Pre-estimationAsync planning poker: estimate before the meeting
Most backlog items don’t need a debate — they need five people to quietly agree it’s a 3. Collect those estimates in advance and spend the meeting on the tasks where the team really disagrees.
How it works
The same planning poker — private cards, simultaneous reveal — stretched over a day instead of an hour.
In a live session everyone votes on one task at a time. In pre-estimation everyone gets the whole list and estimates it at their own pace — between meetings, in another time zone, after lunch. The facilitator only steps in at the end, to review.
- Prepare the session. Add the tasks for a new session (type them, import a CSV or Excel file, or paste a list) and turn on the Pre-estimation switch.
- Start the collection. Click Start pre-estimation and share the room link. Each person opens a private dialog with every task: pick a card and, if your room uses them, rate the estimation arguments. Choices are saved as you click and can be changed until the end.
- Watch the progress. The facilitator sees how many tasks each person has estimated — never the values.
- Close and review. Close pre-estimation & review reveals everything at once: cards, statistics and a verdict for every task.
- Accept or go live. Accept the estimates you trust, mark the rest for live voting, then click Go live — accepted tasks stay locked and the meeting starts with the first open one.
Nothing about the method changes: votes are still private and still revealed together, so nobody anchors on the first number. What changes is when the thinking happens — and that people get time to read the task instead of hearing it once.
When it helps, when it hurts
Pre-estimation is a filter. It works when most tasks are clear and fails when the clarity has to be created in the room.
The point of estimation is the conversation that disagreement triggers. Async voting keeps that — the disagreements surface in the review — but it removes the conversation from tasks that don’t need one. That is a great trade for some backlogs and a bad one for others.
Use it for
- Routine, well-understood work: bugs, small improvements, familiar features
- Long backlogs — 15 tasks or more
- Teams spread over time zones
- Teams where quieter people rarely speak first
- Re-estimating items that were estimated long ago
Avoid it for
- A new product area nobody has worked in yet
- Tasks that need a demo or a walkthrough to make sense
- Very short lists — five tasks are faster live
- Thin task descriptions (“fix the thing”)
- Teams that won’t find time to vote before the meeting
The last point is the most common failure. If half the team votes five minutes before the meeting, you get the downsides of both formats. Async estimation needs a deadline and a facilitator who checks progress — see timing and invitations.
A worked example
Twenty tasks, six people, one refinement meeting.
An illustration with typical numbers. A team of six has 20 backlog items to estimate before sprint planning. Live, at about five minutes per task, that’s a 100-minute meeting — in practice two meetings, because nobody stays focused that long.
Instead, the facilitator imports the tasks on Monday, starts pre-estimation and gives the team until Wednesday morning. On Wednesday the review looks like this:
| Review result | Tasks | What the facilitator does | Time |
|---|---|---|---|
| Consensus | 14 | Accepts them during the review | ≈ 5 min |
| Discuss | 3 | Cards match, an argument is contentious — moves them to live | ≈ 35 min |
| Spread | 3 | Cards differ — moves them to live | |
| Total | 20 | Six tasks discussed live, fourteen accepted | ≈ 40 min |
The meeting shrinks from 100 minutes to about 40, but that’s not the main win. The 40 minutes are spent entirely on the six tasks where people actually see the work differently — the conversations that justify estimating at all. Nobody sits through fourteen rounds of “5, 5, 5, 5 — next”.
Your ratio will differ. A mature team working in a familiar codebase may accept 80% of tasks; a team in a new domain may accept a third. If fewer than a third of tasks reach consensus, the backlog wasn’t ready — refine descriptions first.
Create a room, add tasks, switch on pre-estimation. Free, no sign-up.
Timing and invitations
Async doesn’t mean open-ended. Give a clear window and a reason to finish.
How long to collect
One to two working days is the sweet spot. Shorter, and people vote in a hurry; longer, and the collection drifts until the morning of the meeting. Start it right after the backlog is groomed, close it an hour or so before refinement — or at the start of the meeting itself, with everyone watching the review.
How to invite
Post the room link in the team chat with three things: the deadline, the number of tasks and the expected effort. For example:
smartpokerplanner.com/room/ABCD1234
During the collection
- Check the progress the day before the deadline and nudge people who haven’t started — the facilitator sees counts, so it’s a reminder, not an inspection.
- Use the coffee card for “not my area”. An honest coffee breaks consensus on that task and sends it to the meeting — better than a guess that quietly gets accepted.
- If the facilitator also votes, they estimate from the same dialog (Estimate tasks yourself).
Reading the review
Every task gets statistics, a verdict and a preselected final value. Your job is to decide what to trust.
After you close the collection, each task shows the average, the median, the spread and the number of votes, plus a badge. The final value is preselected as the median (snapped to your scale); pick another card if the team’s history says otherwise.
| What you see | What it means | Usual decision |
|---|---|---|
| Consensus | Same card from everyone, no coffee, arguments aligned | Accept |
| Discuss | Same card, but people rate one of the risks very differently | Look at the heat map; move to live unless the reason is obvious |
| Different cards | The team imagines different work | Move to live — this is what the meeting is for |
| Few votes | Only two of six people estimated the task | Don’t trust a “consensus” of two — move to live |
| Big numbers (13+) | Even when everyone agrees, the task is large | Consider splitting before you accept |
Accept, move to live, go live
Accept locks the estimate with the selected value and moves the task to the top of the list. Move to live marks a task for the meeting. Go live ends the review: accepted tasks stay locked, everything else goes into the usual live flow, and voting starts with the first open task. You don’t have to decide on every task — anything you leave unaccepted simply goes live.
One link for the whole team. No accounts, no installation.
Frequently asked
What is async planning poker?
The team estimates tasks on their own time instead of in a live meeting. Everyone picks a card for every task privately; the facilitator then reviews the results and brings only the unclear tasks to the meeting.
Can I change my estimate after I’ve submitted it?
Yes. Your choices are saved as you click, and you can change them until the facilitator closes the collection.
Can teammates see my estimates during pre-estimation?
No. During the collection the facilitator only sees how many tasks each person has estimated. The cards are shown in the review after the collection is closed.
Can I turn pre-estimation on for a session that has already started?
No. Pre-estimation is a switch you turn on when you add the tasks for a new session. For a running session, use live voting.