Audit Trunk Club's Product Photos Before the Visual-Search Rollout
Objective: Given a 6-image export of Trunk Club product photos with filenames, alt text, and format metadata, apply the lesson's image SEO checklist to flag which images are visual-search-ready and which need rework.
Trunk Club is prepping its catalog for a Pinterest Lens shopping pilot. Before greenlighting the pilot, you've been handed a 6-image export from the menswear catalog to audit against the lesson's checklist.
Score each image against the checklist (filename, alt text, format, single-subject framing) and flag the ones that need rework before the pilot launches.
Which of these product photos are actually ready for a visual-search pilot, and which need rework first?
Before you start
What you'll need
- —Familiarity with basic image SEO fundamentals, filenames and alt text
- Visual search
- search where the query is an image, not text, matched by the pixels and shapes in the photo itself rather than its metadata.
- Single-subject framing
- a photo composition showing one clear product with no competing objects or clutter in frame.
Free path (everything below is enough to finish)
Free, enough structure for a 6-row scorecard
Paid upgrades (optional, faster/deeper)
Manual review doesn't scale past a handful of images
The process
1 step
Step 01 of 01
The lesson's checklist calls for descriptive hyphenated filenames, alt text under 15 words describing content and purpose, WebP/AVIF formats, and clean single-subject framing, since visual search matches pixels, not metadata.
Given this 6-image export, which images are ready for the Lens pilot, and which need rework before launch?
Procedure
- List all 6 images with their current filename, alt text, and format
- Score each against the checklist: descriptive filename, real alt text, WebP/AVIF, single-subject framing
- Flag any image missing 2 or more checklist items as 'needs rework' before the pilot
- Approve only images that pass on framing, since that's the one visual search can't compensate for
File Alt text Format Framing Verdict IMG_5510.jpg (none) JPEG Cluttered desk bg FAIL, rework navy-wool-blazer-front.webp 'Navy wool blazer, front view' WebP Neutral bg, single PASS IMG_5512.jpg (none) JPEG Model + 2 products FAIL, rework brown-leather-belt-detail.webp 'Brown leather belt, buckle detail' WebP Neutral bg, single PASS grey-chinos-side-view.jpg 'Grey chinos' JPEG Neutral bg, single PARTIAL, convert format IMG_5515.jpg 'image123' JPEG Cluttered rack bg FAIL, rework
Healthy
3 of 6 images are flagged for rework before the pilot (missing filenames, alt text, or clean framing), and the pilot launches on the 3 that pass.
Unhealthy
Approving all 6 images for the pilot because 'they'll get indexed eventually', shipping cluttered, unnamed photos into a visual-search pilot wastes the launch on images the matching model can't reliably use.
What this means
Framing is the one item on this list a caption can't fix. A perfectly named, perfectly tagged photo of a cluttered rack still confuses the visual match model, filename and alt text only affect whether Google indexes the image at all.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Half the catalog export has generic IMG_ filenames and no alt text | Batch-rename and add alt text before the pilot, not after | half day |
| One image is well-tagged but shot with 2 products in frame | Reshoot single-subject before approving, tagging can't fix framing | half day |
Analyze your findings
What to look for
- Filename quality
- Is the filename descriptive and hyphenated, or a generic camera-generated name like IMG_5510.jpg?
- Alt text substance
- Does the alt text actually describe the content and purpose, or is it missing or a placeholder like 'image123'?
- Framing
- Is the shot single-subject on a clean background, or cluttered with multiple products or a busy backdrop?
- Format
- Is the image in a modern format like WebP/AVIF, or an unconverted JPEG?
Make the call
'grey-chinos-side-view.jpg' has a good filename and alt text, a clean single-subject frame, but is still a JPEG, not WebP/AVIF. Should it be approved for the pilot as-is?
Recommendation · Priority: Medium
“Approve the images that already pass framing and format, reshoot the cluttered or multi-product images since tagging can't fix framing, and batch-convert the otherwise-good JPEGs to WebP before greenlighting the Lens pilot. Framing is the one gate that determines the reshoot list.”
Common mistakes
What trips people up
Approving images because metadata looks fine — good filenames and alt text help indexing, but a cluttered or multi-subject frame still confuses the visual match model regardless of tagging quality.
Treating every failing image the same way — a missing alt text is a quick fix; a cluttered frame requires a reshoot, mixing these into one bucket wastes time on the wrong fix.
Skipping the format check because the photo looks fine visually — an unconverted JPEG can still pass framing and tagging but should be flagged and converted before launch.
Final deliverable
A scored 6-row audit sheet with a pass/rework verdict per image and the specific fix needed for each rework flag.
See a reference example
Applying the same scorecard to a Blue Bottle Coffee bag photo export (illustrative): 'ceramic-dripper-single-origin-pour.webp' passed on all 4 checklist items. 'IMG_9021.jpg' failed on filename, alt text, and format, all three fixed in one export pass before it went back on the sheet.
Success criteria
You're done when you can:
- Flags at least the 3 images missing filename or alt text as needing rework
- Does not approve the cluttered/multi-product framing images despite otherwise-good tagging
- Names framing as the one defect tagging alone can't fix
Key takeaway
Visual search matches pixels, not metadata, so framing is the one checklist item a caption or filename fix can never repair. Sorting each image by which specific checklist item it fails, not a single pass/fail, is what makes the rework list actionable.