Section 1
Key takeaways
• A demo is not judged in the room; it is judged later, when a champion tries to re-sell it internally from memory. Structure for the retell, not the live reaction. • The serial-position effect means the middle of a long, feature-dense demo is forgotten by default, so put your highest-value point at an opening or closing Tell, never buried mid-show. • The closing Tell should compress to a "Key Operational Impact" of three words or fewer : a phrase short enough to survive the trip to the next meeting. • Narrating while you demonstrate (dual coding) beats showing visuals alone by a median 89% on apply-it tests across Clark and Mayer's eleven studies . • Break one long walkthrough into several short Tell-Show-Tell blocks; segmenting instruction into learner-paced chunks lifted both retention and transfer in a 56-study meta-analysis .
Section 2
Why does a "thorough" demo lose to a shorter one?
Start with what a demo actually is. It is not a performance you are graded on live. In most service and software sales, the person watching your demo cannot sign alone. They are a champion, an internal advocate who has to go re-sell your value to a committee, a partner, a CFO, a skeptical operations lead. The demo's real audience is the people who were never on the call. That reframe changes everything about structure. You are not optimizing for "did they nod while I talked." You are optimizing for "can they reconstruct my argument three days later with no slides in front of them." And human memory is brutally selective about what survives that gap. Glanzer and Cunitz's serial-position work is the clearest evidence. In one study, 240 Army enlisted men were shown lists of twenty common one-syllable nouns and asked to recall them . Recall was high for the first few items (primacy) and the last few (recency), and collapsed in the middle. More telling for demos: when subjects had to count backwards for thirty seconds before recalling, a small distraction, the recency advantage vanished completely . The last thing you showed only sticks if nothing interrupts before encoding. A demo that ends on "any questions?" and then drifts into logistics has already let the ending decay. This is why the feature-dump backfires. The more items you cram into the middle, the more of your demo lands in the forgettable zone, and the more you dilute the two positions, opening and closing, that memory actually protects. As the demo-skills firm 2Win! Global puts it bluntly: "The 2% factor in competitive deals isn't talent. It isn't timing. It's whether the value was framed clearly enough to be remembered and repeated inside the buying organization." That is the whole game. Not whether your product is better in the room, but whether your value is portable out of it.
Section 3
The opening Tell: an advance organizer, not a warm-up
The first Tell is not throat-clearing. It is a cognitive instruction. In one sentence you name what you are about to show and why it matters to this specific buyer, before a single pixel moves. "I'm going to show you how a new client goes from signed contract to fully onboarded without anyone on your team touching a spreadsheet, watch how few steps it takes." That sentence does real work. In cognitive terms it is a signal, or advance organizer: a cue that primes attention so the brain knows what to attend to before the visual arrives. In a controlled experiment with 138 university students, adding signals and cues produced a significant small-to-medium gain on transfer tests, the apply-it tests, eta-squared of .05 at p = .01, though notably not on rote retention . The signal does not help people memorize trivia. It helps them understand and reuse the point, which is exactly what a champion needs to do. Concretely, for a service business demoing a client portal: do not open the screen and start clicking. Say "the thing I want you to notice here is that your client never emails you for a status update again, every milestone is visible to them in real time." Now the prospect watches the show looking for that, instead of getting lost in your navigation bar. The framing comes first because the framing is what they repeat. This is the same discipline that makes a discovery call qualify rather than pitch, you set the frame before you show anything, so attention lands where you want it.
Section 4
The Show: narrate while you demonstrate
The middle is where most demos go quiet, the founder clicks around while the prospect reads the screen. That is a mistake the science is unusually clear about. When learners get words plus pictures instead of words alone, the median improvement on transfer tests is 89%, with a median effect size greater than one across the eleven studies in Clark and Mayer (2016) . The mechanism is dual coding: words and visuals encode through separate channels, and when they arrive together they reinforce each other instead of competing. A silent screen-share is words-absent. A screen-share where you narrate what is happening as it happens is dual-coded, and dramatically more likely to survive into the retell. But narration has a failure mode: reading the interface aloud. "So I click here, and then this dropdown opens, and I select this option." That is not narration, that is a screen-reader. Effective narration names the meaning, not the mechanics: "Notice I didn't have to re-enter the client's details, the system already pulled them from the contract, which is the step that used to eat twenty minutes per client on your team." You are pairing the visual (the filled form) with the words (the twenty minutes saved). That pairing is what encodes. Keep the Show tight. One capability, demonstrated cleanly, narrated for meaning, in roughly one to four minutes . The instinct to show three more things "while we're here" is the instinct that fills the dead zone.
Section 5
The closing Tell: compress the outcome into a Key Operational Impact
The closing Tell restates the outcome, and this is where most demos waste their single most protected memory position. The recency slot is where the message survives, but only if you give it something compact enough to carry. 2Win! Global's framework calls this the Key Operational Impact, or KOI: a phrase of three words or fewer that captures the operational result, designed to become the thing the prospect repeats internally . Not "this will improve your onboarding efficiency and reduce manual data entry across your operations team", that is a sentence no one carries into the next meeting intact. The KOI is "zero onboarding spreadsheets." Or "clients self-update." Or "twenty minutes back." Short enough to survive the hallway, the elevator, the three-days-later recap email. Here the serial-position evidence cuts both ways. Recency protects your closing Tell, but only if nothing interrupts before it encodes . So deliver the KOI and stop. Do not immediately pivot to pricing logistics or "let me also show you the reporting module." The thirty-second distraction that erased recency in the lab is the same distraction you create when you bury your closing line under housekeeping. Land the KOI, let it sit, then move on. The way you frame and anchor that outcome is the same muscle that makes a price feel like an obvious yes rather than a number to negotiate.
Section 6
Why one long demo loses to several short ones
There is a structural decision above the individual Tell-Show-Tell unit: how much to cover in one block. The answer from the research is to chunk. A meta-analysis of segmenting, breaking instruction into short, learner-paced units rather than one continuous stream, found significant gains in both retention and transfer and reduced cognitive load across 56 investigations and 88 pairwise comparisons (retention d = 0.32, transfer d = 0.36, both p < 0.001) . Translated to demos: three short Tell-Show-Tell blocks beat one long fifteen-minute walkthrough, even if the total content is identical. Each block gets its own protected opening and closing position. Each block lets the buyer's working memory reset before the next load arrives. For a service business, that means structuring the demo around the buyer's two or three real pains, not around your product's menu. Block one: onboarding. Tell, Show, Tell. Pause, let them react or ask. Block two: reporting. Tell, Show, Tell. Pause. Block three: the renewal moment. Tell, Show, Tell. Three clean memory anchors instead of one undifferentiated blur. If you've mapped those pains correctly upstream, the blocks practically write themselves, which is why demo structure and handling objections before they surface are the same body of work.
Section 7
The BGA framework: Tell-Show-Tell (Context, Capability, Impact)
Here is the structure as a repeatable build. Run it once per buyer pain, not once per demo. 1. Write the opening Tell as a signal (30–90 seconds). One sentence: what you're about to show and why it matters to this buyer. Name the thing to watch for. Rule of thumb: if your opening Tell could be reused verbatim for any prospect, it is too generic, it should reference their specific pain. The signal is what lifts transfer . 2. Show one capability, narrated for meaning (1–4 minutes). Demonstrate exactly one thing. Narrate the meaning of each action, not the clicks. Pair every visual with the words that explain its value, that is the dual coding that nearly doubles apply-it recall . If you catch yourself saying "and I can also…", stop. That belongs in a different block. 3. Close with a Key Operational Impact (30–60 seconds). Compress the outcome into three words or fewer . Deliver it, then say nothing for a beat. Do not chase it with logistics, the interruption erases the recency that makes it stick . 4. Chunk the full demo into three or fewer blocks. One Tell-Show-Tell unit per buyer pain. Pause between blocks. Resist the fourth block; segmenting helps because it limits load, and a fourth block reintroduces the load you just removed . 5. Pressure-test the retell, not the demo. After the call, ask your champion: "If your CFO asks what you saw, what do you say?" If they can hand back your KOIs unprompted, the structure worked. If they reach for the feature list, you over-showed. This is the only test that matters, because the retell is the actual sale. A metric to track: count how many of your KOIs come back in the prospect's own follow-up email or in their internal-forward language. Zero means your value never left the room. The tooling that captures and routes that follow-up signal sits in the systems that keep a deal warm after the call, structure gets it remembered, systems make sure the memory gets acted on. If you want the Tell-Show-Tell structure as a build-ready asset, the Template Pack gives you fill-in blocks for the opening Tell, the narrated Show, and the closing KOI per buyer pain, and the deeper mechanics live in the ConvertOS playbook.
Section 8
You're running Tell-Show-Tell right when…
You're running Tell-Show-Tell right when your demo is shorter than it used to be and your close rate is higher. When you can name, before you share your screen, the two or three KOIs the buyer should walk away repeating, and you've cut everything that doesn't serve them. When you narrate meaning instead of clicks, and you catch yourself stopping after the impact line instead of barreling into pricing. When your champion's follow-up email contains your three-word phrases instead of a vague "it looked good." And when, asked what you showed, the person who wasn't even on the call can repeat your value back accurately. That last one is the only scoreboard. The demo was never about the room.