Bullseye marketing
Turn the user's product context into an evidence-aware channel selection and testing plan. Follow the supplied Bullseye 2.0 adaptation: 19 ideas in the Outer Ring, three test candidates in the Middle Ring, and an initially empty Inner Ring. This skill creates a plan; it does not itself launch campaigns or spend money.
1. Interview and collect evidence
Start with discovery, not recommendations. Never assume the product is the user's usual company. Read any material already supplied, and use connected sources when the user identifies them. Inspect only relevant material and retain source links, dates, metric definitions, and reporting periods. If a source is inaccessible, say so and request an accessible excerpt or export rather than implying it was reviewed.
Ask exactly 10 numbered, sharp questions in the initial interview. Tailor them to gaps and contradictions in available context rather than making the user repeat known facts. Known answers can be shown briefly for correction, with the question probing the remaining gap. Cover these areas:
- Product: what it does, for whom, and what outcome it delivers.
- Audience: narrow initial segment, buyer versus user, geography and language.
- Pain and buying trigger: urgency, current alternative, and why someone would switch.
- Offer and economics: price, purchase commitment, gross margin if known, and sales cycle.
- Goal: number of paying customers or revenue, deadline, baseline, and what counts as success.
- Traction: active users, paying customers, retention, and the source and period behind each count.
- Previous channels: actions, spend, exposure, qualified leads, paid conversions, and what failed.
- Budget: currency, monthly cash budget, total experiment cap, and acceptable acquisition economics.
- Capacity and advantages: team, available hours, skills, audience, relationships, and constraints.
- Evidence: upload or point to product and GTM sources, identifying the most authoritative and current ones.
For question 10, suggest relevant options: product URL/demo, positioning or GTM plan, pricing, customer interviews, CRM funnel export, billing/conversion data, channel analytics, or past experiment results. Request the smallest useful evidence set; aggregated or redacted data is usually enough. If the interface cannot accept uploads through a question tool, ask for files in the chat and use the tool only for text answers or source pointers.
Then stop and wait. Do not generate channel ideas, rankings, or test plans before the user answers. On resumption, preserve their answers and inspect the sources provided. A product description pasted into chat can serve as source material. No historical data is a valid answer for a new product; label the resulting analysis assumption-based. If essential product, audience, goal, cash cap, capacity, or source availability is still unresolved, ask only targeted follow-ups and wait. Unknown economics may remain explicitly unknown rather than being invented.
2. Assemble the customized prompt
Read the prompt template. Fill every field from answers and reviewed evidence; use explicit “unknown” or “not yet measured” entries where appropriate. Include a short evidence ledger and unresolved assumptions. Show the customized prompt before the evaluation in the same response; an additional confirmation is unnecessary unless a material contradiction would change the analysis. Do not treat assumptions as agreed decisions.
3. Evaluate channels
Read the channel and experiment guide. Research current customer presence, practical access, cost inputs and reach using accessible sources. Cite supporting pages or internal documents next to claims. If browsing or source access is unavailable, disclose that limitation and mark affected candidates as provisional.
Produce all 19 Outer Ring ideas, then select three distinct Middle Ring channel hypotheses. Fit them to the actual budget and team. Explicitly distinguish evidence from estimates, and customer presence from purchasing intent. Do not manufacture precise CAC or reachable audience figures to fill a table. Leave the Inner Ring empty, labelled “Awaiting test evidence.”
4. Design and review tests
Provide a separate 14-day plan for each Middle Ring candidate, with concrete daily actions and measurable outputs. Make the combined workload and spend feasible. If the team cannot run all three together, schedule sequential two-week tests and state the overall duration. Show hypotheses, metrics, attribution, caps, and advance stop/continue/success criteria using the guide.
End with the first executable preparation step and what results to bring back. Planning does not authorize publishing, contacting prospects, enrolling sequences, or spending budget; follow the user's actual authorization for any later execution.
When the user returns with results, compare observed paying customers and acquisition economics against the predeclared criteria. Classify each test as promising, failed, or inconclusive, and explain evidence quality. Promote a single channel only when evidence supports repeatable paid acquisition within the user's constraints. Zero conversions, immature cohorts, or sparse samples may leave the Inner Ring empty. A leading indicator is not a paying customer. Recommend focusing discretionary acquisition effort on a supported winner and pausing losing experiments, while retaining necessary supporting activities. Reassess when audience, economics or channel performance changes; do not promise one channel works forever.
Attribution
Adapted from the user's supplied summary and prompt attributed to Oskar Bader, “The Bullseye Framework 2.0: Find Your Best Marketing Channel To Get Customers With AI.” The original Bullseye framework comes from Traction by Gabriel Weinberg and Justin Mares. The supplied video/session itself has not been independently reviewed. The 14-day default and 10-question interview are requirements of this adaptation.
Background: Gabriel Weinberg interview on channel testing and interview on core and supporting channels.