AI Landing Page Lab — customer validation kit

Date: 2026-08-27 Status: Ready for use after EXTERNAL_OUTREACH approval

Experiment

Target only individually known or permissioned contacts. Do not send bulk messages.

• Segment: people responsible for a new B2B offer or landing-page production at a startup, operating company, or web production company.
• Trigger: they have a service document, proposal, or existing service URL but no campaign-specific LP ready for review.
• Promise: turn one supplied source into a grounded, editable PC/SP LP preview; unsupported claims are visibly excluded.
• Test unit: one 15-minute observation session or asynchronous preview trial.
• Primary behavior: the participant supplies or nominates a real source document for their own preview.
• Revenue signal: the participant selects one export option without being prompted toward the highest price.

30-second demo storyboard

1. 0–5s — Show a source PDF entering the product. Caption: “資料を入れる”.
2. 5–10s — Show acquisition route, audience, CTA, proof, and tone choices. Caption: “売る前提を選ぶ”.
3. 10–16s — Show three headlines and the grounded proof field. Caption: “言えることだけで構成”.
4. 16–22s — Show three genuinely different visual directions. Caption: “色違いではない3案”.
5. 22–27s — Switch the editable preview between PC and SP. Caption: “Webコードが正本”.
6. 27–30s — End on: “あなたの資料でも無料Previewを1本作りますか？”

Do not claim conversion improvement, time reduction, or customer adoption unless separately evidenced.

Individual invitation draft

新しいサービスや施策のLPを作る際に、資料の読み込みから訴求・コピー・デザイン・編集可能コードまでを一続きにする試作を検証しています。30秒のデモを見て、もし今お持ちのサービス資料で試せそうなら、無料Previewを1本作らせてください。営業連絡ではなく、15分以内の使用感確認が目的です。合わなければ返信不要です。

Before sending, replace the generic first sentence with one factual reason the recipient fits. Never infer confidential work or scrape personal contact details.

Observation prompts

Ask in this order without explaining the intended answer:

1. “最初の10秒で、何を作るサービスだと思いましたか？”
2. “普段この作業は誰が、何を使って、どのくらいの頻度で行いますか？”
3. “このPreviewのどこまでなら実際の案件で任せられますか？”
4. “任せられない部分は、品質・根拠・編集性・社内承認のどれですか？”
5. “次の案件で使うなら、どの受け取り方を選びますか？”

Do not ask “便利ですか” or treat compliments as demand.

Price card shown after the trial

• Preview only — ¥0: watermarked browser preview, one direction.
• Editable export — ¥9,800 per LP: HTML/CSS/JS, PC/SP, one revision instruction.
• Multi-angle export — ¥24,800 per LP: three FV/message variants, brand settings, decision log.
• Team plan hypothesis — ¥39,800/month: five exports, shared brand settings, history. Show only to teams with repeated monthly need.

These are hypotheses, not published prices. Do not collect payment before payment/contract approval.

Evidence ledger

Record one row per participant:

Never store unnecessary personal information in the ledger.

Decision rule after five valid contacts

• Continue: at least 3 watch the demo, at least 2 nominate a real source, and at least 1 selects a paid export option.
• Iterate positioning: demo completion is at least 3 but real-source nominations are 0–1.
• Iterate product: at least 2 nominate a source but all reject export because of the same product limitation.
• Stop or pivot wedge: fewer than 3 complete the demo after five relevant, permissioned contacts.

Views, compliments, and “looks useful” are not counted as demand.