The B2B Roadmap That Survives Q3

5 MINS

Every B2B PM has a moment, usually around Q3, where the roadmap they presented at the start of the year stops looking like the roadmap they're actually shipping. New deals come in. A big customer escalates. Sales wants a feature. A regulator publishes a memo. Engineering capacity halves because a critical hire took longer than planned. By October, the slide deck is fiction.

After a few cycles of this, I've stopped being upset about it. I've started designing roadmaps that *expect* this to happen. Here is what I now do differently.

Plan in three time horizons, not quarters

Quarters are a finance unit. They are not a product unit. The roadmap I actually defend now is in three horizons:

Now — what the team is shipping in the next six weeks. This is concrete, named, and almost not negotiable.

Next — what we plan to start in the following six to twelve weeks. This is directional, with explicit assumptions written down.

Later — strategic bets we believe in, but won't pre-commit to. Stakeholders see them, but they don't get a date. This sounds obvious. It is. The reason it works is that **the conversation about what to add stops being "should we shift the roadmap?" and becomes "should this enter Now, or wait for Next?"** That second conversation is much healthier.

The three questions I ask every time the roadmap moves

When something tries to break into the Now bucket — a sales request, a customer escalation, a leadership ask — I run three questions before I move anything:

Is this signal, or is it volume? A loud account is volume. A pattern across three accounts is signal.

Does this kill an entire class of problem, or one instance of it? Class fixes go in. Instance fixes get parked.

What are we *not* shipping if we do this? This is the question that gets skipped most often, and it is the one that matters most. Every roadmap insertion is a roadmap deletion. Make the deletion explicit. If a request can't survive these three questions, it goes to the Next or Later bucket — politely, but firmly.

The customer voice problem

In B2B SaaS, the temptation is to use "customer-driven roadmap" as a synonym for "whichever account shouted last quarter". That is not a roadmap. That is a queue.

The fix is less glamorous than people think. I keep two artifacts that I update monthly:

A patterns document. Themes I'm hearing across accounts, not the accounts themselves. The act of writing this forces me to find the underlying job.

A no-list. Things we have explicitly chosen not to do, with the reason. This list saves me an hour a week, because the same requests come back. The combination is what lets a B2B roadmap stay coherent under pressure. Without them, every quarterly review feels like starting over.

A short closing thought

A good B2B roadmap is not a promise. It is a hypothesis with dates attached, and a procedure for what to do when reality disagrees. The PMs who are best at this don't fight the disagreements. They build the procedure.

If your Q3 looks nothing like your Q1 plan, that's not necessarily a failure of planning. That can mean your planning is alive. The failure is when you ship the Q1 plan anyway, just to feel consistent. Don't do that. Customers don't want consistent. They want correct.

Lean Six Sigma Made Me a Better PM (Even Though I Don't Use the Charts)

Lakshmi skipped presentations and built real AI products.

Lakshmi Prasanna was part of the March 2026 cohort at Curious PM, alongside 17 other talented participants.