measurement Philosophy
• What Happens to Junior Marketing Ops When AI Eats the Entry-Level Work
• How to Turn Marketing Ops From Ticket-Takers Into Business Partners
• How to Use One Interview Question to Screen for Team Fit

About Kelsie Dube

Kelsie Dube is the Director of Marketing Operations and Automation at Unanet, where she joined in 2026 to rebuild the company’s measurement and reporting foundation. Before that she spent 7 years at Sophos, climbing 5 roles from Marketing Operations Specialist to Director and leading the team through much of that run. She started her career in data entry, scrubbing lead lists and learning the VLOOKUPs she still uses most days, and that ground-floor view shapes how she thinks about the craft.

Outside of work she’s a mom of 2, an amateur basketball player, and a snowboarder who guards the hours between 5 and 7 for her kids.

Why New Marketing Ops Leaders Should Audit Before They Act

The first 90 days at a new company come with a quiet dare. Ship something. Plant a flag. Prove they were right to hire you. Most new marketing ops leaders answer that dare by ripping out a tool they never liked and installing one they used at their last job, all inside the first 2 weeks.

Kelsie did the opposite when she joined Unanet. Her first move was to change nothing.

“You have to have that knowledge before you start changing things, otherwise you’re just band-aiding, and you might not be focusing on the biggest rocks.”

The reasoning was twofold. Without a real map of how work gets done, any change you make is a guess, and you’ll probably aim it at the wrong target. And Unanet had said in the job description that they wanted someone who’d slow down and understand the business first, which is part of why she took the role. For a marketer who came up through ops and now leads it, that patience is the whole point. You have to understand what’s happening well enough to be dangerous, she says, before you can tell anyone what to actually fix.

The harder version of this is resisting your own experience. A new CMO walks in and wants a different marketing automation platform because they used one at the last company. Kelsie feels that pull too, since she’s learning a few systems here for the first time. Her answer is to make the existing stack work before spending a dollar to replace it. Right now she’s rebuilding how campaigns get captured and tagged to campaign members, and she’s solving it with a Salesforce workflow or Power Automate instead of a purchase. Replacing the tool would cost more and probably solve less.

Then comes the part nobody warns you about, the clock. Listening mode has a shelf life, and everyone, including you, is quietly counting the days until you produce something. Kelsie gave herself a hard deadline of 30 days to finish the audit, then engineered small wins into every week so the learning never looked like stalling.

New hires have one thing veterans don’t, an empty calendar, and she guarded it. A recurring block every Friday at 2:00 to turn that week’s discoveries into something usable. One of those blocks became a 30-minute campaign taxonomy diagram she walked the leadership team through, a quick and visible proof that better structure meant better reporting down the line.

The instinct to ship in week one is the most expensive habit in operations, because a fast fix aimed at the wrong problem still has to be undone later. The operators who compound value are the ones who can sit in not-knowing long enough to find the actual bottleneck.

Key takeaway: Give yourself a fixed audit window, 30 days is a reasonable default, and block a standing hour each week to turn what you’ve learned into one small, visible win. Keep a running log of quick wins as you find them so you always have something to ship while the deeper work is still in progress.

Back to the top ⬆️

How to Run a Marketing Ops Audit When You Join a Company

“Do an audit” is the most repeated and least defined phrase in marketing ops. Everyone nods along. Almost nobody can tell you what’s actually in one, which funnel stages it covers, or where it starts on a Monday morning.

Kelsie’s version starts before day one. The moment she accepted the Unanet offer, she began mapping ownership lines: what marketing ops owns, what rev ops owns, what IT owns. She wrote out the questions she had and sent them to her leader ahead of her start date, partly out of genuine excitement and partly to get a head start on the RACI.

“Where are the lines that marketing ops owns? What does rev ops own? And what does IT own?”

From there it’s a series of conversations that snowball. She interviews everyone from her own team to her peers to the level above to her business partners, and she ends every conversation the same way: who else should I talk to? The list never stops growing, but the patterns start to show. Along the way she confirms the systems in her swim lane against the job description, asks what she’s missing, and books dedicated working sessions where someone walks her through a live task like a lead import. She records those sessions, then feeds the transcriptions to an LLM to pull out the patterns, a tip she picked up from a mentor that turns hours of shadowing into a usable roadmap.

She also runs the work herself.

Joining a much smaller team than she was used to, she asked people to hand her a lead import and a campaign creation task so she could run them personally. Part of it was capacity, since a director who can still execute is extra hands when the team is slammed. Part of it was fluency. As she says, a director isn’t really a “do-rector,” but being right there in the work is how you understand it well enough to change it later.

Data privacy rides shotgun through all of this. Coming out of the cybersecurity world, she asks about data retention and opt-in policies early, then files them under “needs structure, not on fire” so the genuinely urgent problems keep her attention first.

The audit that produces a good roadmap is the one where the leader does the grunt work themselves for a few weeks, because you can’t prioritize a process you’ve only heard described. Handing the discovery phase to someone else is how new leaders end up fixing the loudest problem instead of the biggest one.

Key takeaway: Before you start, write down every question you have about ownership and send it to your manager. Then book walk-me-through sessions for each core process, record them, and run the transcripts through an LLM to surface patterns. Ask to personally complete at least one live task per system so you understand it by doing, not just by watching.

Back to the top ⬆️

Speaking of marketing audits… are you still operating like it’s 2013 and you haven’t updated your chatbot? We recently started collaborating with the Docket team and what they are building with their Inbound Demand Agent is super cool. Check it out.

How to Balance a Full Audit Against Your Boss’s Priorities

Every new operator gets pulled in 2 directions at once. You want the full map, every system, every handoff, every place a lead falls through the cracks. Your boss wants impact, and wants it visibly, soon. Those 2 goals fight each other, because a real understanding takes time and a quick win needs context you don’t have yet.

Kelsie handles the tension with communication that’s almost aggressively clear. Rather than quietly disappearing into discovery, she tells leadership exactly what she’s building and when: a roadmap that runs 2 quarters out, and she needs their alignment on it.

“Here is the roadmap now that I’m working to build. It’s gonna be two quarters out, and I need alignment from the team.”

She got a break at Unanet that doesn’t come every time. After her first 30 days, leadership asked her to educate them on what she’d found. So she presented the whole audit: what impressed her, what was working, and the surprises that would need digging into. Organized into categories, that presentation walked straight into her roadmap, which happened to deliver the exact outcomes leadership was chasing. Now those buckets are landing in Q3 planning as OKRs, a framework she likes and one Unanet already runs on.

Kelsie was kind enough to build and share her Marketing Operations Assessment Framework, grab your copy here: https://drive.google.com/file/d/1x3IJf9tHku4-fASS9AQE_F6sJ1FLK3PN/view?usp=sharing

The non-negotiable underneath all of it is sequencing. When leadership says they want better reporting, she agrees, then explains why the data foundation has to come first. Building an executive dashboard on top of bad data just automates the wrong answer faster. Getting alignment on that order of operations, before anyone builds a single report, is what keeps the quick-win pressure from steering the whole ship.

The operators who survive their first year make their sequencing legible to leadership before they start. When executives understand why the boring foundational work has to precede the dashboard they asked for, patience stops looking like inaction and starts looking like a plan.

Key takeaway: Turn your audit into a dated roadmap that runs 2 quarters out and present it to leadership as findings plus a sequence, not a list of complaints. Name the dependencies out loud, especially where a data foundation has to precede the reporting they want, and get explicit alignment on that order before you build anything.

Back to the top ⬆️

How to Spot a Process Problem Disguised as a Tool Problem

Sometimes the audit turns up a clean stack. The tools work. Nothing’s obviously broken. And the problem is still there, sitting one layer down where no software can reach it, in how people actually work.

Kelsie’s tell for this is a leader’s willingness to question the sacred cow. The sentence she listens for is “that’s how we’ve always done it,” which she calls the worst thing you can say in operations. A former leader drilled that into her, and she says it matters even more now, a couple years into AI, when almost every process deserves a rethink.

“I’ve had an amazing leader that’s always said, ‘Don’t ever say this is how we’ve always done it,’ which is even more true now a couple years into AI.”

Here’s how it plays out. The business asks for a new tool to fix a problem. She digs a layer deeper and finds there’s no clearly defined process underneath it. Buying software there would be a Band-Aid, stacking cost and a rip-and-replace on top of a broken workflow that stays broken. The honest fix is usually cheaper and more annoying: educate the team, redesign the process, or challenge people to think about it a different way.

When the gap is a skills gap, she has a specific prescription. Go build your digital twin, riff with it, and come back with an idea worth rumbling on as a group.

Building a Digital Twin to Pressure-Test Your Thinking

Kelsie built hers during a company-wide project with AI advisor Liza Adams. It’s a custom GPT loaded with everything she could feed it about how she thinks and what she values, and she gave it a deliberately silly name, the “Doobie Strategist.” You could build the same thing as a Claude skill or in whatever tool you like. What matters is what it does for you on a bad day.

Its main job is to argue. When she has an idea, restructuring her team, say, she tells it to poke every hole it can find. It asks the questions she’s too close to see, pulls in deep research on how else the problem could be solved, and stays level when she’s ready to declare the whole thing broken. It also works as a rehearsal partner. Before a big presentation she’ll tell it who’s in the room and have it react from their point of view.

“I want you to challenge this idea and poke all the holes you can possibly poke in it.”

The most useful AI in operations right now is the kind that disagrees with you. A model tuned to your blind spots catches the expensive mistakes before they reach a slide deck, a better return than another tool that automates a task you never thought hard about.

Key takeaway: Build a digital twin as a custom GPT or Claude skill, loaded with how you think, your past projects, and what you value. When you face a real decision, tell it to poke every hole in your plan, then use it to rehearse high-stakes presentations by having it respond as the specific people who’ll be in the room.

Back to the top ⬆️

How to Surface Shadow Processes and Duplicate Tools in an Audit

The most dangerous parts of any stack are the ones that don’t show up in a stack. The spreadsheet on someone’s local desktop. The tool a team pays for on a personal card. The workaround somebody built because the real system never did what they needed. None of it appears in the admin panel, and all of it breaks the day that person is out sick.

Kelsie’s take is that undocumented work is almost always a symptom. If a process only lives in one person’s head, that’s a single point of failure hiding in plain sight. So instead of policing it, she gives it a deadline. Document this fully enough that if you’re out tomorrow, I can jump in and run it. Then she does exactly that, on a call, following their guide step by step while they watch, which usually surfaces the one step they forgot they even do.

“I need you to document this to the fullest so that if you are out of the office tomorrow, I can just jump in and do it.”

The Curious-Before-Critical Test for Duplicate Tools

Then there’s the overlap problem, the one every operator recognizes on sight. One team’s on Revenue Hero, another on HubSpot, another on Calendly, another on Cal.com, 4 tools doing one job while everyone acts like it’s fine. The rookie instinct is to walk in hot and start consolidating.

Kelsie flags the duplicates, then does something harder than cutting them. She stays curious. She asks each user to walk her through what their tool does from their seat, then moves to the next person, and more often than not there’s a genuine niche reason the “redundant” tool exists. The trust she builds by wearing what she calls her listener hat is what eventually earns her the room to consolidate. Coming in hot would have burned it. The exception is when the money is loud enough to override the diplomacy, and she draws that line herself: 3 webinar tools would get fixed fast, because event platforms are brutally expensive.

“I’m really getting curious before I get critical.”

Tool consolidation fails more often from bad timing than bad analysis. The duplicate stack a new leader spots in week one is real, but the political capital to fix it has to be earned first, and the operators who lead with a spreadsheet of wasted spend usually get overruled by the people they embarrassed.

Key takeaway: When you find an undocumented process, set a hard deadline for the owner to write it up, then run it yourself from their guide while they watch to catch the missing steps. For duplicate tools, interview each user about why their version exists before proposing a cut, and prioritize by cost, tackling the expensive overlaps like event platforms first.

Back to the top ⬆️

How to Choose a New Martech Tool With a Tiger Team of Power Users

Every “don’t buy another tool” evangelist has a story about the time they had to buy another tool. Kelsie’s starts with the worst kind of homecoming.

She came back from maternity leave to a promotion and a mess. While she was out, the org had gone live on a new event platform, and the marketers who had to use it every day couldn’t stand it. The ops team that bought it was happy. The people living in it were miserable. That gap between buyer and user was the whole problem, and it was now hers.

She ran the same play she runs everywhere: curious before critical. She set up one-on-ones with the actual users and made sure they felt heard, which bought her the time to figure out what was really going on. The team poured effort into enablement, retraining, office hours, a dedicated person embedded with the department. Adoption got better. But the tool’s reliability fell on its face anyway, and no amount of training fixes software that doesn’t work.

“We gathered a tiger team of those power users that were very passionate about why the tool wasn’t working and what they really needed.”

So they replaced it, and the way they replaced it is the part worth copying. She pulled the power users, the ones most fired up about what was broken, into a tiger team. They wrote real requirements. They ran a full RFP. When it came down to the top 3 vendors, the tiger team sat through those demos and helped make the final call. The people who would live with the decision owned the decision.

The payoff shows up in a single moment she’s learned to watch for. Somewhere in a demo, a side message pops up from someone on the team: I want to use this today. That’s when she knows the choice is right, because the buy-in is already there before the contract is signed. She’s run this same process on repeat since, and it pays off every time. Even the enablement went about 75% faster than it had with the tool they were leaving, because nobody had to be talked into it.

The failure mode in martech buying is picking any vendor without the people who’ll use it in the room. A tool chosen by the buyer and handed to the user starts every implementation with a trust deficit, and that deficit costs more than the license ever will.

Key takeaway: When a replacement is genuinely necessary, assemble a tiger team of your most frustrated power users and put them through requirements gathering, the RFP, and the final vendor demos. Buy-in built during selection makes enablement dramatically faster, so watch for the moment a user says they want to start today and treat that as your green light.

Back to the top ⬆️

How to Turn a Marketing Ops Audit Into a Prioritized Roadmap

An audit ends with a pile. 30-plus problems, half of them tangled together, all of them wearing the same “urgent” costume. Kelsie found exactly that at Unanet: over 30 things she wanted to fix inside her first 30 days, and the honest first reaction of staring at the list and thinking, where do I even start.

The way out is weighting. She scored each item on 2 axes at once, the order of operations (what has to be fixed before anything else can move) and effort versus impact. That scoring is what turns a pile into a roadmap, and it sorts everything into 3 honest buckets:

• Big rocks, the multi-epic initiatives that move the real outcomes and get their own dedicated runway.
• Quick wins, the small, high-visibility fixes she slots into Q3 and Q4 to keep momentum showing.
• Parked problems, the FY27 items that genuinely matter but aren’t on fire and won’t move the needle yet, so they wait.

“In the 30 days, I identified over 30 different things that I wanted to tackle. And just looking at the list, I was like, ‘Oh gosh, where do I even start?'”

The reporting example shows how deep a single item can run. She was hired to build a proper reporting suite so marketers could see what’s working. But going a layer down, the data underneath wasn’t ready. She could ship a report tomorrow and it would confidently repeat the same wrong numbers everyone already had. So the real work moved upstream, into UTMs, capture, and how campaign source gets tagged, before a single dashboard gets built. The campaign taxonomy piece was a quick win. A self-serve form that spits out a campaign code, something she’s built before and loved, got parked, because the manual version takes a few seconds today and automating it can wait.

Then there’s the political problem every audit creates: you were hired to fix one thing and found 5 others. When lead routing and lead scoring light up red but measurement was the mandate, she makes the case by translating. Think about who you’re presenting to and what’s in it for them. Framed right, fixing routing becomes faster leads to the right reps, fewer angry BDRs, and more revenue, which is the language leadership was speaking the whole time.

The roadmap is where operations leaders either earn trust or quietly lose it, because a list of everything wrong reads as criticism while a sequenced plan reads as leadership. The skill that separates the 2 is translation, turning “your routing is broken” into “here’s the revenue we’re leaving on the table and when I’ll get it back.”

Key takeaway: Score every audit finding on 2 axes, dependency order and effort versus impact, then sort into big rocks, quick wins, and parked items with real dates. When you find problems outside your original mandate, pitch them in the outcome language leadership cares about, faster leads, less friction, more revenue, rather than as new problems you dug up.

Back to the top ⬆️

How to Align Your Team on a Measurement Philosophy

Hire 3 sharp measurement people and you might get 3 religions. One’s a multi-touch attribution truther. One will die on the MMM hill. The third runs everything through experimentation and thinks the other 2 are kidding themselves. All 3 can be right, and all 3 can add value, which is exactly why the philosophy fight has to happen out loud before anyone builds a report.

Kelsie’s starting point is that measurement, like everything else in the role right now, is up for a rethink. AI is already forcing operators to reimagine how they work, so she brought that same posture to reporting. In her intake interviews she skipped straight past the model and asked what people actually couldn’t get today. The answer that kept coming back had nothing to do with attribution theory.

“It’s not really about the attribution. It’s more so like we just wanna know what’s helping drive what.”

The team knew what engagement looked like. What they couldn’t see was whether any of it drove pipeline. So the real question was a mechanical one: what’s helping drive what.

To answer it, she got lucky with timing and had an offsite early in her tenure. They ran a full recorded workshop walking through how every person wanted to report, which turned out to differ by channel, and hunted for the synergies that could become one clear path. She pressure-tested her own defaults with her digital twin and did deep research on how other B2B companies handle it, specifically to avoid locking the team into last-touch just because it was familiar.

The thing she won’t skip is alignment. The whole point of the workshop was that everyone walks out ready to use the same model until the day they decide to change it, and everyone knows what they’re looking at when they open a report.

Most teams need a shared measurement model more than they need a better one. A mediocre framework everyone trusts and reads the same way beats a sophisticated one that half the org quietly ignores.

Key takeaway: Before choosing an attribution model, interview your team about what they can’t see in today’s reporting, and you’ll usually find the real need is connecting activity to pipeline. Run a working session to align everyone on one shared approach per channel, and make the goal a model the whole team will actually use and read the same way.

Back to the top ⬆️

What Happens to Junior Marketing Ops When AI Eats the Entry-Level Work

Kelsie started where a lot of good operators started, scrubbing lead lists and learning VLOOKUPs. That grunt work built the logic she now leads with. And that exact grunt work is what AI does for free today, which sets up an uncomfortable question: if the entry-level reps disappear, where does the next generation of judgment come from?

She takes the worry seriously. There’s a real version of the future where nobody gets the reps she got, and you end up with a missing rung on the ladder, no juniors learning the fundamentals and eventually no seniors who ever did. Her advice for anyone starting out now is concrete: build your digital twin, learn AI deeply enough to code a little yourself, and get fluent in what these tools can and can’t do.

“The advice I would give to someone coming fresh out of college is build out your digital twin.”

The role she describes for a new junior barely resembles the one she started in. Picture someone hired today to manage a series of data-cleanup agents, one agent runs a VLOOKUP and hands off to the next. That’s a real, reimagined job. But she’d hold that person to a specific sequence before trusting the output:

1. Do the process yourself first, by hand, so you actually understand what correct looks like.
2. Test the agent’s work against your own, and cross-reference until you trust it.
3. Run the agent, watch the outcome, and when it gets something wrong, dig into the logic until you know why.

The failure she’s guarding against is someone who never learned step 1, passing bad work through a pipeline they can’t evaluate.

Then there’s the optimistic case, which Phil heard from a former CEO now on the board at GrowthLoop. His take flips the anxiety on its head. The grads walking in today grew up on AI tools and are already native to them in a way most of us aren’t, so the move is to build new roles around how they work instead of mourning the training grounds we came up through. Kelsie lands in the same excited place. If you’re willing to rethink the work from scratch, anything’s possible, and a single creative operator pulling a team of agents now covers ground that used to take a room full of juniors.

The missing-rung problem is real, but it assumes the old ladder was the only way up, and it wasn’t. The operators who thrive in the next 5 years will be the ones who can define correctness well enough to supervise a machine doing the work they once did by hand.

Key takeaway: If you’re early in your career, build a digital twin and learn to code enough to understand what AI can and can’t do, then practice the fundamentals by hand before you hand them to an agent. If you’re hiring juniors, structure the role around supervising agents, but require them to run each process manually first so they can catch what the machine gets wrong.

Back to the top ⬆️

How to Turn Marketing Ops From Ticket-Takers Into Business Partners

Marketing ops runs the engine of go-to-market and gets almost none of the credit. It’s the hidden machinery, the function that shows up in a meeting only when a lead routes wrong. Mike Rizzo wrote about this recently, and Kelsie’s reaction was immediate agreement, because a team that only ever hears about its failures starts to believe that’s all it produces.

Her fix starts with a reframe she repeats to her teams constantly: you’re a business partner with your stakeholders, there to build the relationship, not just to work the queue.

“We’re a business partner. We’re not just behind the scenes taking a ticket, moving on to the next thing.”

The mechanics of that are relational. An operator who meets with their stakeholders regularly shows up to the ticket already knowing what’s coming, and when the rush request lands, the rapport is already there to say yes. When something breaks, whoever’s behind the ticket becomes a person who can add a human element to the mess. That relationship outlasts the job, too, and turns into a network connection when either person moves on.

So she challenges her team to know every single person they take requests from. At a big org that might be 30 people, which is why she leans on structure to make it scale:

• A monthly ops workshop or open office hours where stakeholders bring their tool problems or hear what’s new.
• Deliberate airtime for wins in team and marketing meetings, celebrating the leads that routed correctly instead of only autopsying the 2 that didn’t.
• A strategic project for every person on the team, so there’s always work that feels like moving the needle sitting next to the day-to-day queue.

That last one is the pressure valve. Business-as-usual work is a treadmill, and the junior folks buried in tickets are the ones most likely to burn out on it. A strategic project is how someone gets to feel like they’re making an impact rather than just clearing a backlog.

The visibility problem in operations is a story problem, and teams that only narrate their failures train everyone, including themselves, to see ops as a cost center. The leaders who fix it manufacture proof of impact on purpose, because the work being essential has never been enough to make it visible.

Key takeaway: Reframe your team as business partners by having each operator build real relationships with the stakeholders they serve, and run a monthly office-hours session to make it scale. Deliberately celebrate wins in team meetings instead of only reviewing what broke, and give every person a strategic project so no one lives entirely in the ticket queue.

Back to the top ⬆️

How to Use One Interview Question to Screen for Team Fit

Kelsie has one interview question she leans on above the rest: tell me about the best leader you’ve ever worked for, and why. On the surface it sounds warm. Underneath, it’s a compatibility test.

She takes competence as a given. What she’s testing for is whether the person will actually thrive under how she leads. Someone who needs a defined structure and a daily list of exactly what to do won’t gel with her, because her style runs on autonomy: get your work done, come to me when you want to brainstorm or you’re stuck, and otherwise don’t expect to be watched.

“I want you to be autonomous, you get your work done, and come talk to me when you need me.”

The deeper bar is bigger than fit. She wants every person on her team to do the best work of their careers, and the question is really asking whether that’s possible under her. So she digs, asks about a leader they didn’t enjoy, and reads that answer for what it tells her about how they want to be managed.

Fit is only half of it, though, because you don’t always get to hire your team. Sometimes you inherit one, and then the job is gelling people who didn’t choose each other. Her tool there is feedback, and she’s adamant it starts with her. In one-on-ones she asks what she could do or stop doing to make the job easier, and it takes a while before anyone believes her enough to answer honestly. Once that door opens, feedback runs both ways.

She’ll even take it live. If someone says a process is broken in front of the whole team, her response is “great, tell me why,” because working through it together in the room usually surfaces the better version faster than any private conversation would. The only line is professionalism.

She has a favorite example of this working. Her team was standardizing a process, so they ran a virtual workshop on sticky notes, flagging any disputed step in a “needs review” color. They separated the must-haves from the up-for-negotiation steps, ended up with 10 or 12 contested notes, and then took 2 a week: whoever proposed a step had to stand up in the team meeting and argue for it while the group voted. One team member owned something like 90% of the notes and presented nearly every week. Some made the cut, some didn’t, and the team came out the other side running one genuinely shared process, which matters most when people cover for each other across regions.

Hiring for competence is table stakes, and the leaders who build durable teams screen for whether someone can be autonomous and take feedback in public. A process everyone argued over and voted on holds up under pressure in a way a mandated one never does, because the team defends what it helped build.

Key takeaway: Ask candidates about the best and worst leaders they’ve worked for, and use the answers to gauge fit with your actual management style, not just their skills. To align an inherited team, invite feedback about your own leadership first in one-on-ones, then standardize contested processes out in the open by letting people argue each step and vote, so the final version is one the whole team owns.

Back to the top ⬆️

How a Marketing Ops Leader Decides What Deserves Her Energy

Every guest on the show gets the same closing question: how do you decide what deserves your energy, and how do you stay aligned with what actually makes you happy? For Kelsie, a marketing ops leader, mom of 2, amateur basketball player, and snowboarder, the answer runs on 2 rails at once.

The first is being challenged. She checks in with a simple gut question, am I doing the best work of my career right now, and she’s happiest when she’s trying something new rather than running the same old play. Protecting room for that at work is half of what keeps her engaged.

“I’m always happiest when I’m trying something new and I’m not just doing the same old.”

The other rail is the boundary. She holds a hard gap between 5:00 and 7:00 where the laptop stays closed, because that’s kid time and she wants them to see it protected. The line moves when a hard quarter demands it, but she names that out loud and prepares the family for it rather than letting it creep in silently. Underneath it is a blunt piece of perspective: the job might not be there tomorrow, so the boundaries are worth holding.

Her new role shifted the daily rhythm. She used to start at 6:30 AM, and a more global company gave her more normal hours and, better than that, the chance to wake her kids up in the morning. The routine now is simple. Move her body, sometimes a real lift and sometimes just a walk on the walking pad while she clears a few emails. A good cup of coffee, which she refuses to compromise on. Then the kids, then the day.

“We work too hard to have shitty coffee.”

The operators who last in a function this relentless are the ones who decided in advance what they won’t trade. Boundaries only hold when they’re named out loud and defended before the crunch, not negotiated during it.

Key takeaway: Audit your own energy with one honest question, whether you’re doing the best work of your career, and treat trying something new as a requirement rather than a luxury. Set a hard daily boundary, protect it visibly, and when a tough season forces you to bend it, say so in advance and bring your family or team along instead of letting the creep happen quietly.

Back to the top ⬆️

Episode Recap

Kelsie Dube makes one argument that runs through everything else: when you join a new company, the fastest way to lose is to move fast. Her whole method is built on auditing before acting, understanding how work actually gets done before changing a tool, a process, or a report. The discipline underneath it is patience, and the payoff is fixing the biggest problem instead of the loudest one.

The tactics ladder up from there. She maps ownership lines before day one, interviews everyone in sight, records the sessions and runs them through an LLM, and does the frontline tasks herself so she understands them well enough to change them. She scores every finding on dependency order and effort versus impact, then translates the scary red flags into the revenue language leadership already speaks. When a tool genuinely has to be replaced, she hands the decision to a tiger team of the people who’ll actually use it.

The bigger theme is that most marketing ops failures come down to people and process rather than technology. Duplicate tools, broken measurement, shadow spreadsheets, and burned-out ticket-takers are problems no purchase fixes. Her answer is operational discipline plus a genuine investment in the team, treating operators as business partners and manufacturing the visibility the function rarely gets on its own.

She’s honest about the parts that are still open. She’s only a couple months into Unanet, so the audit-then-roadmap approach doesn’t have results to point to there yet. She sits in the real tension between preaching “you don’t need another tool” and building a career on automation, and she doesn’t pretend the missing-rung problem for junior operators is solved. What she offers is a way of working through it: build the fundamentals by hand, then supervise the machine that does them.

Connect with Kelsie on LinkedIn.

Full episode ⬇️ or Back to the top ⬆️

✌️

—
Intro music by Wowa via Unminus
Cover art created with Midjourney (check out how)

Apple •Spotify• Pocket Casts •Youtube •Overcast •RSS

Related tags

See all episodes

Future-proofing the humans behind the tech

Apple •Pocket Casts•Google •Overcast •Spotify •Breaker •Castro •RSS

• LinkedIn
• Twitter
• Instagram
• YouTube
• RSS Feed