Match your product demo’s promise to visible proof

A cursor moves. A dashboard appears. A headline promises a result. Does the recording actually show that result? Use this product demo video audit checklist to turn vague feedback into a specific, reviewable edit brief.

Updated October 16, 2026 · 8 min read

Download the free claim-to-proof CSV worksheet. No email required. Use it with a video you own or have permission to review.

What should a product demo audit answer?

For a creative strategist or small agency, the useful output is an edit decision the producer can find and understand. “Make the demo stronger” does not identify a shot. “The narration says sent, but the recording ends on a draft; show the sent confirmation or change the line” does.

Define the audience, placement and one task before reviewing. A homepage visitor deciding whether to try a product needs different detail from an existing customer learning an advanced setting. Record the exact source filename and version. An observation from yesterday’s cut may no longer apply to today’s export.

Wistia’s product-page video guidance distinguishes explaining a product’s purpose from demonstrating its operation. That distinction makes a useful review question: does this scene explain a feature, show someone using it, or show the outcome? A polished interface shot can do the first without doing the other two.

1. Replay the first three seconds

Watch the opening at normal speed, first with its intended audio and then muted. Do not infer a spoken promise from a silent still. Write four short observations:

• Opening promise: copy the visible wording and separately note any spoken line you actually heard.
• Visible proof: identify the product state or action present in the image. A logo establishes identity; it does not establish the claimed outcome.
• Transition: describe how the opening leads into the demonstration. Record approximate ranges for navigation, then replay to check actual cuts.
• Audible beat: note a relevant spoken cue or sound only after listening. If sound was not reviewed, leave it Not reviewed.

These are human review fields. AdScanVideo may provide useful candidate observations, but this checklist does not imply that the product generates a numeric Hook Scorecard. See the first-three-seconds hook review for a broader opening worksheet.

2. Separate action from outcome

Write what changed before and after the action. “Clicks Export” describes an action. “A downloaded file opens with the expected content” describes a visible outcome. If the clip cuts away after the click, the outcome remains unconfirmed from this recording.

Check continuity too. Does the result belong to the same input? Does a time jump conceal a required step? A shortened workflow can be appropriate, but an edit should not imply that a slow or conditional operation happened instantly. If timing matters to the promise, preserve a continuous reference or document how the duration was measured.

3. Work through one concrete example

The following invoicing example is constructed editorial teaching material, not a real analyzed customer video or an AdScanVideo report. Its ranges are invented for the exercise, not measured cut boundaries.

A practical review row would say: Visible action checked; sending outcome Unclear. Proposed edit: show the actual sent confirmation, or change the promise to creating a draft. Timing claim: Not reviewed until measured. This does not say the product cannot send invoices. It says this cut does not establish it.

Make the hypothesis testable: “Will showing the confirmation help intended viewers correctly identify the completed task?” Ask a few representative reviewers what happened before explaining it. Record their answers and the tested version. Comprehension feedback is useful, but it is not a measured conversion lift.

4. Use decisions instead of an unexplained score

For each check, choose Confirmed when the reviewed evidence answers the specific question, Unclear when you reviewed it but the evidence does not settle it, or Not reviewed when the check has not happened. A blank cell is not approval.

Then choose a separate action: Keep, Revise or Obtain evidence. For example, a clearly visible unsupported headline can be Confirmed as wording while its promised business result remains Unclear. Do not collapse those into one reassuring number. This rubric is editorial and belongs to the worksheet; it is not an automated product rating.

5. Use AI analysis to find places to inspect

1. Upload the permitted source file and choose Ad intelligence for an ad-oriented first pass. Choose Shot breakdown when you need visual scene notes.
2. Read the actual returned report. Keep its wording and time reference in a separate field from your own replay notes.
3. Replay important ranges in the original with context on either side. Native video analysis can miss or misinterpret brief actions and audio; timestamps are navigation aids, not frame-exact edit instructions.
4. Write the evidence gap and an editorial edit hypothesis. Do not present your hypothesis as generated output.
5. Review the revised export and record whether the requested change is visible. Leave campaign results unmeasured until there is appropriate performance data.

The public animated-cat shot report shows the kind of descriptive source notes a report can contain. It is a shot-description example, not proof that a product demo converts. For your own demonstration, use the report returned from that actual clip.

Download and use the product demo worksheet

The free CSV includes eight blank review rows: opening promise, visible action, visible result, workflow continuity, spoken claim, performance claim, CTA and revised export. It stores source/version, candidate range, generated excerpt, human replay notes, decision, evidence gap, proposed edit and recheck result. It is a manual planning document, not an analyzer export.

Duplicate rows when a demo has multiple claims. Keep one claim per row so a producer can fix one issue without losing the rest. Use the creative brief workflow to turn the reviewed rows into production instructions. Use the report-sharing guide when preparing the handoff.

Check the placement and measure the next step

Google Ads documents different video ad formats, so review the intended placement rather than treating every video as the same viewing experience. For a short demo, inspect mobile readability and the final destination in that placement. Our placement review checklist helps record those checks without promising a universal safe zone.

After launch, keep retention and qualified downstream behavior separate. For a SaaS demo, that might mean completed trials, useful reports or verified purchases. An export click is not a payment. Record the version, reporting period and denominator; aggregate referral traffic cannot establish that one video caused a sale.

For a complete worked example, replay the licensed product-demo source beside its actual Ad intelligence report. The sample preserves uncertain product identity, absent audio and an approximate-duration discrepancy rather than presenting the output as verified performance evidence.

Frequently asked questions