The Slack message arrives at 4:47 PM on a Friday: "The new dashboard is live! Great work team." A few claps emoji. Someone posts a GIF. The PM closes their laptop feeling accomplished.
Two weeks later, adoption is at 3%. The PM is confused. They shipped a solid feature -- good design, clean implementation, no major bugs. But almost nobody is using it.
They shipped it. They did not launch it. And the difference explains why so many well-built features fail to find their audience.
The Distinction
Shipping is the act of deploying code to production. It is an engineering milestone. The feature works. The tests pass. It is accessible to users. Done.
Launching is the act of ensuring users know about the feature, understand its value, know how to use it, and have a reason to try it now. It is a go-to-market event. It involves messaging, timing, enablement, and measurement.
Shipping is necessary for a launch. It is not sufficient.
Most product teams are excellent at shipping. Sprints have clear deliverables. CI/CD pipelines enforce quality gates. Engineering managers track velocity. The entire development process is optimized for getting code to production.
Almost none of that infrastructure exists for launches. There is no CI/CD pipeline for "did users notice this feature?" There is no velocity metric for "did the launch messaging resonate?" The launch process, when it exists at all, is ad hoc and under-resourced.
What a Launch Actually Requires
A launch is a coordinated effort across multiple dimensions. Missing any of them turns a launch into a ship.
Audience Identification
Who specifically needs to know about this feature? Not "all users." Which segment? At Intercom, they distinguish between features for new users versus existing users, power users versus casual users, and buyers versus end users. Each audience needs a different message through a different channel.
Value Messaging
What problem does this feature solve, stated in the user's language? Not "we built a new reporting module." Try: "You can now see which campaigns drive revenue without exporting to a spreadsheet." The feature's existence is not news. The problem it solves is news.
Discovery Mechanism
How will users find this feature? In-app announcements? Email campaigns? Onboarding flow changes? Tooltip tours? Each mechanism has different reach and different intrusiveness. A critical launch might warrant an interstitial modal. A minor improvement might only need a small badge in the sidebar.
Enablement
Can support answer questions about the feature? Can sales demo it? Has documentation been updated? If a user tries the feature and gets confused, what happens next? Launch without enablement creates a surge of confused users and a swamped support team.
Timing
When you launch matters almost as much as what you launch. Launching on a Friday afternoon means most users will not see it until Monday, and by then the newness has faded. Launching during a customer's busy season means they are too distracted to adopt. Launching the week after a competitor's big announcement means you are fighting for attention.
Measurement
How will you know if the launch worked? Not just "did usage go up?" but specifically: What adoption rate are you targeting by day 7? By day 30? What does success look like, and what is your plan if you do not hit it?
The Launch Spectrum
Not every feature deserves the same launch effort. Here is a rough framework:
Tier 1: Major launch (2-4 per year). New product capabilities that change the value proposition. Full go-to-market: blog post, email campaign, in-app announcement, sales enablement, press if relevant. Example: Notion launching their database feature.
Tier 2: Standard launch (monthly). Meaningful improvements to existing capabilities. In-app announcement, targeted email, changelog entry, support documentation. Example: Adding a new chart type to an existing analytics feature.
Tier 3: Quiet release (weekly). Bug fixes, performance improvements, small UX polish. Changelog entry only. No active push. Example: Improving load time on a settings page.
Tier 4: Feature flag rollout (ongoing). Gradual rollouts to percentage-based user cohorts. No announcement until full rollout. Used for risky changes where you need to monitor impact. Example: A new search algorithm that might affect result relevance.
Classifying each release into the right tier is itself a product decision. Upgrading a Tier 3 to Tier 2 treatment can significantly boost adoption. Downgrading a Tier 1 to Tier 3 wastes a big product moment. Spend 10 minutes at the start of each sprint asking: "What tier is each release this sprint, and are we treating it accordingly?"
The mistake most teams make is treating everything as a Tier 3. They ship features that deserve Tier 1 treatment with a Slack message and a changelog entry. Then they wonder why adoption is low.
The Launch Checklist
Here is the checklist I use for Tier 1 and Tier 2 launches. Adapt it to your context.
Two weeks before launch:
One week before launch:
Launch day:
One week after launch:
Why Teams Conflate Shipping and Launching
Three structural reasons:
PM incentives are misaligned. Many organizations evaluate PMs on delivery -- "did the feature ship on time?" This creates a strong incentive to optimize for shipping velocity and a weak incentive to invest in launch quality. The PM who ships 12 features with mediocre adoption looks more productive on paper than the PM who ships 6 features with strong adoption.
Launch work is cross-functional, and nobody owns it. Engineering owns the ship. Marketing owns the blog post. Support owns the documentation. Sales owns the enablement. The PM is supposed to coordinate all of this, but it often falls through the cracks because each team has their own priorities. Without a clear owner and a clear process, the launch coordination simply does not happen.
The dopamine is in the deploy. Shipping feels good. It is concrete, visible, and immediate. The deploy notification is a reward signal. Launch work -- writing emails, briefing sales, creating help docs -- feels like administrative overhead. It is the broccoli after the dessert. Teams naturally gravitate toward the rewarding work.
The Adoption Gap
Here is the uncomfortable truth: most features have terrible adoption rates. Industry data from Pendo and Amplitude suggests that the median feature adoption rate for B2B SaaS products is between 20% and 30%. That means 70-80% of users never touch most features.
Some of that is expected -- not every feature is for every user. But a significant portion of that gap is caused by poor launch execution. The feature exists. It works. Users just do not know about it, do not understand it, or do not have a compelling reason to change their current workflow.
Closing the adoption gap is not about building better features. It is about launching them properly.
Companies That Get This Right
Notion treats every major launch as a campaign. When they shipped databases, it was not a changelog entry. It was a landing page, a video demo, template gallery, creator partnerships, and coordinated social content. The feature was the same quality either way. The launch is what made it a moment.
Linear releases features with focused, well-written blog posts that explain not just what they built but why they built it and how it fits into their vision. (For a full breakdown of how Linear's approach benefits product teams, see Linear for product teams.) Each post reads like a product essay. Users read them not because they need to learn the feature, but because the narrative is compelling. The launch content itself becomes a distribution mechanism.
Stripe embeds launches into their developer documentation from day one. When a new API endpoint ships, the docs, code examples, and integration guides are already live. For developers, the documentation is the launch. Stripe understood that their users do not read blog posts about features -- they read docs when they need to implement something.
Each company adapted their launch approach to their audience. There is no single right way to launch, but there is always a wrong way: silently.
The Shift PMs Need to Make
The mindset shift is this: the PM's job is not done when the feature ships. It is done when the feature achieves its intended outcome.
If you defined success as "25% of target users adopt this feature within 30 days," then shipping is the starting line, not the finish line. The launch plan, the adoption monitoring, the iteration based on early usage data -- that is the work that gets you across the line.
Start allocating your time accordingly. For a major feature, I budget roughly 60% of my PM time for discovery and development, and 40% for launch preparation and post-launch measurement. Most PMs spend 95% on the first and 5% on the second. That ratio is why features land with a whisper.
A Mental Model
Think of it this way: shipping is building a bridge. Launching is telling people the bridge exists, explaining why crossing it is worth the effort, and making sure the on-ramp is smooth.
If you build a bridge and nobody crosses it, the problem is not the bridge. The problem is that nobody knew it was there.
The next time your team finishes building a feature, ask yourself: have we shipped it, or have we launched it? If the answer is only the first, you are halfway done. The second half -- the launch -- is where the user impact happens. Do not skip it.