Product Launches
Most of what kills a launch is knowable weeks ahead. Here's how to find it before the money goes out.
By Shannon Kearns, published 2026-10-01, 10 min read.
You prevent a failed product launch by surfacing its risks before the budget is committed, while they're still cheap to fix. Almost every launch failure I've looked at was knowable in advance.
And while our reflex is to prevent failure with a bigger launch plan (more assets, more channels, a longer checklist) that just makes the launch louder, and more work for everyone. It does nothing to test the decisions underneath it.
Prevention means checking the most common failure-point decisions, on purpose, before launch day.
Run a pre-mortem before you build the launch plan. Ask each person who touches the launch to imagine it's 90 days after launch day and it flopped, then explain why. The answers surface the risks people see.
The method comes from psychologist Gary Klein, who described it in Harvard Business Review in 2007. It builds on 1989 research by Deborah Mitchell, Jay Russo and Nancy Pennington, which found that imagining an event has already happened increases people's ability to correctly identify reasons for the outcome by 30%.
It works for a simple reason. "What might go wrong?" invites people to be polite. "It failed, why?" gives them permission to say what's been bugging them for weeks.
In a launch, the pre-mortem does a second job. When you ask everyone separately, you find out how differently they see the launch. Sales, product and marketing will each name a different cause of death. That's the first step of avoiding a flopped launch: defining how it could possibly go wrong.
Hold 30-minute one-on-one conversations with four to six stakeholders, turn what they say into one page, and walk the team through it in a single alignment meeting. Start to finish, it only takes a few days.
The steps:
Don't skip that last step and just email the page around. A document that lays out every gap in black and white can set off a lot of worry when people read it alone. The meeting is where it turns into decisions.
I walk through the full process, including the one-page template I use, in Do This Before Your Next Product Launch. The full set of stakeholder questions is in How to Align Stakeholders Before a Product Launch.
Before anyone opens a launch plan template, write down four decisions: who the launch is for, the one thing you'll claim, what success looks like at 30, 60 and 90 days, and what you will not do. The pre-mortem usually shows which of these are still open.
Every one of these gets made eventually. If you don't make it on purpose, it gets made by default, one planning meeting at a time. I laid out a way to capture all four on a single page in the one-page GTM doc.
Split the product release from the market launch, roll the product out in phases (I call it a Crescendo Launch, some call it a Rolling Thunder Product Launch), and decide in advance which number at each phase tells you to keep going or hold. Then a problem found late becomes a paused campaign instead of a failed launch.
Most teams fuse the two into one date because every function wants certainty. Product wants a ship date, sales wants a date they can start talking about it, and leadership wants something to announce. The trouble is that one shared date turns every product issue into a market issue.
How to set it up:
I watched this save a launch. Three days before general availability, paid conversion dropped significantly in the power-user phase. Because the criteria were already agreed, the call to hold was fast and nobody had to argue for it. We paused the acceleration and kept the narrative running.
One more question for this stage: does this change look or work differently for any customer segment, tier or account size? Missing that answer is how a launch breaks a workflow for your biggest customers on day one.
Test the core of the launch at small scale first: the buyer, the message, and at least one channel. You want a small, repeatable result before launch day, because a launch turns up a motion that already works. It's a bad place to find out whether it works at all.
What to test:
Then widen in stages. Start with the segment where the evidence is strongest, build early wins there, then expand. (I made the full case for this in The Crescendo Product Launch.) Early-stage teams should also read why early-stage product launches fail before this step, because the biggest risk there is usually demand that was never validated.
This is part of preventing a bad launch plan as well. Plan the first 30 days after launch with the same care as the lead-up: one place for signal to land, an owner for each leading indicator, and enough runway to act on what you learn.
This is also where the pre-mortem pays off a second time. The risks you rated high are the ones you watch first.
And if you're reading this after the fact, the same method works in reverse. Stop the spend that feeds the broken step, gather evidence, run a blameless post-mortem, and trace the gap back to the decision that caused it.
You prevent a product launch from failing by checking its riskiest decisions before the budget is committed. Run a pre-mortem with each stakeholder, lock who the launch is for, what you'll claim, what success looks like and what you won't do, then roll out in phases with go/no-go criteria and test small before launch day.
A product launch pre-mortem is a pre-launch exercise where you ask each stakeholder to imagine the launch has already failed and explain why. It's based on Gary Klein's project pre-mortem and on research showing that imagining an outcome has already happened improves people's ability to identify its causes by 30%.
Run it before the launch plan gets built, as soon as the product scope and rough timing are known. That's when the answers can still change who you target, what you claim, and how you roll out.
Go/no-go criteria are the numbers, agreed before rollout, that decide whether each launch phase moves forward or holds. Pick two or three per phase, like activation rate, paid conversion, or support ticket volume, set the threshold that means "hold," and agree in advance who makes the call.
The clearest early signs are stakeholders defining success differently, the team naming more than one target customer, nobody being able to say what the launch won't do, KPIs added after the plan was built, and launch day being the first real test of your channel and message.
The pre-mortem itself takes a few days: four to six 30-minute conversations, about an hour of synthesis, and one 30 to 60 minute alignment meeting. Testing demand, message and channel at small scale takes longer, usually a few weeks, and should happen before the launch budget is committed.
One person should own the launch end to end, including the risks, even though every function contributes to them. Usually that's the product marketer or launch lead. Their job is to surface the risks, get each one an owner and a decision, and hold the go/no-go calls during rollout.
Read more articles, explore services and pricing, or take the free Go-to-Market Readiness Assessment.