Attribution
Attribution Windows: Choose a Useful Conversion Lookback
Choose a conversion lookback using observed buying delays, compare campaigns at similar maturity, and document window changes before reviewing performance.
Tracly Team · October 2, 2026 · 4 min read
Direct answer
Choose an attribution window by examining the delay between eligible marketing interactions and verified outcomes, then selecting a period that fits the decision and buying cycle. Keep window definitions visible, compare equally mature cohorts, and distinguish the time allowed for credit from the model that assigns it.
A campaign launched yesterday has had less time to produce sales than one launched last month. Window settings add another difference: the same delayed purchase may qualify in one report and fall outside another. Those differences need explaining before a team reallocates spend.
| Window | Approach one | Approach two |
|---|---|---|
| Window | Defines which interactions are eligible by time | Model determines how eligible interactions receive credit |
| Short period | Highlights faster recorded outcomes | May exclude meaningful delayed purchases |
| Long period | Includes more delayed eligible outcomes | Does not prove the interaction caused the purchase |
Separate time eligibility from credit assignment
A lookback or conversion window defines the time relationship allowed between an interaction and an outcome. An attribution model defines how credit is assigned among eligible interactions. Changing from last-click to another model and changing the window are different actions, even if both alter channel totals.
Google Ads describes its conversion window as the period after an ad interaction during which a conversion is recorded. Different event types and systems can have different settings. Record the event, interaction type, window, and reporting date basis for each view instead of assuming one account-level label makes all reports comparable.
Study buying delays without hiding excluded outcomes
Review the time between recorded interactions and verified purchases or qualified leads. Segment materially different journeys, such as an immediate offer and a product requiring a sales conversation. A single average can hide both fast purchases and a long tail of delayed decisions.
Remember that a report constrained by its current window cannot reveal every outcome beyond that window. Where appropriate, use your own matched business records to study longer delays. State the coverage limits of the available data. A missing late conversion is not evidence that no late conversion happened.
- Define the event being measured
- Review the distribution of delays
- Identify where the current report limits visibility
Compare cohorts after similar time has elapsed
Group traffic by acquisition period and give each group similar time to mature before comparing performance. In an illustrative example, ten sales from last week's clicks and six from this week's clicks cannot establish a decline if the newer group still has several days of typical buying delay ahead of it.
Separate an early operational view from a matured performance review. Early monitoring can detect a broken link or absent event stream without declaring the campaign unprofitable. Label preliminary numbers and schedule a later review rather than waiting silently or treating unfinished results as final.
Document changes and preserve a reporting baseline
Before changing a window, record the old definition, the new one, and the effective date. Check the platform's current rules for how changes affect future collection and historical reporting. Do not assume that a longer setting will automatically restore every previously excluded event.
Keep a consistent comparison baseline when reviewing Tracly campaign performance alongside ad-platform or order-system reports. Explain differences in timing and eligibility before interpreting the totals. A longer window can assign more credit, but only experiments and appropriate business evidence can help assess whether a campaign created additional demand.
Frequently asked questions
Is a thirty-day window always the best choice?
No. The useful period depends on the buying cycle, event, system constraints, and decision. Review observed delays and explain what your chosen period includes.
Why do recent campaigns show fewer conversions?
They may have had less time to mature. Check tracking health immediately, then compare performance after similar elapsed time and with consistent definitions.
Does extending the window prove an ad caused more sales?
No. It changes time eligibility for attribution. Causal impact requires additional evidence, such as a suitable controlled experiment.
Make campaign decisions with clearer data
Tracly brings attribution, traffic quality, testing, and performance signals into one practical workflow.
Explore TraclyRelated articles
Attribution
Marketing Attribution: What It Is and How to Choose a Model
Learn what marketing attribution is, how common attribution models work, and how to choose a model that supports better campaign decisions.
Attribution
Last-Click vs. Multi-Touch Attribution: When Each Model Fits
Compare last-click and multi-touch attribution, understand their tradeoffs, and choose a reporting view that fits your marketing decisions.
Conversion Tracking
Why Conversion Numbers Do Not Match Across Marketing Tools
Reconcile conversion discrepancies across ad platforms, analytics, and checkout records by checking event definitions, timing, attribution, and duplicates.