Grow A Product

grow-a-product

ship-a-product ends at the announcement. That is deliberate: it is a launch orchestrator. This is what happens next, and it is a loop rather than a pipeline — you will run it many times on the same product.

0. product-analytics       live product -> you can see activation, funnel, retention, cohorts
1. user-onboarding         signup -> the moment the product first delivers its value
2. (hold retention)        the retention curve flattens instead of going to zero
3. saas-billing-stripe     a human can pay, on a plan that makes sense
   + pricing-page          the price and packaging are decided and presented
   + regional-pricing-stripe  only if the market spans wildly different incomes
4. launch-seo, geo-aeo     found by search engines and answer engines
   + content-engine        what to publish and how often, on the channel you own
   + community-organic     other people's spaces, where being a guest is the whole rule
   + content-rabbit        the queue that keeps the cadence
5. founder-led-sales       the first customers, by hand, for anything not cheap self-serve
   unit-economics          before any paid spend: fully-loaded CAC, gross margin, payback
6. ad-conversion-hub       one canonical conversion event, consented and deduplicated
   + google-ads / meta-ads / the platform adapter you actually buy on
   + ad-experiments        a test with a decision rule written before it runs
   + ad-auto-optimizer     only once there is enough volume to optimise anything
7. lifecycle-email         welcome, activation, trial-ending, abandoned checkout, dunning, win-back
   + release-notes         shipping visibly is retention work, not marketing
8. incidents               you find out it broke before your customers tell you
   + support-inbox          and when they tell you first, which is most of the time
   + dependency-hygiene     the slow failure: a product you can no longer build

The order is the whole point

Most growth work fails by starting at stage 6. Paid acquisition into a product with broken activation is the fastest way to spend money proving you have a product problem. Each stage below is cheap to skip and expensive to have skipped.

• Nothing before stage 0. You cannot fix a funnel you cannot see, and you cannot reconstruct last month's cohort from a tool you installed today. launch-analytics (stage 6 of ship-a-product) measures traffic; product-analytics measures whether anyone came back. You need the second one here.
• Activation before acquisition. If people sign up and never reach the value, more signups produce more churn and a worse reputation. Find the activation moment in product-analytics, then move that number.
• Retention before spend. A retention curve that goes to zero means paid acquisition is renting users, not buying them. Look at whether the curve flattens, not at how high it starts.
• Charging before scaling. A free product with no billing path has no signal about value and no money to fund acquisition.
• Economics before paid. unit-economics turns "is this channel working" into a number. Without a fully-loaded CAC and a payback period, ad-experiments and ad-auto-optimizer optimise a figure that is not tied to whether the business works. Payback usually binds harder than the LTV:CAC ratio at this size, because it is about cash rather than profit.

Deciding where to enter

• Just launched, nothing is happening? Stage 0. Almost always the answer is that nobody knows what "happening" would look like, because nothing is instrumented past pageviews.
• Signups but no usage? Stage 1. This is an activation problem, and buying more signups makes it worse.
• Usage but nobody comes back? Stage 2. Stop here. Do not spend on acquisition until the curve flattens.
• People stay but nobody pays? Stage 3. Usually packaging, not price.
• Product works, nobody knows it exists? Stages 4 to 6, in that order. Organic compounds and costs time; sales teaches you the objections; ads cost money and teach you the least.
• Paid is running and losing money? Go back to stage 0 and check the conversion event is real before touching the campaign. A hub event that never fires looks exactly like a bad campaign.

What this skill does NOT replace

Stage 2 is product work, not marketing work. user-onboarding covers stage 1 -- shortening the path to the activation moment -- but there is no skill for "make the product good enough that people come back", because that is the actual job. This orchestrator makes sure the surrounding stages are not skipped or reinvented; it cannot substitute for the product being worth returning to.

The measurement that decides everything

One number ends most arguments: does the retention curve flatten? A curve that flattens at any height is a business, because the flat part is the population that keeps the product. A curve that reaches zero is a leaky bucket, and every stage after 2 pours water into it.

Build that curve in product-analytics before agreeing to any spend.

Cost and delegation shape

Same pattern as ship-a-product: judgment stays with the highest-tier model — what the activation moment is, what to charge, whether a retention curve is good enough — and mechanical execution goes to a cheaper one. Do not spend on a later stage until the earlier one is measured, not merely believed. Confirm activation before funding acquisition; confirm the conversion event fires before scaling a campaign.

Related

• ship-a-product — everything before this: name, build, repo, SEO, launch.
• validate-an-idea — earlier still. If stage 2 keeps failing, the answer may be that the idea was never validated.
• agentification — a distribution channel this loop does not cover: being discoverable and usable by agents rather than people.