marketing-ideas

Overview of marketing-ideas skill

What marketing-ideas actually does

The marketing-ideas skill helps an AI agent turn a vague growth question into a short, relevant set of marketing options for a SaaS or software product. Instead of freeform brainstorming, it works from a curated library of 139 marketing ideas grouped by channel and use case, then narrows them by product type, stage, budget, and team constraints.

Who should install marketing-ideas

This skill is best for founders, product marketers, growth generalists, and operators who need direction before committing to a channel. It is especially useful when the real problem is not execution yet, but prioritization: “what should we try next?” It is also a strong fit for marketing-ideas for Product Marketing workflows where product context matters more than generic campaign templates.

The job-to-be-done

Users usually install marketing-ideas when they are stuck, spread too thin, or unsure which growth motions fit their situation. The skill is designed to produce 3–5 plausible ideas first, then expand only the chosen paths with implementation guidance. That makes it more decision-oriented than a simple brainstorm prompt.

What makes this skill different from a generic prompt

The main differentiators are:

• a predefined catalog of 139 ideas, so outputs are less random
• a built-in habit of checking product marketing context first
• idea selection based on stage, audience, budget, and team size
• a practical output shape seen in evals: idea, why it fits, how to start, expected outcome, resources needed

If you want channel-specific execution right away, this is not the last skill you will use. It is the front-end triage skill that helps you choose where to go deeper.

Best-fit and misfit cases

Best fit:

• early-stage SaaS deciding where to focus
• Product Marketing teams needing option sets by audience or motion
• bootstrapped teams filtering for low-cost ideas
• teams that already have product context documented

Misfit:

• “write the campaign for me” requests
• highly technical execution inside one channel
• brand strategy, positioning, or messaging work without enough product context
• businesses outside software/SaaS unless you adapt the ideas carefully

How to Use marketing-ideas skill

Install context for marketing-ideas

The upstream SKILL.md does not publish an install command, so install from the repo path used by your skill runner. In Claude Code-style setups, teams commonly add the parent repo and invoke the marketing-ideas skill from there. If your environment supports direct skill installs, use the repository source:

https://github.com/coreyhaines31/marketingskills

Then verify the skill folder exists at skills/marketing-ideas.

Read these files first

For a fast adoption decision, read:

1. skills/marketing-ideas/SKILL.md
2. skills/marketing-ideas/references/ideas-by-category.md
3. skills/marketing-ideas/evals/evals.json

This order matters. SKILL.md tells you when to trigger the skill, the reference file shows the idea inventory, and the evals show what “good” output looks like in practice.

The most important setup detail

Before asking questions, the skill explicitly checks for .agents/product-marketing-context.md or .claude/product-marketing-context.md. This is the biggest quality lever in the whole workflow. If that file exists and is current, the AI can skip basic discovery and recommend ideas that fit the actual product instead of generic SaaS advice.

What input marketing-ideas needs

At minimum, give the skill:

• product and category
• target audience
• company stage
• pricing or deal size
• budget range
• team size
• main goal: leads, trials, demos, expansion, awareness, retention
• known constraints: no designer, no ad budget, long sales cycle, compliance, weak domain authority

Without these, marketing-ideas usage becomes broad and repetitive.

Turn a rough request into a good prompt

Weak prompt:

• “Give me marketing ideas.”

Better prompt:

• “Use the marketing-ideas skill for our B2B SaaS. We sell compliance software to mid-market healthcare teams, ACV is $18k, sales-led, team of 5, budget is $4k/month, and we need more qualified demos in the next 90 days. Prioritize ideas we can test without a full content team. For each idea, explain why it fits, how to start, expected outcome, and resources needed.”

That structure matches the eval expectations and usually produces more actionable prioritization.

A practical workflow that matches the repository

A strong marketing-ideas guide looks like this:

1. Load product marketing context if available.
2. Ask only for missing information.
3. Select 3–5 ideas from the catalog that fit the stated constraints.
4. Rank them by likely impact and feasibility.
5. Expand the top 1–2 ideas into first steps.
6. Hand off to a more specialized execution skill if needed.

This keeps the skill focused on idea selection instead of drifting into shallow execution across too many channels.

How to ask for better filtering

The reference file covers many categories: Content & SEO, Paid Advertising, Partnerships, PR, Product-Led Growth, Developer, Audience-Specific, and more. To get useful output, tell the model how to filter:

• “low-budget only”
• “works for enterprise sales cycles”
• “non-founder-dependent”
• “good for a developer tool”
• “best for a launch in 30 days”
• “exclude paid channels”
• “favor ideas with compounding effects”

This is where the catalog becomes more valuable than generic brainstorming.

What good output should look like

Based on the evals, the best marketing-ideas usage output is not a long list of 25 tactics. It is a short prioritized list with consistent fields:

• idea name
• why it fits
• how to start
• expected outcome
• resources needed

If the first response lacks prioritization or ignores budget/team constraints, ask for a rerun with explicit scoring by impact, speed, and effort.

How to use marketing-ideas for Product Marketing

For Product Marketing, frame the request around segment, buying motion, and proof. Example:

• “Use marketing-ideas for Product Marketing for an AI note-taking SaaS selling to RevOps leaders. We need ideas that create market education and product proof, not just traffic. Recommend 4 ideas that support category understanding, objection handling, and pipeline creation.”

This pushes the skill toward comparison content, proprietary data, webinars, partnerships, proof assets, and other motions that help buyers understand the product.

When to hand off to another skill

Use marketing-ideas to choose the play. Then switch tools when you need execution:

• paid campaign buildout
• email sequence writing
• landing page copy
• social content
• PR outreach
• SEO production planning

The skill itself says it is a starting point whenever someone is stuck. Treat it as strategy triage, not the full growth stack.

marketing-ideas skill FAQ

Is marketing-ideas useful if I already know my channel options?

Yes, if your problem is prioritization rather than discovery. The value of the marketing-ideas skill is not just “more ideas”; it is matching ideas to stage, budget, and product context so you stop considering tactics that are wrong for your situation.

Is this better than a normal AI brainstorm prompt?

Usually yes, because the skill has a fixed idea library and an expected workflow. Generic prompts often produce obvious suggestions with weak filtering. marketing-ideas is more consistent when you want SaaS-relevant options narrowed by constraints.

Is marketing-ideas beginner-friendly?

Yes, especially for solo founders and early product marketers. But it works best when you can answer a few basics about audience, pricing, team, and goal. If you cannot describe your product clearly yet, outputs will be broad.

Can I use marketing-ideas outside SaaS?

Possibly, but it is clearly optimized for SaaS and software products. Some ideas transfer well; others depend on software economics, product-led loops, or content/search dynamics that may not fit services, ecommerce, or local businesses.

Does the skill include execution templates?

Not really. The repository is stronger on idea selection than implementation assets. It gives categories, fit logic, and examples of expected structure, but not deep playbooks for every channel.

When should I not install marketing-ideas?

Skip it if you need:

• one-channel tactical depth right now
• messaging or positioning work before growth ideas
• detailed attribution planning
• a non-software-specific marketing system

In those cases, a channel or strategy-specific skill may be a better first stop.

How to Improve marketing-ideas skill

Give marketing-ideas richer business constraints

The fastest way to improve results is to provide sharper boundaries. Good constraints include:

• budget ceiling
• time horizon
• founder availability
• current traction
• sales motion
• available content or data assets
• channels you have already tried

The skill improves materially when it can eliminate ideas, not just include more of them.

Maintain a strong product marketing context file

Because the workflow checks .agents/product-marketing-context.md or .claude/product-marketing-context.md first, keep that file current. Include:

• ICP and segments
• product category
• pricing and motion
• top objections
• main competitors
• proof points
• current traction
• team resources

This reduces repetitive questioning and improves recommendation quality across sessions.

Ask for ranking, not just brainstorming

A common failure mode is getting a flat list of ideas with no tradeoffs. Fix that by asking the model to score each option on:

• likely impact
• speed to first signal
• resource intensity
• fit for current stage

That turns marketing-ideas from an inspiration tool into a decision tool.

Prevent generic outputs with exclusion rules

Another common issue is broad advice like “do SEO” or “run ads.” Ask for exclusions such as:

• no paid acquisition
• no long-term brand plays
• no channels requiring daily founder content
• no enterprise events
• no tactics needing engineering support

This forces the skill to use the catalog more intelligently.

Iterate from chosen ideas into test plans

After the first pass, do not ask for “more ideas” unless the list clearly missed the mark. Instead say:

• “Take idea #2 and turn it into a 30-day test.”
• “For the top-ranked idea, define first 5 actions, owner, budget, and success metric.”
• “Compare the top 3 ideas by expected time-to-value.”

This is the best way to improve marketing-ideas usage without losing focus.

Use the reference and evals to calibrate quality

If outputs feel weak, compare them against:

• references/ideas-by-category.md for breadth
• evals/evals.json for expected structure and filtering behavior

This helps you spot whether the issue is bad prompting, missing context, or the model failing to follow the skill.

Know the main tradeoff of marketing-ideas

The central tradeoff is breadth versus depth. marketing-ideas gives you a wide, curated option set and decent prioritization, but not deep implementation systems. The best improvement path is to use it early, choose a direction quickly, and then move into a narrower execution workflow.