Revenue Builders · 27 Aug 2026 · From the week of 24 August
Why Great CROs Stop Managing Numbers and Start Building Systems with Sam Costello
These are notes on the conversation, checked against its transcript. The episode itself has the full discussion.
In brief
Sam Costello, CRO at Harness, joins hosts John McMahon and John Kaplan to discuss how he moved from carrying a number to building the go-to-market machine that produces revenue, as Harness grew from under $1M to $100M ARR. The episode covers the champion-first Duplo sales process, multi-champion technical deals, proof-of-value gating, MEDDPIC reviews run as hunts for gaps, and interviewing for coachability. The central argument is that revenue is an output of disciplined recruiting, process and culture, and that qualification has to be triangulated rather than checked off.
For founders
- Sam Costello argues that a large deal is pulled in by a champion who has gained confidence, so the champion and business case should be built before the deal grows.
- Sam Costello says a champion built for a smaller deal may need to be replaced as the opportunity grows, because the right champion needs the influence to answer criteria, process and paper questions.
- Sam Costello says Harness gates POVs on at least one champion, agreed success criteria, economic buyer alignment and known competitors, and that ideally each POV is tied to a business case with a before-and-after metric.
- Sam Costello describes the ingredients of great companies (a great founder, a big market, differentiated technology and inspiring leadership) as hard to replicate, and as the reason he wants to keep building inside a company like Harness.
- Sam Costello says Harness hires for coachability by reading a candidate's career trajectory from the start of their resume or LinkedIn, rather than relying on the interview story.
For revenue leaders
- Sam Costello says Harness runs its sales process by finding a champion first and then doing everything else, and that a typical deal needs two to three champions across the platform, developer and security teams.
- Sam Costello says the economic buyer should be seen before a POV, and that the POV should be tied to a business case with a before-and-after metric.
- Sam Costello says MEDDPIC reviews should be hunts for gaps rather than compliance exercises, with at least one action from each review shared between the leader and the rep.
- John Kaplan says reps should ask what the customer's biggest business issue is on every deal, and Sam Costello says new AI committees and AI addenda can slip deals late in the quarter.
- John McMahon says MEDDPIC elements should be triangulated as the story of the opportunity rather than checked off, and John Kaplan adds that a deal being in Salesforce does not mean it is qualified.
What was said 23, most useful first
Sales process at Harness starts with finding a champion, then doing everything else Listen
Sam Costello calls Harness's process the Duplo sales process, after the bigger, simpler version of Lego made for young children: find a champion, then do everything else. Reps are taught to identify a champion, build one into a champion, and test during the sales process whether that person is who they appear to be. Pain, economic buyer access and the rest of the process follow from that starting point.
“find a champion and then do everything else”
A champion for a smaller deal may not be the right champion for a much larger one, so the champion has to be reassessed as the deal grows Listen
Sam Costello describes a trap where a champion built for a $500,000 deal is still relied on when the opportunity has grown to $5 million. That champion may lack the influence to answer questions on decision criteria, process and paper, so the team gets stuck at the end of the deal. He says the team needs to find a new champion with the right influence.
“the champion you built was for a $500,000 deal, but now you're trying to do a $5 million deal”
A technical platform deal usually needs two to three champions across the platform, developer and security teams, and without mindshare in each the deal gets stuck Listen
Sam Costello says Harness sells a technical platform of 14 products, so it usually has to win two to three champions per deal. These are a platform team that built the tooling, a developer or engineering team consuming it, and a security team with standards and governance it wants followed. He says that without technical mindshare in each of those places, the deal will get stuck.
“we must go win usually two to three champions to do a deal”
About 60% of Harness's proof-of-value work is now installed in the customer's environment, with the rest run as workshops that show features without installation Listen
Sam Costello says Harness always ran POVs in its early years because it was the only way to build mindshare and confidence. With about 1,000 customers and a Magic Quadrant leader position, roughly 30 to 40% of engagements are now workshops, and about 60% of the time the team installs the technology in the customer's environment to prove it solves the complex problem there.
“like 60% of the time we're going in and installing the technology in there to prove that it works in their environment”
A POV should only go ahead when there is at least one champion, defined success criteria, economic buyer alignment and known competition, including the customer doing nothing Listen
Sam Costello lists the checks Harness runs before a POV: whether there is at least one champion and ideally several, whether the POV criteria are well defined and the team is highly differentiated, whether the economic buyer would discuss changes if the criteria changed, and whether the team knows its competitors. He says the global SE leader and he both sign off on each POV, and that competition is often a customer choosing to do nothing.
“do we have champion, at least one, but multiple champions ideally?”
Tie the POV to a business case with a before-and-after metric so each technical result maps to a business outcome Listen
Sam Costello says the ideal POV is tied to a business case that states the value of solving the problem, even if it is not fully baked. The POV is then designed to prove the points in that case, with a metric such as how long a task took before and how long it takes now. He describes creating this linkage as the most important step.
“it used to take this long to do it, now it takes this long”
See the economic buyer before running the POV to confirm the problem, the criteria and the straw-man business case Listen
Sam Costello says that before the POV, the team should check with the economic buyer whether this is the problem they care about, whether these are the criteria they want to see, and whether they have faith in the straw-man business case, which may only need to be directionally accurate. He also asks whether other people need to be involved.
“is going to see the economic buyer before you do the POV”
Asking what the customer's biggest business issue is on every deal tends to produce technical answers, so reps need to be pushed back to the business question Listen
John Kaplan says that when he asks anyone inside a selling company what the customer's biggest business issue is, he usually gets a technical answer, and the question has to be reframed repeatedly to reach a business issue. He suggests asking it on every deal and counting how many times it has to be reframed, and thinks it is getting worse partly because of the focus on AI. Sam Costello adds that it has never been easier to solve, since an LLM can identify a company's biggest business problems.
“Ask on every deal that you have. Just start off with what's the biggest business issue.”
AI committees and AI addenda in buying processes can slip deals that looked committed, so reps need to track how each buyer's process has changed Listen
Sam Costello says buying processes now include AI committees that review all investments, and AI addenda that may be needed on existing contracts. As an illustration, he describes how a deal forecast in commit, with an MSA already in place, could slip if purchasing finds no AI addendum with two weeks left in the quarter and sends it into a legal process. He says the team triangulates with peers to find such pitfalls, though it is still difficult.
“there's now an AI committee that wants to review all the investments”
MEDDPIC reviews work as hunts for gaps, and a rep who leaves a review not excited means something did not go right Listen
Sam Costello says in a compliance-style review the rep only tries to prove they already have what the reviewer wants so they can get off the call, while in a hunt the team assumes there is something it does not know and looks for what could go wrong. He says if a rep does not leave a MEDDPIC review excited, something went wrong, since the review should uncover gaps. He adds that if a rep follows a compliance process and still loses, the process loses mindshare.
“If it's a compliance effort, then you're just trying to prove that you already got what that person wants”
No single MEDDPIC element is the most important, because the qualification comes from triangulating pain, economic buyer, competition and criteria as one story Listen
John McMahon says he has refused for about 30 years to name the most important MEDDPIC component, because any single element is irrelevant to a unique deal. He triangulates whether the pain is big enough, whether the economic buyer will care at that level, what the competition is and whether the decision criteria are differentiated. He says this is how the pieces work together as the story of the opportunity, not a checkbox.
“It's how it all works together as the story of the opportunity.”
Coachability can be assessed by reading a candidate's career from the start of their resume or LinkedIn, looking for times they were coached and followed direction Listen
Sam Costello says Harness teaches leaders to read candidates bottom-up on their resume and LinkedIn, because trajectory predicts who someone is better than the interview story. He looks for coaching in any setting, including sports, debate or band, and whether they followed direction. He then asks what their favorite thing about a leader they worked with was, and about a deal they lost and what they learned, looking for ownership.
“what is the deal you're most proud of?”
A large deal is pulled in by a champion who has gained confidence, not pitched to the company as a big idea from the start Listen
Sam Costello says that when reps find pain, build a champion, reach an economic buyer and build a business case, the company and champion start to pull them into a large deal. He calls pitching a large deal early a trap. Harness is comfortable landing where it makes sense, delighting the customer and expanding from there.
“they start to pull you into a large deal. It's not the other way around.”
When a champion will not introduce a rep to the economic buyer, the usual cause is a lack of confidence on the champion's side Listen
Sam Costello says that when a champion resists taking a rep to their boss, it is always down to confidence, because a champion who is sure of the deal is happy to show the rep to their executives. He treats that reluctance as a signal that something is unresolved and needs to be found and solved. When a host tied this to weak discovery, Sam agreed and added that a skipped step in the sales process leaves the champion unsure.
“it's always down to confidence. They lack confidence.”
Reps often find pain the company does not rank highly, and quantifying it depends on the metrics the buyer actually tracks Listen
Sam Costello says reps are trained to find pain, but the company or champion may not care about it, or it may sit outside the company's top priorities, so the team has to wait for it to rise. Quantification usually depends on getting the metrics the buyer measures that pain by. A host added that a lack of trust with the champion is often why reps cannot get that information, and Sam agreed.
“I actually think a lot of times it's because haven't built any trust with the champion.”
Early on, a champion deposit is education and awareness, which builds enough trust for the team to ask for information in return Listen
Sam Costello describes a give-get dynamic in which champions are out in the market trying to learn the best way to solve their problem. Giving them education and awareness early builds some measure of trust, after which the team can start getting the information it needs to understand whether there is something to solve.
“champion deposits are usually education and awareness”
A technical coach is not the same as a true champion, because a true champion has influence and access across the organization Listen
Sam Costello says a technical contact who can validate the product may not be a true champion. A true champion understands what is happening in the business and is in the room where problems are discussed, and can link the solution to a problem the economic buyer has said they solved, even without tying it to NPS or revenue. If the contact cannot make that link, he says, they are likely a technical coach.
“you're not talking about a true champion because a true champion has influence and access across the organization”
Ask leaders at non-competing software vendors that sell to the same buyers for real-time information on where those buyers are in their processes Listen
Sam Costello says Harness calls leaders at non-competitive software companies to learn what is playing out in shared accounts, since those firms may be weeks ahead in a buying process. When a host noted that such vendors still compete for funding, Sam called that fair but said that in the current environment knowing where you stand and what the pitfalls are matters more, and that there is funding for projects tied to the right things.
“That's more important than like, will they steal my funding?”
Priorities at large buyers can shift within days, so deal assumptions need frequent testing and a cadence with the economic buyer, who sees changes first Listen
Sam Costello describes a large financial institution in the middle of an unexpected $60 million data center cost overrun, where the priorities changed in about 10 days. He says that what the team knew about a company even three or four weeks earlier could now be materially different. The economic buyer knows what is happening at the top first, so the team needs cadence with them to learn early.
“but what we're seeing is priority shifting really fast.”
Each MEDDPIC review should leave at least one action, with the leader taking part of the work instead of leaving it all to the rep Listen
Sam Costello says that when he runs a MEDDPIC review he always leaves with at least one action and a play of his own, such as looking at pricing or unpacking the business case, so reps do not feel they must do all the work alone. He circles back in about two weeks to see where the deal stands. A host added a leader's line that it is not a crime not to know an answer in a review.
“I'll always have at least one action on it and my play that I'm going to go do for that”
Just because a deal's details are filled in Salesforce does not mean it is qualified, and AI that automates CRM fields can give faster answers from unqualified data Listen
John Kaplan says many AI companies are trying to automate whatever sits in Salesforce to give faster answers. He says a deal being in Salesforce does not mean it is qualified, and illustrates by saying he would ask 20 questions of a rep who claims a $2 million deal with a champion. He calls this what AI slop means to him in sales.
“just because it's in Salesforce doesn't mean it's qualified”
Staying at one company long enough to build something matters more than frequent moves, and job-hopping every 18 to 24 months limits what a seller learns Listen
Sam Costello says he wants to stop people moving across companies every two years, arguing that building a company is very hard and that the skills and knowledge gained over time propel people furthest. He says that if you skip jobs every 18 to 24 months you are probably not going anywhere. A host added that it is the company's responsibility to offer a reason to stay.
“can we stop moving people every two years across companies and can we actually stay somewhere and build something?”
In the talent war, a company has to offer candidates more than money, such as building something durable and developing people Listen
Sam Costello says the current talent war involves OTE and guaranteed compensation from companies with strong product-market fit, which is different from the earlier hiring war when VCs were flush with cash. He argues that a company needs to offer more than money, such as building a durable company, a leadership tree and a culture of radical candor. A host recommended asking people at QBRs to say why they are there, and Sam agreed that knowing their why matters for attracting and keeping people.
“if you don't have your value proposition down, and you can't honestly say, why are you here?”