The setup: POCs are the most expensive stage nobody talks about
Enterprise SaaS funnels get optimised aggressively at the top — and then deals quietly stagnate at the POC after 12–18 months of nurturing. It’s the longest, most expensive stage in the funnel, and the one most under-invested in relative to what it decides.
The numbers, roughly:
- Enterprise sales cycle: ~10–18 months, conservatively.
- POC duration: 4–8 weeks — and the longest single stage of the cycle.
- Cost of a single POC: conservatively $35k, often $60k+, once you count AE, SE, CS, enablement, infrastructure.
- Median CAC for six-figure ACV deals: nearly $2 spent for every $1 in revenue.
- 40–60% of pipeline dies at “no decision.” Not lost to competitors. Not lost to bad product-market fit. Lost because value wasn’t shown well enough to force a decision.
Put a price tag on every POC. Kaminarov’s CRO-hat advice: when you view a POC as a $50–70k investment, your team’s posture toward qualification, kickoff, and wrap-up changes. It’s not another stage. It’s a make-or-break moment where everything upstream is either converted or lost.
The biggest myth: “the POC is a technical validation”
Both speakers hammered this. The technical win matters — table stakes. If the product can’t solve the problem, nothing else matters. But teams that stop there lose over and over to no decision.
“The technical win is a precursor to the business win, often preceding it by five or six weeks. The technology and the business are intertwined like DNA — you need both strands to win.”
— John Care
Kaminarov’s framing: the POC is not a validation exercise, it’s an alignment exercise — alignment on value, impact, what the customer really needs to move forward, who has budget, who’s the executive sponsor, what “success” even means. Technical success and commercial success are separate wins. Both must be achieved.
“Do Nothing Incorporated” is your biggest competitor
Care’s line: ~42% of enterprise deals are won by DNI — Do Nothing Incorporated. When you run competitive analysis, add a column for “no decision” alongside your standard competitors. It’s usually easier to pull a deal away from no-decision than from a rival vendor — because you’re competing against inertia, not against features.
Consensus building, landing, and expanding inside the account are how you actively beat DNI. Treat it as a competitor. Prepare against it.
The 6 killers of a POC
The six patterns they see across dozens of engagements a year:
No executive sponsor / economic buyer
If no exec is engaged before the POC begins, you’re running a technical exercise, not a buying process. Kaminarov’s rule as a CRO: run a full MEDDPICC pass before entering any POC. If the champion can’t identify an exec who’ll approve budget on a successful outcome — don’t enter.
No alignment on objectives, success criteria, scope
The classic kickoff-call failure: everyone dives into screens and configuration before agreeing on what “success” means. Also fuels scope creep — when a customer keeps adding features to test, it usually means they don’t know what would count as a decision. POCs then extend indefinitely and die.
Generic environment
In 2026, there’s no excuse for showing a customer a demo of somebody else’s data. Stand up an environment that mirrors their industry, their vocabulary, their edge cases. Every time the customer has to translate “widget” into their real domain, you lose momentum.
No POC management
Care’s pick for the deadliest one. No project plan, no owner, no cadence, no visibility. If you’re not doing this, you kind of deserve to lose the POC. POCs need who-does-what, checking cadence, in-product analytics, escalation paths, and shared status — internally and with the customer.
Zero visibility (engagement + usage)
Two flavours. Engagement visibility: a shared Slack/Teams channel where questions get answered in hours, not days. Product-usage visibility: analytics inside the POC environment so you can see whether the customer is actually running the agreed use cases. If they’re not, you engage before the POC runs out of momentum — not after.
No POC wrap-up (or a bad one)
Some teams don’t do a formal read-out. Others do one, but let the champion see the deck for the first time on the day. Best practice: dry-run the wrap-up with your champion first, gather testimonials from POC users, present against the success criteria you defined at kickoff, and land on a firm next-step advancement. No soft advancement.
The pre-POC alignment framework
The killers are systematic, not technical — so the counter has to be systematic too. The framework the panel proposed is a five-piece checklist. Two must happen before POC entry; the other three run through it.
Before kickoff, the customer fills out a structured questionnaire covering their objectives, use cases, stakeholders, current-state pains, and success criteria. Two effects: you get the context you need, and you filter out tire kickers who won’t engage. If they refuse to fill it out — that’s a signal.
A branded, customer-specific kickoff deck presented on the first call. Contents: objectives, measurable success criteria, stakeholder map (see next), timeline, scope, operating cadence. Kaminarov shared that they generate these with a Claude prompt tuned to their design system — standardises the artifact so no team ships an ad-hoc kickoff again.
Three people mapped on each side: vendor side — account executive, solutions engineer, executive sponsor. Customer side — business champion, technical champion, executive sponsor. The more of these six seats are filled and speaking to each other, the less deal risk. Any empty seat is a risk to name and mitigate.
Confine the POC to the specific use cases aligned to the objectives. When a customer asks to test another feature mid-POC, the honest answer is: “That’s post-deployment work. Testing it now would prove nothing about your original success criteria.” Hard conversation. Non-negotiable.
The read-out is a designed event, not a debrief. Dry run with the champion. Present against the success criteria you defined at kickoff, tick by tick. Include user testimonials gathered during the POC. Close with a firm action plan and named next-step milestones. Never accept a “we’ll let you know.”
The best story in the transcript — and why it matters
Care told a story from his time as a CIO evaluating programmer-productivity software. Three vendors ran POCs. Scores: 74, 70, 58. The 58 was out. In a normal universe, you pick the 74.
They picked the vendor that scored 70. Half a million dollars.
Why? Because the winning vendor did better discovery upstream. They understood that the CIO’s company had a product launch on a critical path, that IT infrastructure delays were about to slip that launch by eight weeks, and that the delay would cost $8M. During the final read-out, they tied their features directly to that delay: “By doing this, you save 2.5 weeks. By doing this, you save another 12 days.” By the end of the presentation they’d saved eight weeks of the schedule. What a coincidence.
“It’s not always the best product that wins. It’s the people who link the tech back to solving the business problem.”
— John Care
Kaminarov added the Challenger Sale data point: research across millions of sales processes shows 53% of the purchase decision has nothing to do with the product and everything to do with the customer’s experience of the evaluation itself. The prospect isn’t just evaluating your product — they’re evaluating you as a partner.
Other worth-keeping details
- Why is every POC 30 or 60 days? Because the AE didn’t think. If the actual work fits in 14 days, make it 14. Care has seen a one-day POC succeed — because that was all the customer needed to see impact.
- Who owns the POC? Care’s answer: AE and SE joined at the hip. AE owns the frame and the kickoff, hands off to SE for the technical execution, AE steps back into the driver’s seat for the wrap-up. The SE is the person the customer trusts most — use that trust.
- Skin in the game. The panel disagreed. Care sees paid POCs in ~single-digit-percent of enterprise deals. Kaminarov sees the trend growing, often in the $10–15k range. Kaminarov’s clever structure: sign a 12-month framework agreement with a 3-month opt-out — the customer is paying, but they can walk if the first three months don’t hit criteria. Trades harder MSA negotiation upfront for faster close-out at the end.
- Benchmarks. Win rates for POCs vary wildly with product complexity — from ~30% for simple products to ~92% for complex, highly-differentiated ones. The number to aim for is 10% higher than yours today. Meter it, know why you win, know why you lose.
- PLG doesn’t remove the POC. For simple SMB use cases, a free trial is enough. For complex enterprise deployments — custom integrations, security audits, real data volumes — you still need a structured POC on top of whatever PLG motion you have.
What we take away for VeehiveLabs
Every enterprise AI engagement we run has a POC or discovery-sprint shape. The failure modes above are the ones that quietly kill our deals too. Five principles we’re internalising:
A working demo is table stakes. The deal is decided by whether we’ve aligned the customer’s exec sponsor, champion, and stakeholders on value and a firm next step. If we only shipped a technical win, we’ve done half the job. Design every discovery sprint to close both wins.
On every pursuit we should have three named competitors: the actual rival vendor(s), and DNI — the client just not deciding. Our discovery, kickoff, and wrap-up all have to be built to defeat inertia specifically. Assume the buyer’s default action is to do nothing until proven otherwise.
We adopt Kaminarov’s CRO rule verbatim: no discovery sprint enters a build phase without a named economic buyer and measurable success criteria signed off by the customer’s champion. If the customer resists this, that’s our signal to walk — or to demote the engagement to something cheaper than a full sprint.
Every AI system we demo or POC uses the customer’s data, their vocabulary, their edge cases. Not sample data. Not a Fortune 500 case study in a different industry. If we can’t get real data yet, we synthesize it against their glossary — but nothing generic ships to a decision-maker.
Every discovery sprint ends with a scheduled read-out with the exec sponsor. We dry-run it with the champion first. We present against the success criteria we agreed at kickoff, tick by tick. We include quotes from the users who ran with us. We land on a firm next step with a named milestone and date. No “we’ll get back to you” endings.
Open questions we’re holding
- How much of this transfers cleanly to custom AI builds, where the “product” is being built during the sprint rather than evaluated? Our POCs are less about validating an existing product and more about de-risking a new build. Success criteria still apply. Success measurement is harder.
- Should we adopt paid POCs by default for six-figure engagements? The framework agreement + opt-out structure is elegant. Worth piloting on the next enterprise sprint that qualifies.
- How do we measure our own POC win rate cleanly given how varied our engagements are — MVP builds vs. team augmentation vs. consulting? Care’s answer applies: meter it, and aim for “10% higher than today.”
Source: Rev Genius × Demo Stack webinar — Why your POCs are set to fail — and the framework to win them. Featuring John Care (Mastering Technical Sales), Gilad Kaminarov (3x CRO), Noam Harel (Head of Growth, Demo Stack), hosted by Jared Robin (Rev Genius). Notes compiled by the VeehiveLabs team; corrections welcome.