Your Gantt chart was accurate for about four days. Then a dependency slipped, a hotfix ate the sprint, and the roadmap you presented to the VP now describes a timeline that no longer exists. Nobody updated the slide. Everybody quietly stopped trusting it.

This is the core problem with date-based engineering roadmaps: they promise precision you don't have. A commitment to ship auth in week 6 reads as a contract to the people above you and a fiction to the people writing the code. The fix isn't a better Gantt tool. It's dropping the pretense that you know exact dates twelve weeks out.

The Now/Next/Later format does that. Three columns instead of a timeline. It communicates sequence and confidence without fabricating dates you'll have to walk back. This post is a build guide for it as a live board — not a slide — with confidence labels and stakeholder permissions baked in, so leadership can read it anytime without a status meeting.

Why Now/Next/Later beats the Gantt chart

A timeline roadmap answers "when." That's the wrong question for anything past your current cycle, because the honest answer is "we don't know yet." Now/Next/Later answers "in what order, and how sure are we" — which is the question stakeholders are actually asking when they poke at your roadmap.

  • Now — in active development this cycle. Scoped, assigned, moving.
  • Next — the on-deck queue. Agreed direction, not yet started, order can still shift.
  • Later — directional bets. Real problems worth solving, deliberately un-scoped.

The columns absorb reality. When a Now item slips, it stays in Now — you don't repaint a timeline. When priorities change, a card moves one column left or right. The roadmap updates by dragging a card, not by rebuilding a chart.

The one thing this format drops on the floor is time-based confidence. "Later" tells a stakeholder when-ish but says nothing about how likely. That gap is where roadmaps still lose trust. So we add a second axis.

Add confidence labels as the second axis

Columns show sequence. Labels show conviction. Put a confidence label on every card so a reader can tell the difference between "we're doing this, scoped and certain" and "we think this matters but haven't validated it."

A three-level scheme is enough — more just invites debate about the difference between 60% and 70%:

  • Committed (green) — scoped, resourced, we'd be surprised if it changed.
  • Likely (amber) — direction agreed, details soft, order may shift.
  • Exploring (grey) — a bet. Might get cut after discovery.

In Zoobbe you can do this two ways, and they're not mutually exclusive. Board-level color-coded labels are reusable across every card and read at a glance on the board — good for the visual scan. A single-select custom field called "Confidence" is better if you want the value to sit in a structured place you can filter and sort on later. For most teams the label is enough; reach for the custom field when you start slicing the roadmap by confidence.

The combination is what makes this honest. An Exploring card in Now is a spike — you're actively de-risking a bet. A Committed card in Later is a known future dependency you've already locked. Two axes describe your real roadmap; one column never could.

Rule of thumb: if a stakeholder asks "is this definitely happening," the answer should already be on the card. That's the label doing its job.

The build: a roadmap board in Zoobbe

Here's the concrete setup. Adjust list names to your cadence.

  1. Create three lists: Now, Next, Later. That's the whole structure. Resist the urge to add "Done" — a roadmap isn't a task tracker, and shipped work should graduate off the board, not pile up.
  2. One card per initiative, not per ticket. A roadmap card is "Self-serve SSO," not "Add SAML callback handler." Keep cards at the altitude leadership reads at. The engineering breakdown lives on a separate delivery board or in the card's checklist.
  3. Add the confidence labels: Committed / Likely / Exploring, with distinct colors. Apply one to every card. A card with no confidence label is a card nobody owns an opinion on.
  4. Set priority (Normal / High / Urgent) if you want a within-column ranking that's separate from confidence. Confidence is "how sure," priority is "how much it matters" — don't conflate them.
  5. Use custom fields for the detail leadership asks about: a text field for the target quarter (not a date), a single-select for owning squad, a number field for a rough t-shirt-size estimate. These render on the card and answer the follow-up questions before they're asked.
  6. Write a real description on each card: the problem, who it's for, what "done" looks like directionally. This is what turns a roadmap from a list of feature names into something a stakeholder can actually reason about.

Stakeholder permissions: read access without the status meeting

The entire point of a live roadmap is that the VP, the sales lead, and the support manager can look at it whenever they want instead of pinging you for an update. That only works if you can give access without giving edit rights — nobody wants a stakeholder dragging cards.

Two levers control this in Zoobbe:

  • Board visibility. Set the roadmap board to Workspace-only so everyone inside your org can view it, or Public if you need to share it beyond the workspace. Keep the granular delivery boards Private. Stakeholders see the roadmap; they don't see the sausage being made.
  • A companion roadmap page. If you'd rather present the narrative — the "why now" behind Now/Next/Later — build a Page and share it with specific roles: viewer for most stakeholders, commenter for the folks whose input you actually want, editor for co-owners. Permissions inherit from the workspace or the parent page, so you set them once.

The result: leadership self-serves. When someone asks "where's the roadmap," you send a link once, not a fresh export every month.

Keeping it alive

A roadmap dies when updating it is a chore. Two habits keep it breathing.

First, review it on a fixed cadence — end of each cycle, ten minutes. Move what shipped off the board. Pull one or two cards from Next into Now. Re-check every confidence label: did anything you marked Likely earn Committed, or should it drop to Exploring? The label review is the meeting.

Second, let the board notify instead of you. Watchers on a card get pinged when it moves, so a stakeholder who cares about SSO subscribes once and hears about it when the card crosses into Now — no broadcast email required. If you run a lightweight automation on the roadmap board, a card moving into Now can drop a comment or notification automatically, so the transition is logged without anyone typing it.

And because the board updates in real time, two staff engineers reworking the Next column during planning see each other's changes as they happen. The roadmap reflects the conversation you're having, not the one you had last Tuesday.

Key takeaways

  • Drop date-based timelines past your current cycle — they promise precision you don't have.
  • Now/Next/Later columns show sequence; confidence labels (Committed/Likely/Exploring) show conviction. You need both axes.
  • One card per initiative, at leadership's altitude — not one per ticket.
  • Set the board to Workspace-only or Public so stakeholders self-serve; keep delivery boards Private.
  • Review on a fixed cadence and re-check every confidence label. The label review is the update.

FAQ

How is Now/Next/Later different from a sprint board?

A sprint board tracks execution of committed work over one or two weeks. A roadmap tracks direction and sequence over quarters. Keep them separate: roadmap cards are initiatives, sprint cards are tickets. Linking them by putting the ticket breakdown in a roadmap card's checklist is fine; merging the two boards isn't.

Should confidence labels or priority drive the order?

They answer different questions. Confidence is how sure you are something happens; priority is how much it matters. A high-priority, low-confidence item is exactly the thing to run a discovery spike on. Show both and let readers combine them.

How do I stop stakeholders from editing the roadmap?

Use board visibility and page roles. Set the board to Workspace-only for view access, and share any companion roadmap page as viewer or commenter rather than editor. They can read and comment without moving your cards.

How often should the roadmap change?

The columns should shift a little every cycle and the confidence labels should tighten over time. If nothing moves for a month, the roadmap has quietly become a slide again.

Building your first Now/Next/Later board? Spin one up in Zoobbe, add the three confidence labels, and set it Workspace-only so leadership can read it without a meeting. Start here.

Photo by Kelly Sikkema on Unsplash