gooseworks-ai/goose-skills / skills/marketing/where-to-start
where-to-start
VerifiedFirst-move advice for a broad growth or marketing ask with no job named: looks at what GooseWorks already knows and at the site, asks a few short questions in one message, names the one thing in the way, and sends the marketer to the right next job. Use when the marketer says "help me grow my brand", "I need to grow my business", "how do I get more customers / sales", "help me with marketing", "where do I start", "sales are flat", "what should I focus on this month", or "I'm back, what's next?". Recommends only; never runs the work. Not for a named job like "make me ads" or "grow my Instagram" (go straight to it); not for a campaign with a launch, budget or date — use plan-campaign; not for why this week's numbers moved — use analyze-performance.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 12.4 KB |
| references/base-rates.md | markdown | 5.1 KB |
| references/output-shape.md | markdown | 7.0 KB |
| references/playbook-map.md | markdown | 19.8 KB |
| references/questions.md | markdown | 5.1 KB |
| references/situation-lens.md | markdown | 9.7 KB |
| skill.meta.json | config | 3.2 KB |
| tests/test-by-hand.md | markdown | 9.9 KB |
references/situation-lens.md
View on GitHub# The advisor's lens
Load at steps 3 and 4. This is how a senior advisor reads a confused founder's situation. It is a set of questions and defaults held loosely, each with the signs that override it. It is not a decision table: no "stage X means playbook Y". Your own judgment picks the move; this page makes sure the judgment looks at the right things.
## 1. Find the binding constraint first
Ask: *if this one thing were fixed, would growth follow?* Name exactly one. The picks serve the constraint, not the channel the founder wishes for.
| Constraint | Signs |
|---|---|
| **Fit** (the product doesn't yet keep or delight the people who try it) | First sales then churn or no repeat; users try once and vanish; the founder can't say who loves it; B2B design partners not converting to paid |
| **Knowing the buyer** (who buys and why is still a guess) | Generic messaging; "everyone" as the audience; the founder describes features, not the job the customer hires it for |
| **Demand / awareness** (people who try it love it, but too few hear of it) | Good repeat or retention, good reviews, little traffic; word of mouth works but slowly |
| **Conversion** (people arrive but don't buy or sign up) | Traffic but few orders; ads get clicks, not sales; no clear first-order reason to buy; weak store page |
| **Activation** (people sign up or install but never reach the first moment of value, so they don't come back) | Many free sign-ups or installs, few active a week later, few converting to paid; "they try it once and disappear"; nobody knows where they drop off. Ask what the first useful result is and how many get there; if those who get there also leave, it is fit. People who paid and then cancelled are the strongest evidence: talking to them is often the first move. No event tracks the first useful result → add it (blindness) alongside |
| **Retention / lifecycle** (customers buy once and are never brought back) | List exists but no flows; repeat rate low for a repeatable product; no welcome or cart emails |
| **Creative supply** (paid works but the ads wear out faster than they are replaced) | Few ads running for weeks, frequency rising, click rate sliding, cost per sale climbing |
| **Blindness** (nobody can say what is working) | "I don't know my numbers"; platform revenue and store revenue disagree; tracking never checked |
| **Economics** (a customer cannot be paid for at this margin or price) | Thin margin, low order value, no repeat purchase, and the plan is to buy traffic; cost per customer already above what a first order earns |
| **Supply** (marketplaces: the scarce side can't meet demand) | Requests go unfilled, waitlists grow, existing providers are fully booked or turning work down |
| **Distribution access** (the buyers sit behind a gate: retail buyers, marketplaces, enterprise procurement, app-store search) | Sales depend on getting listed, stocked or found in a store's search |
When two seem tied, choose the one upstream: fit before demand, blindness before *more* spend, conversion before more traffic, economics before paid. Weigh the calendar too: a constraint that can be eased before a big date (e.g. retention flows before Black Friday) may beat a deeper one that takes months.
Place the constraint from evidence the founder gave (or the brain holds), and say which evidence in one clause. The constraint is about customers or the business. A missing setup step (claims not signed off, an ad account not connected, a list not exported) is never the constraint; it is something to bring. A plan saved on the brand by an earlier session is evidence of what someone once intended, not of what is in the way. The founder's own guess ("where do you think it's stuck?") is a lead, not a verdict: it is often a symptom.
If the constraint is not a marketing problem (the product doesn't work yet, pricing is broken, the business needs a salesperson), say so. That is a valid and valuable answer. Product-scope calls (who the product is for, which users to stop serving) belong to the founders: show them as a decision with what each choice costs.
## 2. Match the stage
Place the company in one of four stages by evidence (customers, repeat, a channel that works), not by how long ago it launched, and never guessed:
- **Pre-launch** — nothing sold yet; a launch date may exist.
- **Launched, searching for fit** — some customers or users, no repeatable pattern yet.
- **Early traction** — sales or signups come steadily from at least one repeatable channel; the founder wants more, or wants a second channel.
- **Stalled or leaking** — it was growing and now flattens or falls, or spend rises while results don't.
Defaults, each overridable by the evidence:
- Pre-launch and pre-fit companies **earn attention and learn the buyer's own words**; they don't buy attention. Paid spend before fit makes sense only as a small, deliberate learning test.
- With traction, **buy more of what already converts**, and fix the leak before adding traffic.
- **Re-reaching people who already signed up or bought** (email, a founder note, ads shown only to that list) is not buying attention; it is often the cheapest first move when a list exists. Paying to reach new people waits until the leak is understood.
- When stalled, **diagnose before prescribing**. A falling number has a cause; find it before spending more.
## 3. One or two, deep
Recommend one playbook. Recommend a second only when it is sequential (the first feeds the second) or a quick fix that runs alongside. Everything else the founder mentioned goes under "not yet", each with when it becomes a yes.
## 4. Use what they already have
An existing audience, email list, network, community, press angle or partner beats any cold channel. Ask about it and use it first.
Weigh the founder's real hours. A move that needs ten hours a week of posting is wrong for someone with two. A move that needs a budget is wrong for someone with none. **Sequence, don't stack:** one founding engineer can't run social at 5 hours a week, 50 cold emails and a creator wave at once. Add up the hours, pick the order, and give each later move the trigger that starts it.
## 5. Fix the basics before spending
Before recommending anything that spends money on traffic, ask (or read) whether:
- they can see sales by source (tracking);
- the site or store gives a clear first reason to buy (offer and page);
- in regulated categories (health, beauty, supplements, finance, food, kids), there are claims they are allowed to make.
A broken basic is the first move even when the fix is off the map and done by the founder or their developer. Name the playbook that helps alongside it.
## 6. Recommend only what the founder can feed
A playbook whose inputs the founder cannot supply this week is not a first move. Check each pick's "must bring" line in playbook-map.md. Common dependencies:
- plan-campaign needs a budget number and a date.
- make-static-ads, email-lifecycle and make-content need signed-off claims to write anything that states a fact: signed-off lines on file, or 3–5 lines the founder OKs inside that job.
- run-paid-campaigns needs a live ad account and an agreed cost per customer.
- analyze-performance needs something live that produces numbers.
- research-audience needs roughly who they sell to (it firms up the definition with them), and the founder's own tool to send from and verify addresses.
## 7. Respect the calendar
A launch, a seasonal peak, Black Friday / Cyber Monday (Black Friday is the Friday after the fourth Thursday of November; in 2026, 27 November), a holiday season, a funding deadline or a trade show can override the natural order. Name the date and work back from it. Compare against today's date; never assume the season.
## 8. Company type sets which channels are even in play
Use it as a lens on where buyers are, not as a route:
- **B2B SaaS / sales-led:** knowing the buyer, outbound lists, content and search, founder-led social.
- **DTC e-commerce:** paid social and creative volume, email and SMS flows, product imagery, creators.
- **Consumer app:** app-store search, community launch, short video, lifecycle; retention first.
- **Marketplace / two-sided:** the scarce side first (usually supply), then search.
- **Marketplace seller (Amazon, Etsy, Walmart):** the listing (images, title, price, reviews) before ads; most listing and marketplace-ads work is off the map today, so say so.
- **Local / multi-location:** listings, reviews and local search; offers; referrals.
- **Services / agency / consultant:** positioning, referrals, thought-leadership content, outbound.
- **Creator / media:** the audience itself, newsletter, sponsorships.
- **Non-profit:** donor lifecycle, storytelling content, partnerships.
Don't apply DTC logic to everyone.
## 9. More axes worth weighing
- **Sales motion:** self-serve, sales-led, retail or wholesale, walk-in, app store. It decides whether ads, outbound or listings are even the lever.
- **Where it is sold:** own site, marketplace, retail shelves, a local storefront. Off-site sales mean many on-site levers don't apply.
- **Repeat purchase:** consumables and subscriptions make email and retention worth far more than for a once-in-a-lifetime purchase.
- **Regulated claims:** health, beauty, supplements, finance, food, kids. Claims must be signed off (inside the job, through its claims check) before any playbook writes copy that states facts.
- **Who does the work:** the founder alone, a junior, an agency. A playbook that ends in "the human posts / sends / loads it" needs someone who will.
## 10. Ad bias is a known risk
GooseWorks sells ad, video and photo production. The right first move for many companies is not an ad. Saying "don't make ads yet, and here is why" is part of this job.
## Base rates
Rough ranges per company type, for calling a target unrealistic and setting the "worked if" number, are in `references/base-rates.md`.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |