Signal timing projects are often judged by anecdotal feedback: one driver says conditions improved, another insists they got worse. While public perception matters, transportation agencies need a more objective way to determine whether a timing change actually delivered the intended benefits.
A defensible before-and-after study uses observed traffic performance metrics to measure outcomes, separate operational changes from normal day-to-day variability, and determine whether improvements are consistent over time. By comparing equivalent periods and focusing on metrics such as control delay, arrivals on green, and excessive delay, agencies can confidently answer one of the most common questions that follows a retiming project: Did it work?
When to Use This Approach
This workflow is useful when:
- You’ve recently retimed signals and need to verify the results
- You’re preparing documentation for a consultant, grant application, council presentation, or public update.
- You want to avoid drawing conclusions from a single unusually good or bad day.
- Field resources are limited and new counts aren’t feasible
- External factors, such as return-to-office trends, new developments, or changing travel patterns, may be affecting traffic conditions.
What You Need Before You Start
You can run a solid before‑and‑after study with:
- Defined before and after date ranges
- Consistent time of day windows (e.g., weekday PM peak)
- Movement level metrics:
- Control delay
- Arrivals on green
- Excessive delay
- A defined list of intersections or corridors you want to evaluate
You do not need:
- New tube counts
- Floating car runs
- Detector reconfiguration
Step-by-step: A Defensible Before-and-After Workflow
Step 1: Lock in Comparable Time Windows
This is where most studies go wrong.
Select:
- The same days of the week (e.g., weekdays only)
- The same peak periods
- Multiple days in each period (not single days)
Avoid periods affected by:
- Construction weeks
- Incident heavy periods
- Weather outliers
If the time windows are not comparable, the results will be difficult to defend.
Step 2: Start at the Corridor Level
If timing changes were coordinated:
- Compare corridor travel time and reliability before vs after
- Look for improvements that are repeatable across days
- Confirm gains aren’t isolated to one short segment
Corridor results provide context before you dive into intersection details.
Step 3: Evaluate the Right Intersection Metrics
For each intersection or movement affected:
Focus on:
- Control delay – did average delay drop?
- Excessive delay – are vehicles waiting longer than three minutes at an individual intersection (i.e., multiple signal cycles)?
- Arrivals on green – did progression improve for key movements?
Look at patterns, not just averages:
- Are improvements or degradations consistent by hour?
- Do the same movements improve or degrade day after day?
Step 4: Verify Sample Size Before Drawing Conclusions
Low volume movements can distort results.
If samples are light:
- Extend the date range
- Stack multiple days
- Focus conclusions on higher volume movements first
A longer window with consistent conditions is more defensible than a short one with volatility.
Step 5: Look for “Shifted Problems”
After changes:
- Confirm improvements or degradations on the mainline didn’t create new delays on side streets
- Look for new excessive delays emerging elsewhere
- Verify gains or losses persist beyond the first few days
This avoids declaring success too early.
Step 6: Document What Changed (and why)
A defensible study includes brief context:
- What was adjusted (splits, offsets, coordination windows, demand, or environmental conditions)
- Which movements were targeted
- Why those locations were prioritized
This turns metrics into a story leadership can follow.
What Success Looks Like
Positive before‑and‑after results usually show:
- Reduced average delay during peak periods
- Fewer and less severe excessive delay
- Higher arrivals on green for coordinated movements
- Improvements that hold across multiple weekdays
Perfect numbers aren’t required; consistent improvement is.
Common Mistakes to Avoid
- Comparing mismatched time periods
- Using a single before day vs a single after day
- Ignoring construction or temporary disruptions
- Focusing only on averages instead of movement level patterns
- Declaring success without checking side effects
Reporting Results
Most agencies summarize studies with:
- A one-page before-and-after snapshot for each corridor
- A short table of key metrics by intersection
- A paragraph explaining what has changed and why
This format works for management, elected officials, and consultants alike.
A strong before-and-after signal timing study is less about finding perfect results and more about producing findings that can withstand scrutiny. By comparing consistent time periods, evaluating multiple days of data, focusing on meaningful performance metrics, and documenting both the changes made and the reasons behind them, agencies can clearly demonstrate whether a timing project achieved its objectives. The result is a credible, data-driven assessment that supports technical decision-making and provides leadership, elected officials, and the public with a clear understanding of what changed, why it changed, and how traffic operations responded.
Learn more about the U.S. Signals Scorecard at the July 29th webinar. Join here.
FAQs
Q: How many days should I include in a before‑and‑after study?
There’s no single number. Use enough days to create stable patterns—often a full week or multiple weeks—especially for lower volume movements.
Q: Should I compare averages or distributions?
Start with averages but validate with hour by hour or movement‑level patterns to ensure improvements are repeatable.
Q: Do I need field counts to validate results?
Not usually. Before‑and‑after studies rely on consistent observation of vehicle behavior. Counts are useful only if results are near warrant thresholds.
Q: What metrics matter most for retiming studies?
Control delay, arrivals on green, and excessive provide a strong operational picture when used together.
Q: How do I explain results to non‑engineers?
Focus on outcomes related to driver experience: reduced wait time, fewer stops, smoother progression—then show simple before/after visuals.
Q: What if results are mixed?
That’s common. Document where improvements occurred and where additional tuning may be needed. Mixed results are still useful.