Continuous Forecasting
Make realistic commitments from actual delivery history.
Nave uses your team’s actual delivery history to calculate the probability of finishing a given amount of work by a date, or when a given amount of work is likely to be done. No story points. No estimates.
Two questions
Every commitment starts with one of two questions.
Nave answers both using your actual delivery history.
Probability
A forecast isn’t a promise. It’s a probability.
Higher confidence requires more time or a smaller commitment.
The percentages help you weigh different commitments based on your team’s actual daily throughput. An 85% forecast means the work finished by that date in 85% of the cases.
See how likely you are to hit each date.
For the same work, an earlier date comes with lower confidence and a later date with higher confidence. You choose how much risk to accept.
A SINGLE DATE
07 Dec
DATES BY CONFIDENCE
- 30%01 Nov
- 50%12 Nov
- 70%24 Nov
- 85%07 Dec
- 95%24 Dec
- 98%05 Jan
The method
No story points. No estimates.
Nave forecasts from what your team has actually completed, not what someone estimated the remaining work might take.
How Nave calculates the forecast.
Start with your actual throughput
Nave counts the work items completed each day in the selected history window, including days with no completions.
Simulate many possible futures
Nave runs 10,000 trials by default, drawing random days from that daily history to simulate possible delivery outcomes.
Calculate the probabilities
Nave calculates how often the work finished by each date, or how many items finished within the available time.
Keep it current
Compare different commitments on the spot as the charts recalculate. The Continuous Forecasting dashboard updates daily, showing whether the target remains likely as the team delivers work.
“Developers didn’t believe the forecasts initially,” Tim shares. “They thought they could complete tasks faster than what the data suggested. But after a few iterations, they realized the forecasts were far more reliable than their own predictions.”
Timofey Yevgrashyn Agile Coach and Trainer, DoodleOne forecast
One forecast. One decision.
Question
Starting 28 Sep, when are 20 items likely to be done?
Nave forecast
Nave shows 50% by 12 Nov, 85% by 7 Dec, and 95% by 24 Dec.
Decision
Use 7 Dec as the 85% confidence date for the commitment.
“We ended up with a 75% probability of completing the features within the client’s desired timeframe, and we felt confident with this. Had the probability percentage been lower, we would have investigated reducing the features for this release. In the end, we actually finished everything two weeks earlier than forecast.”
Francine Souza Agile Project Manager, ArcTouchContinuous forecasting
The forecast doesn’t stop when planning ends.
The Continuous Forecasting dashboard shows whether your commitment is still likely while there is time to act. It keeps the original target date fixed and recalculates daily as items finish, so you can cut scope, improve flow, add capacity, or tell stakeholders early if the likelihood drops.
You’re not comparing reality against an estimate made three months ago. You’re making decisions from the delivery evidence you have now.
Changing scope
What happens when the plan changes?
As your team finishes items, the likely delivery dates and chance of meeting your target update based on the work left and the team’s recent pace. Change the scope or date to compare options.
20 items: 85% by 7 Dec
15 items: 85% by 24 Nov
Negotiate scope, dates, and commitments with evidence from the forecast.
- What if we reduce scope?
- What if this date is fixed?
- How much can we commit to at this confidence level?
Process stability
A forecast is only as useful as the system behind it.
Stable does not mean fast. Stable means predictable.
If the underlying delivery system is behaving predictably, its historical performance gives you a stronger basis for forecasting.
If something meaningful has changed, investigate and stabilize the process before blindly trusting the forecast.
Items do not need to be the same size. Your delivery history already includes the mix of sizes your team handles. What matters is a predictable delivery system.
Check Process ImprovementUse cases
Use the forecast where commitments happen.
Release planning
“We need this by Q4. How much scope can we realistically commit to?”
Backlog planning
“We have 70 items remaining. When are we likely to finish?”
Near-term commitments
“How much can this team reasonably complete in the next two weeks?”
Stakeholder conversations
“Can we give leadership an 85% confidence date instead of a single unsupported deadline?”
Customer story
Frost Design and Consulting Group
“I did some searching around before I found Nave. I signed up for a free trial and right away, I could see that there’s a way to make reliable delivery commitments using past throughput data – which we had plenty of in Trello.”
John Paul de Guzman Founder, Frost Design and Consulting GroupFaq
Questions about probabilistic forecasting
Probabilistic forecasting shows how likely different delivery outcomes are, rather than giving you one date without context. Nave calculates those probabilities from 10,000 possible futures based on your team’s actual daily throughput, including days when nothing finished.
No. Nave uses the number of work items your team actually completed each day. Your delivery history provides the evidence for the forecast, so there is no size, estimate, or story-point input.
No. Nave does not estimate individual work items. For a single item, use the Cycle Time Scatterplot to see how long a given percentage of past items took to finish.
No. Your delivery history already includes the mix of item sizes your team handles. Reliable forecasts depend on a predictable delivery system, not evenly sized items.
The forecast uses daily throughput from the dashboard’s selected date range, which defaults to the last three months. Days with no completions count as zero, so slow days remain part of the team’s delivery history.
Nave counts an item as completed when it reaches one of the board’s Done statuses. If the board has multiple Done statuses, reaching any one of them counts as completion.
Nave counts calendar days, including weekends and holidays. The daily history includes those days as they happened, with days that had no completions counted as zero, so they are already reflected in the forecast.
Yes. Filters for statuses, labels, members, and custom fields determine which items count. This lets the forecast cover the specific work being committed.
Nave sets no minimum. In a stable delivery system, 20 to 30 completed items can be enough. New teams should collect data for a few iterations before making long-term commitments. After a significant team or process change, use only the history since that change.
Choose a confidence level based on the work and how much uncertainty you can accept. A 50% forecast is about a coin flip. Around 70% can suit easy work with stable flow, while 95% or 98% better suits complex or unfamiliar work. Higher confidence requires more time or a smaller commitment.
Include all items that need to be completed, both new work and work already in progress. The Continuous Forecasting dashboard subtracts items finished since its start date, as counted by the dashboard view’s filters.
The Monte Carlo charts recalculate immediately when you change the item count, dates, or filters. The Continuous Forecasting dashboard recalculates once a day from the latest synced delivery data, so you can see whether the commitment remains likely.
Change the item count to reflect the new scope. The Monte Carlo charts recalculate immediately, so you can compare the different commitment on the spot.
Yes. Use Number of Tasks to see how many items are likely to be completed between a start and end date. The Continuous Forecasting dashboard fixes its target date on the first run, then tracks the daily likelihood of meeting it.
Yes. Use Delivery Date to see when a given number of items is likely to be done. The chart shows how the likely date changes at different confidence levels.