Doug Camplejohn
(00:01)
Hello everyone, this is Doug Camplejohn. Welcome to this week’s episode of Revenue Renegades. Today I’m very excited to welcome Jeremy Donovan from Insight Partners, who does everything around revenue strategy for them. Welcome, Jeremy.
Jeremey Donovan
(00:16)
Yeah, super excited to catch up. It’s been a couple of years, maybe dating back to your LinkedIn days. Really good to talk.
Doug Camplejohn
(00:25)
It would take too much time to give Jeremy’s background, but he’s technical with advanced degrees in engineering and data science. I can’t remember if it was when you were at SalesLoft we met, but you’ve had stints at Gartner, Xilinq, CB Insights, and others. We had a chance to chat in a revenue ops round table and reconnect. Excited for this chat.
Jeremey Donovan
(00:53)
Very cool.
Doug Camplejohn
(00:55)
To kick off, maybe tell us a little about Insight Partners and your role.
Jeremey Donovan
(01:02)
Super high level. I actually hadn’t heard of Insight, which is a bit embarrassing when I joined SalesLoft. Now I can do the spiel. It’s $90 billion in assets under management, 556 B2B SaaS portfolio companies. There’s an investor side, which I am not on, and an advisory side, which I am on. On the advisory side, we have about 40 advisors who are ex-operators like myself who help our portfolio companies.
Once they’re on board, it’s a pull, not a push model. People think of PE or VC as coming in and telling people what to do. We’re the exact opposite. We’re here when you need us as advisors and consigliares, if you’ll allow us. And if not, then you go on your way, and that’s great too.
Doug Camplejohn
(01:55)
That’s kind of a dream for the entrepreneur. We love that model. I don’t know if it’s before SalesLoft days or even in those days. I don’t know when RevOps was founded; earlier in your career, the term didn’t even exist. You’ve posted about the shift to collaborative RevOps team being the number one core hire for new CROs. Can you talk about that?
Jeremey Donovan
(02:25)
Definitely happened in SalesLoft days. I didn’t start my career in sales. I consider myself a bit of a fraud because I never carried a bag. I’m geeky about sales: I read every book, listen to as many podcasts as possible, including yours, and talk to CROs and salespeople to learn and absorb.
The reason I found myself in it is my engineering background. I wanted domains where I could apply analytical, statistical, process design rigor into a field that needed it. That’s what drew me to RevOps—sales was considered more art than science. At SalesLoft, I was lucky to work for CRO Sean Murray. Sean told me, “I’m going to work in the business, and you’re going to work on the business.” “In the business” means talent stuff, picking people, performance management, deal reviews, and travel. My job was infrastructure strategy and planning—so we’re prepared months ahead and avoid bottlenecks.
During that time, “RevOps” as a term started to be used. Most RevOps today is still highly decentralized. In the beginning, there was just sales ops; RevOps was the same. There were separate CS ops and marketing ops, but we thought of them collectively. Eventually, the company absorbed CS ops into sales ops, but marketing ops still sat under the CMO.
Jeremey Donovan
(04:53)
I don’t know if it’s the same today, but I think the operations person should be the consigliere to the function leader. I don’t believe everything needs to be centralized. If you have a chief customer officer, head of sales, and CMO as peers, just let their ops people report to them. There’s no harm in that.
Doug Camplejohn
(05:32)
Got it. So you’ve seen it work both ways—centralized RevOps reporting to the CEO and individual functional ops people.
Jeremey Donovan
(05:43)
It works both ways. It’s also important: every one of us has “If I knew 20 years ago what I know now” moments. Earlier in my career at SalesLoft, I was territorial. I didn’t necessarily play nice with others early on. I needed strict boundaries and got upset when people stepped into what I saw as my territory.
One realization as I got older: these boundaries are amorphous. Folks who went into my space mostly did so with the best intentions. Sure, sometimes there are bad actors, but mostly people are trying to do good. If you’re aligned on your North Star—growth, or now “efficient growth”—then it’s just a team. Who cares who does what? That maturity is important in decentralized worlds.
Doug Camplejohn
(06:55)
I was probably similar, not long ago. Jeff Weiner, CEO of LinkedIn, had a phrase: “Seek first to understand.” He was excellent at taking a beat and asking, “Doug seems upset—what’s really the reason?” It often wasn’t what’s on the surface. I’ve tried to apply that. Let’s take a tangent—you mentioned SalesLoft. SalesLoft-Clari merger thoughts?
Jeremey Donovan
(07:13)
There are things I can’t comment on. I can’t speak to anything related to portfolio companies. In general, on consolidation: people want to consolidate their tech stacks overall. AI is accelerating that because you get more benefit from consolidating conversation intelligence, deal intelligence, CRM data—more data means more benefit from crossing silos and providing guidance. Consolidation is a no-brainer.
The key pieces are the usual suspects: account and contact data, meeting intelligence, forecasting and pipeline management, sales engagement. The next piece is post-sale. Sales engagement companies may argue they do CS use cases, but they’re not really there yet. Their capabilities aren’t at the same level of discipline as post-sale folks. That will come together. There’s always the push and pull between best-of-breed and full-suite solutions. There’s always going to be a best-of-breed tech that’s a bit better. We’re in a wave of consolidation now.
Doug Camplejohn
(09:28)
Do you see an advantage to build vs buy, from a data standpoint?
Jeremey Donovan
(09:35)
Build internally vs buying consolidated platforms?
Doug Camplejohn
(09:42)
In general—look at the players in the space. Gong has taken a “build ourselves” approach. Clari has acquired. Also, people build their own stuff.
Jeremey Donovan
(09:58)
In addition to my regular job, I teach a course at University of Chicago Booth for CROs. I asked the recent cohort about their use of AI. Everyone’s using it—they were confused about “agentic” use cases. They gave standard use cases, but most are DIY-ing: building their own.
Buy vs build, I come at it from the economic point of view. Back at SalesLoft, valuations were 20x or 30x or more; so acquiring wasn’t practical except in a few cases. We acquired a conversation intelligence company and a deal technology company, but had to do big codebase rewrites. We gained distribution, brand, knowledge, and iterations from their teams. That worked, but we mostly had to build because buying wasn’t viable.
Now, there are many companies with great tech who struggle with distribution. Right now, I’d probably advise companies to buy—with the understanding that the thing you’re buying may not be ready to sell, but the benefit is mainly the people with experience and iterations. Coding is much faster now with modern tools. I say buy and optimize is probably better now.
Doug Camplejohn
(13:16)
Funny, there was a post on LinkedIn: “Incumbent CRMs are dead, someone coded a CRM in 48 hours.” I thought, “Good luck.” You can get something fast, but getting something deep takes a long time.
Jeremey Donovan
(13:29)
Getting something deep takes a long time. Don’t underestimate UX iterations or the importance of distribution. For instance, why did Zoom beat GoToMeeting and WebEx? Why did Slack beat HipChat? These are fundamentally similar, but UX is just better. Sometimes it’s tech evolution—like HTML5 launching when Zoom came out. Still, UX matters.
There’s another key in acquisition strategy: culture fit. I’ve seen acquisitions go well and others poorly, even on the same day with deals of the same size. Culture fit made all the difference.
Look at John Chambers’ book (maybe “Connecting the Dots”). Cisco’s strategy: way more likely to succeed when acquiring a much smaller company. That strategy works far more than mergers of equals. For SalesLoft, we picked up very small conversation intelligence and deals companies. At the time, that was definitely right. Merging similar-sized companies is much harder.
Another takeaway: Cisco identified acquisition targets by talking to customers about tech they were buying/using or tech Cisco was losing to. If CIOs/CTOs were excited about or buying a company, Cisco bought them. It’s simple, but not many are that disciplined.
Doug Camplejohn
(17:13)
It’s not rocket science, but few do it well. I have to mention—at LinkedIn when Microsoft acquired us, Satya did a brilliant job. They said, “We’re just going to let you run things.” You couldn’t tell you worked for Microsoft except your shares said Microsoft and you had to use Outlook.
Jeremey Donovan
(17:37)
Exactly.
Doug Camplejohn
(17:39)
And I could get discounts on Xboxes—that was it. Nothing else changed.
Jeremey Donovan
(17:42)
There’s probably a lot of people who don’t even know LinkedIn is owned by Microsoft.
Doug Camplejohn
(17:47)
It’s true. What they did well was put Jeff in charge of integration, which got tested multiple times. Microsoft discounts heavily selling the suite, but when I ran Sales Navigator, no discretionary discounting.
Jeremey Donovan
(17:56)
Yeah.
Jeremey Donovan
(18:11)
I remember trying to negotiate a discount on Sales Navigator and was unsuccessful. There was zero competition—no reason to discount.
Doug Camplejohn
(18:18)
Right. Mike Gamson, running sales, pushed for no discounts. Eventually, a small piece was added, like with the recruiter tool, but it was minimal.
Jeremey Donovan
(18:30)
Exactly. At Gartner, I was there 16 years. In 2004, a new CEO asked “Why are we discounting?” By then Gartner had won decisively; Forrester was a distant second. CEO told salespeople to stop discounting. I don’t know if that’s still true, but there was no discounting for years after that.
Doug Camplejohn
(19:16)
Forty-two years in dog years. You see a ton of portfolio companies at all stages. When do you advise bringing on RevOps?
Jeremey Donovan
(19:17)
We’ve looked at the data—usually around 10 sellers. Before that, there’s usually an outsourced Salesforce admin, or someone like an SDR/AE with proficiency who does the work. Around 10+, they hire a RevOps person.
First RevOps hire is usually a Salesforce admin—tactical, not strategic. Second hire lays the foundation for strategy, analytics is key, and they’re a jack-of-all-trades. With every 10–15 more sellers, you add another RevOps. Eventually, you get a commissions person, territory person, and more. At the big companies, you have “RevOps partners” embedded with sales leaders.
Doug Camplejohn
(21:18)
At LinkedIn, working for a big company spoiled me—then I started a startup. LinkedIn had a BizOps team I called “business Siri” (pre-TVT): ex-McKinsey and consulting. You could ask them deep questions about new logos, existing customers, or critical trials—they’d research and come back. As a startup guy, I was amazed. How do you see this strategic ops role evolving in the AI era?
Jeremey Donovan
(22:26)
I’d split predictive ML/AI from generative. Predictive is more useful there. Interface is a bit of both—agentic means you’ve given it authority to act elsewhere (email, CRM, etc.), not just source info but as system of action. You could ask these questions in natural language, but the core is the ML.
It makes sense to embed data scientists inside RevOps. The “go-to-market engineer” term is mostly a tool user, not a data scientist. Smaller companies can’t afford a dedicated data scientist in RevOps—they tap data science from engineering. On a call with one company, for example, the CRO, head of RevOps, and a data scientist worked together on churn and retention. Data scientist is key for modeling, prioritizing accounts, and triggering CSM plays from health scores.
Larger companies might have embedded data scientists in RevOps. Smaller ones tap into engineering expertise.
Doug Camplejohn
(25:40)
Great point: people think AI equals LLMs, but it’s broader. Where do most companies you talk to have their data relative to what’s needed for real data science work?
Jeremey Donovan
(26:01)
Short answer: not good enough. Many companies are still doing foundational data work so it’s usable—garbage in, garbage out is especially true in ML. Super low-hanging fruit is secure filling out security questionnaires—they cut that from days to minutes using their vendor’s AI capability. RFP vendors do that too; good tech available. Harder to DIY that than other things.
Other data is often not good enough for advanced work.
Doug Camplejohn
(27:52)
Which vendor is helping with security reviews or RFID?
Jeremey Donovan
(27:56)
It’s not a portfolio company—I think it was Vanta.
Doug Camplejohn
(28:05)
Vanta, yeah. GTM Fund is all about SOC 2 compliance and ISO, so I expected that.
Doug Camplejohn
(28:30)
At my last company, Fliptop, we did account lead scoring. After connecting hundreds of Salesforce instances, the only data we could trust was when a rep got paid—the deal probably closed. At least one contact was probably correct, dates and size were probably correct. Data quality dropped off dramatically after that; anything relying on manual entry is challenging.
Jeremey Donovan
(29:11)
Yeah, they’re not going to bother.
Jeremey Donovan
(29:20)
Tools to solve that have existed for a while; people.ai’s claim to fame was collecting more data. Conversation intelligence companies like Gong do a good job getting data into your CRM. There’s a new crop: Momentum, Attention, others—better at getting data into the system.
Doug Camplejohn
(29:59)
One of our founding principles was you can’t have good AI with bad data. We automated contact creation/enrichment, captured conversation and email (unstructured data), and noticed most companies start with a forms-over-database architecture (HubSpot, Salesforce, Pipedrive).
Jeremey Donovan
(30:05)
Yeah.
Doug Camplejohn
(30:30)
If you change a stage, old values disappear. At scale, gluing Snowflake or Databricks enables historical tracking. From day one at Coffee, we knew that becoming core data. It isn’t obvious to a customer until they want to compare pipeline over time—like Clari and Gong do, showing pipeline deltas. Unlock happens quarters later when asking ops/rev ops questions: “How did I do this quarter versus the same point last year”?
Jeremey Donovan
(31:25)
Stage changes have been my bane. Generative AI helps—we use it to improve operational efficiency. For diligence in new investments, pipeline data is often messy. We have a partner tool that automaps stage changes, asks for confirmation, and uses AI to estimate likely mappings. Eyeballs still needed. As Amanda Callow from One Mind said, human augmentation is shifting. More and more we trust AI to do autonomously—but at least for big checks, still want people involved.
Doug Camplejohn
(33:05)
Recently someone asked if I’d trust AI to send emails for me. No, not yet—I want a human in the loop. But I trust a Waymo or Tesla autopilot on the freeway; surely B2B software will improve.
Jeremey Donovan
(33:29)
When generative stuff first started, my holy grail was auto-drafting email responses in my voice and style. Maybe it exists now and I just don’t know. Someone mentioned SuperMe, which supposedly could do this by training on all your prior content. I’d love it to not just get my tone but my structured, analytical approach.
Doug Camplejohn
(34:29)
I’ve heard Fyxer does that decently. Haven’t tried it deeply. Coffee has basic abilities to draft emails from transcripts, but it’s early. I’m all about brevity and bullet points, but some want very detailed messages. How do we get more structure and voice in, templatize not just field mapping, but set a general structure? I see lots of opportunity there.
Jeremey Donovan
(35:07)
It should learn. You can go to an LLM and say, “You are John McMahon—how would you answer?” He’s the CRO from PTC, then Blade Logic, BMC, and his team went on to Zscaler, AppDynamics, Snowflake, Rubrik, etc. Lots of content—so you can prompt LLMs to answer as John McMahon. Some talk about having an advisory board of personalities for answering questions. That’s valuable.
Doug Camplejohn
(36:07)
You see interesting queries across portfolio companies. What are the most common ones?
Jeremey Donovan
(36:20)
They’re very seasonal. Lots of annual planning questions now—organizational design, moving jobs between SDRs, AEs, AMs, renewal specialists, CSMs, technical account managers, etc. Planning ratios: how many RevOps, SEs, CSMs per org size/ARR, etc. Comp plans—most AEs have a simple 50-50 comp plan with 4–5x quota on OTE, but companies tend to create crazy complexity. I coach simplicity.
Right now the biggest question is how to get growth and efficiency out of AI. Honestly, I don’t think anyone has a great answer for sales. CS may have a stronger answer. CEO of Dekagon, on a podcast, put it well: sales hasn’t realized AI’s potential yet. People are a bit disappointed; I need to think more about why. Could be data, could be that above a certain dollar threshold—
Jeremey Donovan
(38:42)
You need to interact with a human and it has to feel personal, in business or life. Recently, I replaced my roof—a big ticket purchase. I bought from the person I trusted. The most important thing is when something goes wrong, they’re there. That was key. The salesperson was upfront about fixing mistakes. After the job, he called about venting that should be changed, and did it for free. Amazing.
Doug Camplejohn
(39:47)
Right, he’s your roof guy for life now.
Jeremey Donovan
(40:03)
And I’ll recommend him. When he came, he saw our lawn sign still up. That’s how much I value them. Who would have thought a roof would be my best recent purchase? Shows that for big purchases, you want a human. That threshold rises over time: it could be $100/month, maybe $1,000, maybe $10,000—at some point with a big spend, you’ll talk to a human.
If you’re spending $100,000 at work, you’re going to talk to a human.
Doug Camplejohn
(40:50)
Absolutely. Amanda from 1Mind has a spicy take: we’re all saying jobs are safe, but she’s not. Things will change, especially in customer success as AI takes over more. Layoffs happening there, but I don’t see it in planning conversations yet or in sales headcount reductions. Maybe there’s attrition, reduced hiring. Any impact you’re seeing?
Jeremey Donovan
(41:30)
I agree with Amanda. Reason you don’t see it as much: Regular LLMs make everyone 10% more effective, but don’t fully replace us. If you have five CSMs, you can’t cut half a person. At big companies with 100 people, you can cut 10. So, bigger impact on big companies—obvious, but needs stating. That’s why you don’t see it in smaller companies; AI isn’t fully taking over jobs outside of customer support, maybe.
One thing that drives me nuts is people saying, “AI won’t take your job, people who learn AI will take your job.” It’s low empathy—doesn’t acknowledge people can’t control the tsunami headed their way. Can’t blame customer support folks for getting hit.
Yes, it’s making people more effective. One of my kids is starting law school—I’m a little fearful, since legal tech will make things more efficient. Bigger law firms won’t need as many lawyers; research and writing will be faster.
Doug Camplejohn
(43:45)
Especially on the research side, yeah—my youngest daughter is the same. The idea of sending interns to look for case law seems crazy.
Jeremey Donovan
(43:57)
Some programs hire interns mainly to lock them in for future offers, not for actual immediate work. There will be less need for that work.
Doug Camplejohn
(44:24)
We’re coming up on time, Jeremy; could talk forever. Let’s do a speed round. What do you like to do for fun outside work?
Jeremey Donovan
(44:39)
Fly fishing—super addicted fly angler for 25 years.
Doug Camplejohn
(44:42)
Awesome.
Doug Camplejohn
(44:46)
What’s another thing we’d be surprised to know about you?
Jeremey Donovan
(44:53)
Super cat lover. I have four cats. I love all animals—a dog is too much responsibility. Cat lover is my other passion.
Doug Camplejohn
(45:07)
Last one—I’m a product geek, you’re a product geek. What’s a personal product you love or admire?
Jeremey Donovan
(45:18)
What do I use regularly? No super insightful answer—maybe silly, but I hated seltzer my whole life. Suddenly addicted. Can’t go a day without seltzer and non-alcohol beer. Quit alcohol post-COVID, so now non-alcohol beer—mainstream brands like Peroni and Stella, they’re all good.
Doug Camplejohn
(45:49)
Do you have a favorite?
Jeremey Donovan
(45:58)
Stella, Peroni, all make good ones.
Doug Camplejohn
(46:01)
Last question—how can listeners stay in touch and support your journey?
Jeremey Donovan
(46:08)
LinkedIn is best. I wish LinkedIn would lift the 30,000 connection limit—I have to delete to accept new connections, which I hate. Connect and I’m there multiple times a day; if someone messages me, I’ll respond.
Doug Camplejohn
(46:36)
Fantastic. Thanks for your time, Jeremy. Great conversation. Have a great day.
Jeremey Donovan
(46:40)
Thanks so much. You too, take care.
