What is the Cost of Delay Analysis?

The Cost of Delay Analysis helps you answer a practical prioritisation question: what does it cost us to wait? It estimates how much value you lose by delaying a bet — and turns timing into an urgency signal you can use to sequence work.

It’s especially useful when multiple bets look “high impact,” but only some are time-sensitive. Bet Impact tells you what to do; Cost of Delay helps you decide what to do first.

What Value It Gives

• Clear urgency – Quantify “we should do this now” as value lost per day, not just intuition.
• Better sequencing – Pick the bets that are expensive to postpone, even if several look similarly impactful.
• Fewer missed windows – Surface time-sensitive opportunities (seasonality, churn spikes, competitive moves) before they pass.
• Stakeholder alignment – Replace debate with transparent assumptions about impact, timing, and urgency.
• Stronger planning loops – Estimate urgency before shipping, then measure how delay affected outcomes to improve future planning.

Common Use Cases

• Roadmap sequencing – Decide what to ship first when capacity is limited.
• Time-sensitive opportunities – Capture seasonal demand, launches, or short-lived distribution windows.
• Churn and reliability – Prioritise fixes where every week of delay compounds leakage or risk.
• Revenue timing – Compare a pricing change vs. a conversion improvement when both matter, but one is more time-critical.
• WSJF-style prioritisation – Break ties between bets by urgency per build time (CD3).
• “Can this wait?” – Identify bets that are safe to delay (or where delaying reduces risk or downside).

Two Modes: Estimate Now, Measure Later

Planning Mode (Forecast-Based)

Planning mode compares two scenarios: ship now vs ship after N days, then estimates the value you lose by waiting.

• Uses a baseline forecast (“business as usual”) plus your expected impact (including lag/ramp).
• Applies optional discounting to reflect that later value is worth less than earlier value.
• Produces urgency signals like Cost of Delay per day and CD3 (urgency per build time).

Learning Mode (Observed Data)

Learning mode estimates how delay actually affected outcomes using observed data.

• Fits a regression model to estimate metric change per day of delay, with optional controls.
• Reports significance (p-value) and data-quality diagnostics to prevent false certainty.

How Cost of Delay Is Calculated

Under the hood (Planning Mode)

1. Build two scenarios

   • Immediate: the bet’s impact starts as soon as your lag/ramp allows.
   • Delayed: the bet starts after delayDays.

2. Choose how the delay is modeled

   • Shift within horizon: compare both scenarios over the same calendar window, with the delayed scenario pushing impact later inside that window.
   • Extend horizon: the delayed scenario starts later in the calendar; both scenarios still use the same number of days so the comparison stays fair (with discounting applied from the later start).

3. Compute value over the horizon
   For each day in the horizon, Segflow computes the expected delta (baseline vs impacted) and optionally discounts it.

4. Compute lost value
   lost value = total value (ship now) − total value (ship later)
   Positive lost value means waiting costs you value; negative means waiting avoids a negative impact.

5. Convert lost value into urgency metrics

   • Cost of delay per day = lost value ÷ delay days
   • CD3 = cost of delay per day ÷ duration days (build time)
     CD3 is useful for sequencing because it reflects urgency per unit of implementation time.

6. Score and rank bets
   Segflow converts lost value into a normalized score, weighted by:

   • Confidence (your evidence level)
   • Reach (metric importance: north star > goal > input, or an override) When forecast uncertainty bands exist, Segflow also provides low/median/high score ranges and a risk-adjusted score.

Under the hood (Learning Mode)

1. Estimate the effect of delay
   Segflow fits a regression model where the delay variable is the number of days delayed, optionally controlling for other factors.

2. Report “cost per day” with confidence
   You get an estimated effect per day, a p-value, and a confidence weight — plus data quality warnings when results are weak.

Example: Sequencing Two “High Impact” Bets (Illustrative)

Scenario: You have two strong candidates for the next sprint: a retention fix and a new top-of-funnel experiment. Both look valuable, but you need to decide what to ship first.

How you’d use Cost of Delay:

1. Set an expected delay window (e.g., “What if this slips by 14 days?”).
2. Connect each bet to the metrics it’s meant to move (e.g., churn, retained users, CAC).
3. Run Cost of Delay and compare lost value per day and CD3.

What you might learn:

• The retention fix has a high cost of delay per day because leakage compounds immediately.
• The acquisition experiment still matters, but its cost of delay is lower because benefits are slower to materialize (or more uncertain).

What you do next:

• Sequence the retention fix first, then run Bet Impact to choose the highest ROI follow-up bets once the leak is plugged.
• After shipping, use Learning Mode to validate whether delay truly had the cost you expected.

What You Provide

Required Inputs

• Bet duration – How long the bet takes to implement (in days).
• Bet → metric connections – Which metrics the bet is intended to move.
• Expected impact – Absolute or percentage change, with optional lag/ramp timing.
• Confidence – Your evidence level for the estimate.
• Metric importance / reach – North star vs goal vs input (or an explicit reach override).

Optional Inputs (Advanced)

• Delay window – The number of days you want to evaluate delaying the bet.
• Horizon – How far forward to measure value (default: 90 days).
• Delay mode – “shift within horizon” vs “extend horizon.”
• Discount rate – A daily discount rate to reflect time value of outcomes.
• Controls (Learning Mode) – Other metrics that may explain some movement.

What You Get Back

Core Outputs (Planning Mode)

• Lost value – The estimated value lost by delaying the bet by the chosen window.
• Cost of delay per day – The urgency rate (value lost per day of waiting).
• CD3 – Urgency per unit of build time (useful for sequencing).
• Score + ranking – A normalized urgency score for comparing bets.
• Contribution breakdown – Which metric connections are driving urgency.

Core Outputs (Learning Mode)

• Effect per day of delay – Estimated metric change per day of delay (in the metric’s units).
• Statistical confidence – p-value and an evidence-weighted confidence score.
• Diagnostics – Model fit and data-quality warnings.

Optional Outputs

• Risk-adjusted score – A conservative score when forecast uncertainty is wide.
• Low/median/high ranges – Scenario ranges when forecast percentile bands exist.
• Warnings – Notes when the horizon was truncated or when the delay window makes interpretation less reliable.

How to Interpret Results

• Positive lost value means “don’t wait” – Delaying burns value; the higher the cost per day, the more urgent the bet.
• Negative lost value means “waiting avoids harm” – This can happen when the bet is expected to move a metric in an undesirable direction. Always sanity-check direction.
• Use cost per day to sequence – If two bets have similar ROI, do the one with higher cost per day first.
• Use CD3 to break ties fairly – High CD3 means “high urgency per build time,” which is a strong signal for sequencing.
• Prefer risk-adjusted when uncertain – If ranges are wide, use conservative scores for planning.
• Heed warnings – If the horizon is truncated or delay exceeds the horizon, treat per-day results as directional.
• In Learning Mode, read effect + p-value together – A big effect with weak significance is directional; a smaller effect with strong significance is real but modest.

Best Practices

• Use Cost of Delay to decide when to do something; use Bet Impact to decide whether it’s worth doing.
• Choose a delay window that matches your planning reality (e.g., “one sprint,” “one month”).
• Use discounting when near-term value matters more than later value (e.g., runway constraints or seasonality).
• Keep assumptions explicit and revisit them after shipping — this is how planning gets better over time.
• Keep the number of primary outcome metrics per bet small; too many secondary metrics makes urgency harder to interpret.

Summary

The Cost of Delay Analysis quantifies urgency: how much value you lose each day you wait to ship a bet. Use it to sequence your roadmap around time-sensitive leverage — then validate with Learning Mode so your prioritisation improves every cycle.