50 rules for a SaaS demo that converts, backed by the data

This is for SDRs, founders and SaaS sales teams who run the same demo 5, 10, 20 times a week and want a higher close rate from it.

Every rule below is tied to a number. Gong's analysis of 3 million web demos and 67,149 recorded demo calls. Consensus's 6 million buyer interactions. Navattic's 40,000 interactive demos. Gartner and Forrester's buyer surveys. HubSpot's rep surveys. The year sits next to each figure, because benchmarks drift and a dated number ages better than a vague one.

You'll get 50 rules in 10 groups, from the discovery call to the demo-to-close report. And where a rule changes when the demo is recorded instead of delivered live, I say so, because that's the direction most SaaS demos are heading.

Your demo, on repeat. Webinar Flow is built for exactly the demo this post describes. Record it once. Prospects book into a session that runs like it's live, ask questions mid-playback, and your phone buzzes so you can answer in chat or pause the recording and go live on camera. Every answer trains the AI to handle the same question next time. You only show up when it counts.

Discovery before the demo

The demo is decided before it starts. These 6 rules cover what happens between "book a demo" and the first shared screen.

1. Diagnose before you demo. Gong's study of 67,149 recorded demos found that winning demos mirror the topics raised in discovery and open on the customer's highest-priority problem (2017, refreshed 2026). Peter Cohan calls the alternative "buying it back": showing capabilities the prospect never asked for, then spending the rest of the deal convincing them they don't have to pay for those (2025). Skip discovery and you've skipped the demo. If the calendar won't allow a separate call, run 10 minutes of it at the top.

2. Ask 11 to 14 questions on the discovery call, and spread them out. Across 519,000 B2B calls, Gong found success peaks at 11 to 14 targeted questions and drops to average beyond that, and that top reps pace their questions evenly instead of front-loading a checklist (2017). A 2025 refresh on 326,000 calls moved the winning count to 15 to 16 questions, with losing calls at around 20. The point holds: enough to diagnose, few enough to stay a conversation.

3. Ask about the business problem, and phrase it so a one-word answer is impossible. Gong's discovery research is blunt: one-word responses to discovery questions are the worst outcome, and "Can you help me understand your biggest challenge?" pulls a longer answer than "What's your biggest challenge?" (2021). Sort the pain into latent (they haven't accepted it's worth fixing) and active (they're already looking). The demo for each is different.

4. Go easier on executives, and bring them in at the third touchpoint. Asking senior executives a long list of discovery questions correlates strongly with losing the deal (Gong, 2019). And on more than 1 million executive sales cycles, win rates drop roughly 6% when an evaluation starts with an executive and rise 5% when the executive joins around the third touchpoint (Gong, 2026). Demo to the champion first. Bring the exec in once you know what they'll ask.

5. Respond to the demo request inside 5 minutes. The 2007 Lead Response Management study by James Oldroyd, across 15,000 leads and 100,000 call attempts, found the odds of qualifying a lead drop 21 times between a 5-minute and a 30-minute response, and the odds of even reaching them drop 100 times. Harvard Business Review's 2011 audit of 2,241 US companies found firms that responded within an hour were nearly 7 times as likely to qualify the lead, while the average responder took 42 hours and 23% never replied at all (figures summarised here, 2026). For a recorded demo, this rule collapses to zero: with Webinar Flow's just-in-time sessions, the prospect books the next 15-minute slot and is watching your pitch before an SDR would have opened the email.

6. Book the soonest slot, send 3 reminders, and confirm who's coming. GrowthSpree's 2026 show-rate benchmarks put same-day demos at 78 to 88% show rate versus 22 to 35% for demos booked 14 or more days out, and a 24-hour, 1-hour and 15-minute reminder cadence lifts show rate from 38 to 48% (no reminder) to 68 to 78%. Ziellab's 2026 analysis lands in the same place: same-day meetings no-show around 7%, 8-plus days out past 23%. Calendly found 36% of revenue teams name no-shows as a top blocker (2024). Then confirm attendees, because Gartner puts a complex B2B buying group at 6 to 10 decision makers, each arriving with 4 or 5 pieces of information they gathered on their own, and Forrester's 2025 Buyers' Journey Survey counts 13 internal stakeholders and 9 external participants on the average purchase. You need to know which 2 of them are on the call.

The first 3 minutes

Prospects decide whether to lean in or check Slack in the first 3 minutes. Five rules for the opening.

7. Open on their problem, in their words. Gong's analysis of 3 million web demos says to start with the priority-ranked problems from discovery, never with the product (2019, refreshed 2026). 2Win's Tell-Show-Tell framework makes the opening "Tell" do 3 jobs: title the topic in the customer's language, connect it to their pain, and preview what to watch for (2026). "You said the handoff between sales and onboarding loses about a week. Here's what that looks like fixed."

8. Two minutes on your company, then stop. Gong found that time spent on the "about us" section is fine up to 2 minutes, and after that there's a sharp drop in win rates (2017, refreshed 2026). One slide, or none. They booked a demo to see the product, and your funding round can wait for the case study. (Sorry.)

9. Set the pace in the first 3 minutes. Top performers get prospects to adjust their rate of speech by 13% on average within the first 3 minutes, against 7% for average reps (Gong, 2018). Match their tempo for a minute, then lead it. It's the difference between a call that feels like a meeting and one that feels like a broadcast.

10. Lose the slides. Slide-free discovery calls are 17% more likely to book a follow-up, and winning intro meetings average 9.1 minutes of presentation against 11.4 for losing ones (Gong, refreshed 2026). If you must have a deck, keep it for the recap. Our 100 sales presentation tips covers what to do when slides are unavoidable.

11. Camera on, from the first second. Deals where the seller sells on video have 94% higher win rates, buyers on video 96% higher, and any video in the process makes the deal 127% more likely to close (Gong, refreshed 2026). Storylane's customer Respond.io saw demo engagement go from 3% to 6% after adding presenter videos with faces and voices, with A/B testing still to confirm it. For a recorded demo this is the rule that gets skipped most: people record a screen capture with a voiceover. Record yourself on camera, picture-in-picture, exactly as you'd appear live. More on that in 100 sales video tips.

Show the outcome before the features

The single biggest structural mistake in SaaS demos is building up to the good bit. Five rules to invert it.

12. Do the last thing first. Peter Cohan's Great Demo! rule: show the end result, the finished report or the closed loop, before any setup, because buyers make "a very rapid decision: does this look interesting?" the moment they see it (2025). His analogy is a cookbook: the photo of the dish comes before the recipe. Gong's demo data agrees: winning demos open on the part the buyer cares most about instead of saving it for a finale (2017).

13. Run every topic as Tell, Show, Tell. Context, capability, impact. 2Win names the anti-pattern the "Show-Tell-Tell crime": the SE shares the screen, opens the product and starts narrating each feature as they go (2026). The closing Tell is one sentence of operational impact the buyer can repeat to their boss. Cohan's caveat applies: only run Tell-Show-Tell on capabilities that discovery confirmed they want, or you're structuring a feature tour very neatly (2025).

14. Talk business, cut the feature tour. Top reps spend 52% more time on business topics and 39% less time on features than average reps (Gong, 2018). Gong's shortlist of deal-killing mistakes puts feature dumping at the top: reciting generic feature after generic feature, at length, makes win rates plummet, and the effect gets worse the longer the demo runs (2019).

15. Never prove ROI. Let them say the number. Presenting ROI in a sales call correlates with a 27% drop in close rate, and ROI language in cold email cuts success by 15% (Gong). The buyer doesn't trust your spreadsheet; they trust their own arithmetic. So ask: "How many hours a week is your team on this today?" and let them multiply it out loud.

16. Social proof: late, specific, or never. Generic social proof lowers close rates by 22%, and by 47% when it's used in early-stage calls (Gong). "Lots of companies like yours use us" reads as a red flag. A single named customer, same industry, same problem, dropped in after they've raised the objection it answers, is a different thing entirely. For a recorded demo, that named story becomes a seeded prompt fired at the minute the objection usually surfaces.

Tailoring and personas

Generic demos lose to tailored ones by a wide margin in every dataset. Five rules on how far to go.

17. Build one version per persona. Walnut's 2025 platform data found teams that personalise 50% or more of their demos see 40%-plus higher conversions than generic templates (2026). The same company's customers run an average of 39 interactive demo templates each (2025). Three will do: one for the economic buyer, one for the daily user, one for the technical evaluator, plus the discipline to pick the right one at booking.

18. Use their data, or data that could be theirs. Cohan's practical advice is to pre-stage captured, linked screens with the prospect's terminology and numbers so the "last thing first" moment looks like their business (2025). Demostack lists realistic demo data as a foundational element, with the warning that lorem-ipsum accounts and "Test Company 1" break the illusion instantly (2025). Rename 3 records before the call. It takes 4 minutes.

19. Say their industry and their vocabulary in minute one. Salesforce's research on 7,775 sales professionals found 73% of customers expect companies to understand their unique needs (2022), and its buyer data shows 59% of business buyers say most sales professionals don't take the time to understand them (via ZoomInfo, 2026). Use their job titles, their team names, their word for the thing. If they say "onboarding" and your product says "activation", you say onboarding.

20. Assume the demo will be forwarded to people you'll never meet. Prospects view an interactive demo 3 times without the rep present, look at 3.4 demos per deal, and more than 3 stakeholders view each one (Walnut, 2025). Consensus's 2026 report goes further: an average of 4 previously unknown stakeholders surface per demo, and those discovered stakeholders engage at 61% versus 28% for the original recipient. Every demo needs a version that works with you absent. That's the whole logic of a recorded session with an instant replay link.

21. Bring the sales engineer for the technical buyer. Sales engineer involvement in demos lifts win rates by up to 30%, closed-won selling teams are 67% larger than losing ones (Gong, 2025), and calls with more than one seller are 258% more likely to close than solo calls (Gong, 2018). For a recorded demo, this is the clearest case for the live jump-in: the SE's 5 minutes on the integration question is the part you can't pre-record.

Navigating the product on screen

Screen-sharing is where good demos go to die. Six rules for what happens once the product is up.

22. Stage the environment and test it an hour before. Cohan's pre-opened tabs versus captured, linked tabs trade-off: open tabs are interactive but look clumsy when you switch; captured tabs look real and never break (2025). Demostack's list includes testing the demo beforehand and sharing client insights with whoever's presenting (2025). Noota's SaaS demo checklist puts customising the demo environment ahead of anything about delivery (2023). Log in, load the right account, close the 14 other tabs, mute Slack.

23. Show 2 or 3 workflows end to end. Prospeo's software demo guide says show only 2 or 3 features and finish each one (2026). Guideflow's 2026 list says show only the features that solve their stated problem. Gong's data on focused demos: top performers receive nearly a third more questions from buyers during the demo while asking 30% fewer themselves (2018). A finished workflow is something the buyer can picture doing on Monday. Half of 6 workflows is a blur.

24. Keep each workflow to 5 to 13 clicks, and each explanation to about 25 words. Navattic's 2026 State of the Interactive Product Demo, from 40,000-plus demos, found the top performers run 5 to 13 steps with 25 to 30 words per explanation, and demos split into multiple short flows have 48% higher completion than single long ones. The same constraints work on a live screen-share: if a workflow needs 20 clicks, either you're showing the wrong workflow or the product has a problem the demo can't fix.

25. Say the click before the click, and cap each "show" at 5 minutes. 2Win's original Tell-Show-Tell guidance: keep the screen segments under 5 minutes, because people associate simplicity with brevity and time-consuming things with complexity (2015, updated 2024). Gong's version is the 9-minute rule for the main product demonstration inside a longer call (2019). Narrate what you're about to do, do it, then say what it changed.

26. Stop sharing your screen for the conversation. Prospeo's list includes stop sharing your screen when the demo moves to discussion, so faces come back (2026). Gong's data explains why it matters: in successful demos the rate of speaker switches climbs 36% in the second half, and a shared screen with a mouse moving on it suppresses that (2017). Show a workflow, stop sharing, talk, share again.

27. Have a plan for when the product breaks. Claap's 2026 guide has a whole section on what to do when the demo breaks: acknowledge it in one sentence, switch to the backup (captured screens, a recorded clip), and keep going. Demostack calls it delivering with flexibility. For a recorded demo this rule disappears, which is one of the underrated arguments for recording: the pitch never lags, never 500s, never shows the wrong tab. The only live variable left is you, when you choose to jump in.

Talk ratio and questions

Gong has more data on this than anyone, and most of it contradicts how demos are usually run. Six rules.

28. Aim for 65:35 on a demo, 46:54 on discovery. Gong's 67,149-demo study found the winning demo talk-to-listen ratio is 65:35 rep-heavy, against 46:54 for discovery calls (2017). The older "golden ratio" across all calls was 43:57 (2016), and the 2025 refresh on 326,000 calls found won deals at 57% rep talk time against 62% for lost. Demos are allowed to be rep-heavy. They're not allowed to be monologues.

29. Never pitch for more than 76 seconds without a break. Gong's line: "We didn't find a single demo that led to a closed deal that involved more than 76 seconds of uninterrupted pitching" (2017). Build the break in. A question, a "does that match how you'd do it?", a pause long enough to be uncomfortable. For a recorded demo you get to design every one of those breaks in advance: a seeded prompt or poll at 1:10, 2:20, 3:30, so the viewer is nudged to type just as they'd start to drift.

30. Engineer the back-and-forth to climb. Successful demos have 21% more speaker switches per minute, and the rate rises 36% in the second half (Gong, 2017). Losing demos start as a pitch and stay one. Winning demos start as a pitch and turn into a conversation. Put your best question at minute 15, not minute 2.

31. Answer objections with a question, then pause. Top reps respond to objections with a clarifying question 54.3% of the time against 31% for average reps, pause 5 times longer after an objection, and slow to 176 words per minute against 188 (Gong, refreshed 2026). The clarifying question buys you the real objection; the pause tells them you took it seriously. We've written out 75 sales objections and how to answer each one live.

32. Seed the questions that close. Buyers who ask 3 or 4 pricing questions correlate with the highest win rates; fewer than 3 or more than 4 and win rates decline (Gong, 2017). Read that as a to-do: the questions that close a deal are predictable (price, integration, security, "who else uses this"), and a good demo makes sure they get asked. In a live call, you raise them yourself. In a recorded session on Webinar Flow, seeded prompts and polls fire at set moments and do the same job, and the real questions typed underneath get a real answer from you, by phone, in chat or on camera.

33. Use risk-reversal language, and listen for "probably". Talking about opt-outs, guarantees and SLAs lifts win rates by 32% on average (Gong, 2017). The same study flags "probably" from the buyer as a positive timing signal and "we need to figure out ____" as a negative one. Matthew Dixon's JOLT research on 2.5 million sales conversations found 40 to 60% of qualified deals are lost to no decision, and in roughly 56% of those the buyer wanted to change and froze (2026 summary). Taking risk off the table is a demo skill. Practise it before the contract stage.

Pricing and next steps

The end of the demo decides whether there's a second one. Five rules.

34. Talk price on the first call, around minute 38 to 46. Discussing pricing on the first call produces win rates 10 points higher (42%) than waiting for call 2 (32%) or call 3 (15%), across 11,331 opportunities, and star reps bring it up in the 38-to-46-minute window (Gong, 2020, refreshed 2026). Discussing budget on the first call wins 49% against 7% when budget never comes up. The buyer wants it too: HubSpot found 58% of prospects want to talk price on the first call, while only 23% of reps are prepared to (2016, updated 2025). For a recorded demo, put the pricing section at the same 75%-through mark and make it a timed offer card, then keep negotiation live.

35. Never say "list price". The phrases "list price", "typical price" and "standard price" extend sales cycles by 19% (Gong), because each one signals a discount is coming if they wait. Losing demos also spend 8% more time talking price than winning ones (2017). Say the number, say what it includes, stop talking.

36. Spend the last 10 minutes on next steps. Winning demos devote 12.7% more time to next steps (Gong, 2017). The fastest-closing deals spend 53% more time on next steps in the first meeting, and skipping next steps on a first call makes the close rate plummet by 71% (Gong, 2019). Ten minutes sounds like a lot. It's the 10 minutes that get the second meeting.

37. Book the next meeting on the call, with a date. Once you're in a live sales cycle, a specific CTA with an exact time and date beats an interest-based CTA, the reverse of what works in cold outreach (Gong). Calendly's list puts it as rule 10 of 11: schedule next steps before the end of the call (2024). "I'll send some times" is where deals go to think about it.

38. Multi-thread before the fourth meeting. Gong's 1.8 million-opportunity study: 77% of deals involve more than one buyer contact, won deals have twice as many contacts as lost ones, and multi-threading lifts win rates 130% on deals over $50K (2025). Group calls with several buyers on the line nearly double win rates when they happen on the fourth meeting versus the first (Gong, 2021). And SMB deals without a decision maker are 80% less likely to close, enterprise deals 233% less likely. Use the demo replay as the multi-threading tool: "Who else should see this? I'll send them the session link."

Demo mistakes that kill deals

Some of these overlap with rules above, deliberately. These are the 5 that show up most often in lost-deal reviews.

39. The harbor tour. Cohan's name for the demo with no discovery behind it: "The prospect boards the boat, is driven around the harbor for three hours while continually being asked, 'Have you seen anything you like so far?'" (2024). Juniors give it because they don't know better. Seniors give it when sales hands over nothing. Either way the fix is rule 1, or a short vision-generation demo that leads straight into discovery on the same call.

40. Demoing people who were never going to buy. Consensus's presales research found 30% of demos delivered are unqualified (2025), and its 2026 buyer report costs a manual demo at $288 to $330 per session. SaaStr's rule of thumb: if fewer than 8 to 10% of demos convert to paid, you'll burn out the sales team. Qualify at booking. Or give the unqualified 30% a recorded session and save the live hour for the other 70%.

41. Treating "I need to think about it" as the end. That phrase doesn't hurt win rates, but it stretches the cycle by 173% (Gong). JOLT's research says doubling down on the cost of doing nothing backfires 84% of the time; what works is a recommendation, a shorter shortlist and less risk (2024). "Think about what, specifically?" is the question. Then offer them a path.

42. Going dark after the demo. Won deals average 8.21 emails a week between seller and buyer against 1.87 for lost deals (Gong). Invesp's numbers via HubSpot: 80% of sales take 5 or more follow-ups, and 48% of salespeople never follow up at all (2026). And there's a cliff: deals that close within 50 days of the demo win 47% of the time; past that, win rates fall under 20% (Puppydog, 2026, 939 companies). Send the replay within the hour. Our async selling playbook has 45 ways to keep a deal moving without another live call.

43. Running a group demo like a 1:1. Chris Orlob's line, collected by Everything Design: 90% of salespeople are terrible at group demos (2026). With 6 to 10 decision makers each holding their own 4 or 5 pieces of research (Gartner), the person who booked the demo is rarely the person who'll block it. Ask every attendee what they want to see in the first 2 minutes, then run the demo in that order. And expect more than 3 of them to watch the replay without you (Walnut, 2025).

Recorded and self-serve demos

The buyer research points one way: more of the demo happens without a rep on the line. Four rules for the recorded version.

44. Record it, because that's how they want to buy. Gartner's buyer surveys moved from 61% of B2B buyers preferring a rep-free experience (June 2025) to 67% in the 2026 survey of 646 buyers, with 45% using AI tools during a recent purchase. Gartner also puts buyer time with suppliers at 17% of the journey, and 5 to 6% with any one rep. HubSpot's sales trends data says 90% of prospects are already informed before the first call and 51% of companies now offer free trials (2025). The rep-free preference covers the research phase; the purchase itself still has you in it. Give them the demo on their schedule, and be reachable the moment they have a question. That combination is what Webinar Flow was built to run.

45. Build the recorded demo in chapters, and let it run long. Vidyard's 2024 benchmark on 943,305 business videos found 65% of viewers finish videos under a minute and 20% finish videos over 20 minutes. But Wistia's 2026 State of Video, on 13 million videos and 79 million hours, found webinar replays of 31 to 45 minutes had more than twice the engagement of replays under 30 minutes, and 30-to-60-minute videos got the most clicks. Buyers spend 6.5 minutes on an interactive demo (Walnut, 2025) and 5 minutes 14 seconds on a Consensus demo (2026). So: a 30-to-40-minute session built from 5-to-6-minute chapters, each one a complete Tell-Show-Tell, with a menu so the CFO can skip to pricing.

46. Decide what to record and what to keep live. The 90% of a demo that's identical every time (the problem framing, the 3 core workflows, the pricing walkthrough, the named customer story) is what you record. The 10% that changes (their integration, their security review, their negotiation, the SE's deep dive that lifts win rates by up to 30%) is what you keep live. On Webinar Flow, the split is built in: the recording plays, a real question pings your phone, you type a reply or pause the session and go live on camera, then the recording resumes. The rule for jumping in personally: any question about their setup, their data or their price. Everything else, the AI has probably seen before.

47. Ungated for reach, gated for leads, and decide which one you want. Storylane's 2026 study of 535,763 demo sessions across 50 companies: ungated demos get 19.4% engagement against 9.2% gated, but 43% of engaged gated users convert to leads against 0.3% for ungated. Navattic's data adds that 71% of top-performing demos are ungated and run 10% higher engagement (2025), and Walnut's customers found ungated self-serve demos drove 3 times more qualified pipeline than demo-request forms (2025). A booked session is a soft gate: an email and a time slot. It captures the lead without the form fatigue, and the show-up tells you how warm they are.

Measuring demo-to-close

If you don't know your number, every rule above is guesswork. Three rules for the reporting.

48. Know your demo-to-close benchmark by segment. Puppydog's 2026 benchmark from 939 B2B companies: SMB 32%, mid-market 25%, enterprise 18%; under $10K ACV 35%, $100K-plus 15%, with 30% as the top-quartile entry point. Optifai's cut of the same data: 25% B2B average, 30% for SaaS, 38% where the buyer controls the demo (2026). SaaStr's older founder-stage guidance is 10 to 20%, with 20%-plus meaning the engine works. LevelUp Demo's 2026 range is 15 to 30%, with 30%-plus as high. Pick the row that matches your ACV, then measure against it monthly.

49. Track show rate, engagement and questions asked, not just closes. The industry median demo show rate is 55 to 65%, with the top quartile at 75 to 85% (GrowthSpree, 2026), and RevenueHero's benchmark puts no-shows at 20 to 40% on average, under 15% for high performers (2025). For recorded demos, Navattic's published ranges are 20 to 30% click-through, 35 to 45% engagement, 40 to 50% completion (2025). And two Consensus figures worth putting on a dashboard: buyers who engage with 9 or more demos close at over 55%, and demos with 7-plus stakeholders attach to an opportunity 36% of the time against 19% with none (2026). Questions asked per session is the metric I'd add. It's the one that predicts whether the replay gets forwarded.

50. Review the recordings every week, and let the answers compound. Reps who use AI frequently generate 77% more revenue than those who don't, across 7.1 million opportunities (Gong, 2025), and Gong's AI briefer users saw win rates up 42% and deal duration down 26%. Storylane's 2026 report on 110,257 sessions and 150 deals found prospects who engaged with a demo converted 3.2 times more often and closed 6 days faster (33 days to 27). HubSpot's 2025 survey of 1,000 reps found only 8% use no AI at all. The compounding version of this rule: every answer you give to a real question should make the next demo better without you re-answering it. That's what Webinar Flow's AI does with your transcribed answers, per customer, so the 50th session is sharper than the first.

Put the demo on repeat

Fifty rules is a lot to hold in your head on a live call at 4pm on a Thursday. The honest fix is to stop delivering the identical 90% live.

Record the demo once, with rules 7 to 27 baked in. Let prospects book into a session whenever suits them (rule 5 becomes automatic). Seed the questions that close (rule 32). Then be on your phone for the 10% that needs you, and let the AI handle the questions it's already seen.

Start free with Webinar Flow and run your next demo without being in the room for it.