17 The 4-Week Gap: What a Failed Recency Filter Can Hide
The 4-week gap is the interval between refreshing an ecommerce listing and the point when enough impressions and conversion data exist to judge the refreshed asset. This matters for ecommerce sellers because a recency filter can exclude that learning period and make an untested revision look like a stable winner.
In this article, the term is a practical audit label, not an industry benchmark. It describes the evidence gap created when a dashboard, catalog rule, or creative workflow favors the newest assets while dropping older comparisons, unrefreshed winners, and newly published pages that need time to mature.
What a recency filter gets wrong
A recency filter answers a narrow question: which assets were published or changed most recently? It does not answer whether those changes caused better outcomes. Applied to product photography, it can turn upload dates into an accidental ranking system. A newly uploaded hero image may appear beside a proven image because both are current, even when the new file has lower clarity, weaker mobile loading, or less relevant framing.
The boundary creates a second problem. A file created on the last day of a 28-day window receives the same eligibility as a file with four weeks of exposure. Meanwhile, the previous hero image, which supplied the baseline, may disappear from the report. The comparison becomes new versus nobody, not new versus established. Google Search Central also treats filenames, nearby text, and image relevance as separate inputs; a recency rule cannot replace those signals.
What the gap can hide
1. A false winner. A revised image can lead for one week because a promotion sent more qualified traffic to it. If that week is the only cohort retained, the dashboard rewards timing. It cannot show that the asset still works after the promotion ends or that it merely inherited the promotion's demand.
2. Evergreen evidence. A top-selling product may keep the same clear studio image for years. That image is commercially valuable, but a strict changed in the last four weeks rule excludes it. The seller then loses the control that would reveal whether a new lifestyle image improved results.
3. Seasonal or channel shifts. Four weeks may sit inside one campaign, traffic source, device mix, or inventory state. Stockouts can suppress conversion; a new creative can coincide with more cautious buyers. Google Trends supports custom date ranges, so a four-week result can be checked against a longer baseline instead of being treated as a permanent trend.
Technical failures that look like creative failures
Recency filtering can also admit a broken asset while hiding the working control. If the new file is oversized, cached incorrectly, or missing descriptive metadata, it may be visually correct in the CMS and underperform in delivery. Search Central explains that image sitemaps help discovery but do not guarantee display, so an indexable file is not proof of a sound customer experience.
These thresholds do not diagnose recency by themselves. They provide a quality floor for the pages carrying the images. A recency report should record loading and rendering failures beside conversion results.
A four-step workflow that preserves the evidence
Use recency as a cohort label, not a verdict. This workflow keeps new assets visible without pretending they have the same history as established ones.
- Step 1: Define comparable windows. Record the launch or change date, then compare the latest four weeks with a named baseline such as the prior four weeks and, where volume allows, the same seasonal period. Keep traffic source, device, geography, price, and inventory notes attached.
- Step 2: Prepare controlled variants. Change one meaningful variable at a time. You can edit and resize product photos in a browser-based studio without altering the composition, making crop, export, and file-weight tests easier to interpret.
- Step 3: Build context without losing product clarity. Use a mockup generator for contextual product scenes and an image background remover for clean product cutouts. Label every output because a lifestyle scene and a white-background hero are different creative treatments.
- Step 4: Report evidence by age. Publish a current four-week cohort beside named historical cohorts. Include zero-impression assets, delivery errors, and stockouts. Call a result promising only when the metric improves and the sample remains credible.
Rewarx workflow versus a recency-only competitor
This is a workflow comparison. The competitor column represents a recency-only process, not a claim about every competing product. A broader workflow preserves the control and the uncertainty; a recency-only workflow may display fewer assets while showing less context.
| Decision point | Rewarx | Competitor |
|---|---|---|
| Evidence window | Current four weeks plus a named baseline | Newest four weeks only |
| Missing files | Retained with launch date and zero-result status | Silently dropped |
| Creative changes | Isolated, named, and linked to test dates | Often bundled with upload recency |
| Decision output | Confidence, caveats, and next test | Automatic recent winner |
| Long-term assets | Evergreen controls remain eligible | Disappear unless refreshed |
Audit checklist and safeguards
- ✓ Every new image has a dated cohort and named baseline.
- ✓ Traffic source, device, geography, and price changes are recorded.
- ✓ Stockouts, delivery errors, and zero impressions remain visible.
- ✓ Each result includes sample size and confidence limits.
- ✓ Evergreen controls are compared before replacement.
A recency filter should identify the latest evidence, not decide which evidence is allowed to survive.
Frequently asked questions
What is the 4-week gap in ecommerce?
The 4-week gap is the evidence gap between changing an asset and having enough observed traffic and conversion history to judge it. It is not a universal testing period; traffic volume, purchase cycle, and seasonality determine how long a seller should wait.
How can I tell whether a recency filter is hiding useful results?
Compare the current four-week cohort with a named prior window and inspect files excluded by the rule. Look for proven evergreen images, assets with zero impressions, publication errors, and promotions or stockouts that distort the comparison. Google Trends supports custom date ranges for market-demand checks, while your analytics platform supplies asset-level performance.
Should every product image older than four weeks be replaced?
No. Age alone is not evidence of weak creative. Replace an image when it is inaccurate, unclear, slow to load, poorly descriptive, or no longer reflects the product. Google Search Central recommends relevant images, descriptive filenames, and useful alt text, but performance data should still determine the test.
Audit the assets your filter excludes
Prepare cleaner creative variants, compare them by dated cohorts, and keep the evidence your recency rule would otherwise erase.
Try Rewarx Free