Image Stack Audit — ResultSport Spiky Massage Ball Set of 3

Amazon UK Full stack audit HTML · v1 ASIN B00CCPNDWI · Brand ResultSport · Audit date 22 May 2026
Data sources: live PDP scrape (amazon.co.uk/dp/B00CCPNDWI, 2026-05-22) · SQPR Q1 2026 · Sellersprite reverse 30d · SP search-term snapshot 2026-04-30 · prior slot-2 audit (2026-05-21, superseded).
Rules canon: Amazon Search Algorithm & Listing Optimization.pdf (Mar 2026) + Amazon_Banned_Keywords_Guide.pdf (Apr 2026).
✓ Permanent snapshot. The 6 slot thumbnails in this report are saved as slot1.jpgslot6.jpg alongside this HTML at 500×500 px. The before/after record is durable: the visual reference will not change when the seller pushes the rebuilt gallery, even if Amazon delists the ASIN or rotates the CDN. As long as the product folder stays intact, the audit reads correctly.

1. Executive summary

The current image stack is doing two things badly and one thing well. The bad: two slots (4 and 5) carry banned-keyword violations that put the listing at active suppression risk under Amazon's automated review system — "TOP QUALITY" + "TOP GRADE" in slot 4, "MAXIMUM HEALTH BENEFITS" + "Prevent Health Complications" in slot 5. The wording in slot 5 is the higher-severity one because "Prevent Health Complications" is a borderline medical claim. Pulling and replacing these two slots is the single highest-priority action — even before any design work, the seller should consider taking them offline.

The other failure is structural: slots 2, 3 and 4 all duplicate the "3 sizes" job, while no slot answers the genuine top buying objection ("which size for which body zone / which complaint?"). The keystone slot 2 is generic — small-caps "3 Different Sizes" header in a default font, no risk-kill badge, no anchor on the sub-niche search drivers we found in the SQPR (plantar fasciitis, trigger point, myofascial release). It scores 28 / 80 on the 8-dim keystone scorecard — Broken tier. Slot 7 doesn't exist as a comparison/closer; slot 6 carries packaging which would belong in slot 7.

The one thing this listing does well is the A+ Content's built-in Rufus-style FAQ block ("Can these help with plantar fasciitis?" / "How do they help with muscle knots?" / "Can I use these for reflexology?") — exactly the kind of intent-aligned seeding the user's reference PDFs recommend. The cross-sell comparison module is also above category average. The gap is that the A+ has no anatomical size-guide module, no materials-safety panel, and no brand-story banner.

Stack tier
Broken
3 slots duplicate jobs · 2 carry compliance violations · 0 slots answer top objection
Slot 1 (hero)
62 / 100
Compliant. Pedestrian. 3 push tactics available — all zero/low risk.
Slot 2 (keystone)
28 / 80
Broken tier — rebuild from scratch.
Compliance
2 HIGH-risk slots
Slot 4 + Slot 5 require immediate copy fixes.

2. Listing context

Product: ResultSport Spiky Massage Ball Set of 3 — solid plastic spike balls in 6 cm, 8 cm, 10 cm diameters for self-myofascial release. 227 g full set. PAH-free, phthalate-free, glue-free construction. Listing at amazon.co.uk/dp/B00CCPNDWI.

Buyer avatar (named per COSMO knowledge-node guidance)

  • Adults managing plantar fasciitis pain and arch tension
  • Runners with tight calves and post-training foot fatigue
  • Office workers easing desk-bound back, neck and shoulder posture tension
  • Pilates and yoga practitioners using spike balls for foot mobilisation, posture release and grip work
  • Adults working through a physiotherapy plan between in-clinic sessions

Top 3 buying objections (data-backed from SQPR + listing-rewrite findings)

  1. "Which size do I actually need for what I want to fix?" — biggest objection. The product wins on every "spiky massage balls" plural variant (23.68% purchase share) but loses heavily on bare "massage ball" (0.33%) — the long-tail intent buyers know they want spikes, but the broad-intent ones can't see why a 3-pack is worth it over a single ball.
  2. "Is it firm enough / safe enough for bare skin and barefoot use?" — implicit in the "Is it sharp?" / "Is the plastic safe?" review themes the listing-rewrite v3 surfaced.
  3. "How is this different from a tennis ball or lacrosse ball that I already own?" — long-tail "lacrosse balls massage" (6,037 vol) and the fact that the product's value-prop isn't legible in current slot 2.

3. Keyword evidence (cluster ↔ image coverage)

Mapping the SQPR clusters from the kw-research run against which slot (if any) carries a visual answer:

ClusterVol (top kw)Current image coverageVerdict
Core — "spiky massage ball"15,995Slot 1 (3-ball cluster), Slot 2 (size labels)Covered
Sub-niche — "trigger point massage ball"5,733None — neither image nor on-image text mentions trigger pointMissing
Sub-niche — "myofascial release ball"3,064 (trend)NoneMissing
Sub-niche — "plantar fasciitis"7,387 (trend)None — slot 3 hints at foot use but text says "Large"Missing
Sub-niche — "physio ball"6,403NoneMissing
Sub-niche — "foot massage ball"6,086Slot 3 (small foot illustration)Weak
Comparison — vs lacrosse / tennis6,037None — no comparison-chart slot existsMissing
Material safetyn/a (objection)Slot 4 (claims "TOP QUALITY", banned)Compliance fail

Reading: 5 of the 8 cluster-level intents are not visually answered. The image stack is doing far less work for the keyword graph than it could — and what work it is doing (slot 4 / 5) is non-compliant.

4. Stack-level diagnosis

  • Job duplication in slots 2, 3, 4. All three try to communicate "3 different sizes" with slightly different framings. By the time the shopper reaches slot 4 they've seen the same info three times — and slot 4 adds nothing except a banned superlative.
  • No slot owns the size-to-body-zone mapping. The single clearest visual that would close the biggest objection — "6 cm = arch of foot · 8 cm = calf / hamstring / glute · 10 cm = back / shoulder" — doesn't exist as a dedicated slot. Slot 3 attempts it but the text is illegible at thumbnail size and the labels say "Large / Medium / Small" instead of the specific body zones.
  • Compliance debt concentrated in slots 4–5. Both slots ship copy explicitly listed in the user's banned-keywords PDF as Section 1 (severity HIGH) or borderline medical claim. Suppression risk is real, not theoretical.
  • Hero (slot 1) doesn't use any of the legal CTR push tactics. Straight-on shot, no diagonal isometric, no micro-shadow grounding, no luminosity drop. Pure white background and product-only — fully compliant, but leaving CTR on the table.
  • Slot 7 is empty. The closer / comparison slot is missing. Slot 6 carries packaging which would belong in slot 7.
  • No slot speaks to the sub-niche queries that drive the genuine search-volume opportunity (trigger point, myofascial, plantar fasciitis, physio). These are the same gaps the listing-rewrite v3 identified for the title and bullets — but the image stack is the highest-leverage place to surface them.

5. SLOT 1 — Hero (deep-dive)

Current slot 1 hero
Current slot 1 · pure white BG · 3-ball cluster

Job: CTR + category legibility

Three solid plastic spiky balls (teal, dark purple, green) clustered together on a pure white background. Compliant on every hard rule — no text, no models, no props, no badges. Product fills roughly 65–70% of the frame. The cluster arrangement does communicate "set of 3" but doesn't make the size differentiation obvious — the eye reads "three balls in different colours" before it reads "three different sizes".

4-dimension scorecard

Category legibility
8/10
Crop strength
6/10
Trust cue
5/10
CTR push potential (used)
3/10

Total: 22 / 40 (Compliant, pedestrian). Compliance verdict: pass — all hard rules respected.

CTR push tactics available

TacticRiskAction for this hero
Diagonal isometric rotationZero riskRotate the 3-ball composition along the canvas diagonal so the 10 cm ball sits front-left and the 6 cm trails back-right. Maximises pixel fill in the 1:1 mobile thumbnail and gives a natural depth cue that the current frontal cluster lacks.
Micro-shadow groundingZero riskAdd a tight ambient-occlusion shadow directly beneath each ball's contact point. Currently the balls float — they read as a flat cutout. Grounding makes them read as real objects, which lifts perceived premium.
Luminosity drop (RGB 253,253,253)Low riskDrop the white background ~2 RGB points to 253,253,253. Creates a faint border in Amazon search-result tiles so the listing stops bleeding into the page. Recommended given the spiky-ball category is visually noisy.
Ghost packagingMedium riskMentioned, not recommended for this category. Would mean rendering a "ResultSport Massage Ball Set" branded box behind the three balls with size callouts printed on the package. The technique is widely used in supplements and beauty but rarer in sports recovery, where reviewers tend to be stricter. Skip unless competitor B07YVDWNN2 or B08LVWNLY1 adopts it first.

Top fix

Reshoot the hero with diagonal isometric + micro-shadow grounding. Same 3 balls, same colours (or pick a single representative SKU colour — see note below), but staged along the canvas diagonal with depth, and grounded with crisp ambient-occlusion shadows. This is a zero-compliance-risk move that routinely yields 5–10% main-image CTR lift in competitive UK categories.

Colour decision: the current 3-colour hero (teal / purple / green) reads as "a set of toys" rather than "3 sizes of one tool". Consider a single-colour set (e.g. all green) so the eye sees the size difference, not the colour difference. Or keep the 3-colour set but enlarge the size delta so the 10 cm visibly dwarfs the 6 cm.

6. SLOT 2 — Keystone (deep-dive)

Current slot 2 keystone
Current slot 2 · "3 Different Sizes" header · hand-held scale

Mechanism analysis

Current slot 2 attempts Mechanism #4 — Use-case framing ("here's how the sizes compare in-hand") with a side-of Mechanism #2 — Avatar affirmation (the visible hand suggests "you, holding these"). It does neither well. The "3 Different Sizes" header is in a small-caps generic font, the size labels (Big / Medium / Small) are tiny and stacked confusingly, and there's no risk-kill badge for any of the high-intent sub-niches (plantar fasciitis, trigger point, myofascial).

MechanismCurrently playing?Notes
1 — Category shortcutNoProduct isn't novel enough; not the right play here.
2 — Avatar affirmationPartially (hand visible)Not the recommended mechanism — buyer identity (runner / pilates / physio user) is too varied.
3 — Objection pre-emptNoRecommended mechanism. The size-to-body-zone confusion is the #1 objection. Slot 2 should answer "which ball for what muscle".
4 — Use-case framingPartiallyGeneric "3 sizes" framing without the body-zone specificity that would make it convert.
5 — Proof-of-outcomeNoHard to play compliantly in this category (medical-claim risk).
6 — Differentiation hookNoCould play "Not another single ball" as backup — but #3 is stronger.

8-dimension keystone scorecard (0–80)

1. Mechanism clarity
3/10
2. Thumbnail / 3-sec clarity
3/10
3. Risk-kill strength
2/10
4. Human-scale presence
7/10
5. Brand clarity
2/10
6. Typography craft
3/10
7. Category reframe
3/10
8. Personality / craft
5/10

Total: 28 / 80 — Broken tier. Rebuild from scratch.

Top 3 fixes

  1. Switch mechanism to Objection pre-empt (#3) playing the size-to-body-zone map. The headline becomes a noun phrase from the kw-research keystone — for example: "6cm Arch · 8cm Calf · 10cm Back" with a body silhouette behind it and a numbered ball overlaid at each anatomical position. This answers the #1 objection visually, anchors the sub-niche keywords (plantar fasciitis = arch = 6 cm), and replaces the generic "3 Different Sizes" header. Mobile-thumbnail readable because the three big numbers do the heavy lifting.
  2. Add a risk-kill badge anchored on the sub-niche search drivers. Top-corner pill in a calm brand colour: "Built for Plantar Fasciitis · Trigger Point · Myofascial". This single badge converts the slot from a generic size chart to a category-defining keystone, and seeds Rufus retrieval for the three biggest opportunity clusters.
  3. Bring craft up to brand-tier typography. Drop the small-caps generic header. Pair an oversized numeric (6 / 8 / 10) with a clean sans (Inter / GT Walsheim / similar) for the body-zone label, and a quiet ResultSport™ signature in the bottom corner. The hand-held photography element from the current version is fine to keep — it's the strongest of the 8 dimensions in the current build.

7. SLOT 3 — current

Current slot 3
Current slot 3 · hexagon-pattern background · 3 use-case illustrations

Current job: a second pass at "3 sizes" with use-case illustrations

DimensionScoreNote
Job clarity3/10Duplicates slot 2's "3 sizes" job. Use-case framing is buried in tiny illustrations with generic Large / Medium / Small labels.
Execution3/10Hexagon-pattern background looks like a stock template. Text illegible at thumbnail. Three illustrations crammed into one frame — feature dump pattern.
Compliance10/10Pass — no banned claims.

Biggest fix: repurpose this slot for materials safety (mechanism #3, objection #2). One clean shot of a ball next to a bare-skin contact frame, with one badge: "PAH-free · Phthalate-free · Glue-free · Solid plastic". Replaces what slots 4 currently tries to do — without the banned wording.

8. SLOT 4 — current (COMPLIANCE FLAG)

Current slot 4 with TOP QUALITY claim
Current slot 4 · "TOP QUALITY" and "TOP GRADE" badges visible
HIGH compliance risk. This slot contains the phrases "TOP QUALITY" and a "TOP GRADE" badge. Both are listed as flagged superlatives in Section 1 of the user's Amazon_Banned_Keywords_Guide (severity HIGH). The same audit also flagged "Top Quality Material" in the bullet copy in the listing-rewrite v3 — this is the same violation, just rendered as image text. Subject to listing suppression or auto-edit under the Jan 2025 policy.

Current job: material-quality claim

DimensionScoreNote
Job clarity5/10The intent (material safety reassurance) is correct — this is exactly the right job for a supporting slot.
Execution4/10Check-mark bullet list is on-pattern but compliance issues neutralise any benefit.
Compliance0/10HARD FAIL — multiple banned phrases.

Biggest fix: Pull this slot immediately. Replace with the same material-safety message rebuilt without banned wording. Move the material-safety play to slot 3 (per the slot 3 recommendation above), and use this slot for a clean use-case-in-context shot (mechanism #4) — e.g. office-worker placing the 8 cm against a chair lumbar, captioned "Built for desk-bound back tension".

9. SLOT 5 — current (COMPLIANCE FLAG — highest severity)

Current slot 5 with MAXIMUM HEALTH BENEFITS and Prevent Health Complications claims
Current slot 5 · "MAXIMUM HEALTH BENEFITS" header · "Prevent Health Complications" claim
HIGH compliance risk — highest severity in the stack. Three separate flags on this single slot:
  • "MAXIMUM HEALTH BENEFITS" — banned superlative (Section 1).
  • "Superb Comfort and Effectiveness" — "Superb" pattern matches the banned-superlative class.
  • "Speed Up Recovery for Sports Injuries, and Prevent Health Complications" — borderline-to-explicit medical claim ("prevent"). Section 1 lists "cure, treat, prevent" as the HIGH-severity health-claim trigger words.
The "prevent" wording is the most dangerous element on the listing right now. Amazon's automated review system, the FDA-adjacent compliance bot, or a customer report could trigger immediate suppression. This needs to come off before any other fix.

Current job: outcome / benefit messaging with model

DimensionScoreNote
Job clarity4/10Intent is "lifestyle outcome with avatar" — the right idea for a slot 6 lifestyle play, but executed as a feature-list compliance-fail instead.
Execution3/10Model photograph is fine; overlaid text turns it into a billboard.
Compliance0/10HARD FAIL — multiple banned phrases, including the most dangerous "prevent" wording.

Biggest fix: Pull this slot immediately. Rebuild as a clean proof-of-outcome lifestyle shot — keep the model (she fits the avatar — runner / pilates user demographic), but remove every claim. Replace with a single line in the same brand font: "For runners. For pilates teachers. For desk warriors." — names the three converting audiences without making a health claim. Move the materials messaging to slot 3, the use-case-in-context to slot 4, and let this slot be the aspirational pull.

10. SLOT 6 — current

Current slot 6 packaging shot
Current slot 6 · branded packaging · instruction leaflet · exercise chart

Current job: what's-in-the-box / packaging — but placed in the wrong slot

DimensionScoreNote
Job clarity6/10The shot itself is clear (3 balls + packaging + leaflet). But it's playing the slot 7 job in the slot 6 position — slot 6 should be the lifestyle / proof-of-outcome moment.
Execution6/10Composition is busy; the wooden-floor background fights the brand colour. "EXERCISE LEAFLET" sticker is helpful but underplayed.
Compliance10/10Pass.

Biggest fix: Move this slot's content to slot 7 and rebuild this position as the aspiration / proof-of-outcome lifestyle shot (taking over from rebuilt slot 5). That gives the stack its proper rhythm: 4 (use-case) → 5 (broad audience) → 6 (aspiration) → 7 (packaging closer).

11. SLOT 7 — Missing

Slot 7 doesn't currently exist as a distinct frame. The closer / comparison job is unowned.

Recommended: a 2-column comparison chart — "ResultSport Spiky Set vs. a single lacrosse ball" — with ✓/✗ rows for: number of sizes, body-zone coverage, fascia-layer reach, intent (sport ball vs. mobility tool), price-per-tool. Don't name specific competitor brands; "a single lacrosse ball" is the generic comparator. This slot directly addresses the third top objection ("how is it different from a tennis / lacrosse ball that I already own") and is a near-direct conversion of the comparison FAQ that the listing-rewrite v3 already drafted.

12. A+ Content review

The current A+ block has three components:

  • Product Description with embedded Rufus FAQ. 5 question/answer pairs already match the seeded format the user's reference PDFs recommend — "Can these help with plantar fasciitis? Are they suitable for sciatica relief? How do they help with muscle knots? Can I use these for reflexology? Do these balls help with recovery after running?" This is the strongest element of the entire listing's image work — keep it.
  • Cross-sell comparison module ("Other ResultSport Products"). Shows 7 sibling products (Spiky Set, Lacrosse Massage Balls, Lacrosse & Spike Set, Trigger Massager Ball, Handy Portable Massager, Finger & Hand Massager, Stress Gel Hand Ball Set) with prices, ratings, sizes. Above category average for a brand catalog module. Slightly defensive against single-product comparison though — could be repurposed.
  • Usage-illustration images. 6+ embedded usage illustrations covering body zones. Alt text quality is mixed ("a hedgehog sitting on a mat with a ball of fur" was AI-generated and unhelpful — Amazon's auto-alt fills in when the seller doesn't supply one). The user-supplied alt text isn't reaching the multimodal AI layer optimally.

Recommended 6-module rebuild (cluster-anchored)

ModuleJobAnchored on cluster
1 — Brand banner"3 sizes, one fascia toolkit" — the brand promise as one sentence over a clean product shotCore ("spiky massage ball")
2 — Size-to-body anatomical guideBody silhouette with 3 numbered balls at arch / calf / back. Same payload as the recommended slot 2 rebuild.Sub-niche (plantar fasciitis · trigger point)
3 — Use cases (4-tile)4 image+caption tiles: arch rollout · calf release · lower-back wall release · seated glute. Each tile heading as a question (Rufus retrieval).Foot massage ball · physio ball
4 — Materials & safetySpecs panel — PAH-free, phthalate-free, glue-free, machine-formed solid plastic, blunt spike tips. Replaces the banned slot 4 content compliantly.Objection #2 (safety)
5 — Comparison chartThis vs single lacrosse ball vs tennis ball. Substitutes for the missing slot 7. Same content as the eventual slot 7 build.Comparison (lacrosse / tennis)
6 — Brand story closeResultSport one-paragraph story + the sibling-product cross-sell (keep the existing module — it's strong).Brand defence

Each module needs both mobile and desktop variants. Every image needs human-written alt text (≤100 chars, 2 semantic variations) — don't rely on Amazon's AI-generated alt.

13. Stack-level compliance pass

CheckVerdictNotes
Slot 1 main-image rules (white BG, no text, no models)PASSFully compliant. Only opportunity is unused CTR push tactics.
Slot 2 — banned superlatives / health claimsPASSMechanism is weak but no banned wording.
Slot 3 — banned superlatives / health claimsPASSClean.
Slot 4 — banned superlativesHIGH FAIL"TOP QUALITY" + "TOP GRADE" — Section 1 violations. Pull immediately.
Slot 5 — health claimsHIGH FAIL"MAXIMUM HEALTH BENEFITS" + "Prevent Health Complications" — the "prevent" wording is the most dangerous element on the listing.
Slot 6 — banned wordingPASSClean. Wrong-slot placement, not a compliance issue.
A+ — banned wording / orphan claimsREVIEWAuto-generated alt text is sub-optimal. Embedded usage illustrations need claim audit.
Stack-wide — competitor brand mentionsPASSNo competitor brand names visible.

14. Measurement plan

The listing's monthly sessions sit roughly in the 5–15k band based on the SQPR Q1 impression data. That's below the threshold for clean A/B power on slot-2 only (Manage Your Experiments needs ~5k unique visitors/variant/cell), so the recommendation is 1P-decided sequential swap with pre/post measurement, not A/B testing.

PhaseWindowMetricDirection
Baseline (current state)Now → swap daySessions, unit-session %, main-image CTR (Brand Analytics), Rufus impressionsCapture 4-week average
Phase 1 — Compliance fix (slots 4 + 5 pulled / rebuilt)Day 1 → Day 14Sessions, unit-session %Defensive — watch for suppression risk reversal
Phase 2 — Slot 2 keystone rebuildDay 15 → Day 42Unit-session %, mobile CTR proxy (clicks / impressions)+5–12% unit-session % expected
Phase 3 — Slot 1 hero push tactics + Slot 7 comparison buildDay 43 → Day 70Main-image CTR, comparison-cluster purchase share+3–8% main-image CTR; +1–3 pts comparison cluster share
Phase 4 — A+ 6-module rebuildDay 71 → Day 100Unit-session %, Rufus answer surfacing rate (qualitative)+5–10% unit-session %

All lift estimates are ranges, not point figures, per the skill's house style. Single-variable lifts can compound — the compliance fix on its own is risk-mitigation (avoids suppression), not CVR upside.

15. Next-step action list

This week (compliance triage)

  • Pull slot 5 from Seller Central today. "Prevent Health Complications" is the single most dangerous line on the listing. Replace with a clean lifestyle shot (the existing model photo without the text overlay) until the rebuild lands.
  • Pull slot 4 within the same window. "TOP QUALITY" + "TOP GRADE" — Section 1 superlative violations. Replace with a placeholder of the same image minus the badges if a clean rebuild isn't ready.
  • Audit bullet 4 + bullet 5 in Seller Central against the same banned-keyword list. The listing-rewrite v3 already flagged "Top Quality Material" in bullet 4 — confirm it's been replaced live, not just in the rewrite draft.

Within 2 weeks (slot 2 rebuild)

  • Brief the slot 2 keystone rebuild per Section 6.3. Mechanism: Objection pre-empt. Headline: 6cm Arch · 8cm Calf · 10cm Back. Risk-kill badge: "Plantar Fasciitis · Trigger Point · Myofascial".
  • Brief slot 3 as the new materials-safety frame (the legal version of slot 4's old job).
  • Brief slot 4 as the new use-case-in-context frame (desk-bound back tension scene).
  • Brief slot 5 as the new audience-affirmation lifestyle shot (with no overlay claims).

This month (full stack landing)

  • Brief slot 6 as proof-of-outcome / aspiration. Move current slot 6 packaging content to slot 7.
  • Brief slot 7 comparison chart (vs single lacrosse ball).
  • Apply CTR push tactics to slot 1 (diagonal isometric + micro-shadow grounding + luminosity drop).
  • Run the amazon-aplus-content skill to brief the 6-module A+ rebuild per Section 12.
  • Re-shoot all 7 slots at Amazon-compliant 2000×2000 sRGB JPG.

16. Appendix

Framework: amazon-listing-images skill — slot 1 (CTR push tactics + 4-dim scorecard), slot 2 (6-mechanism + 8-dim 0–80 scorecard), slots 3–7 + A+ (3-dim supporting-stack rubric).

Listing URL: amazon.co.uk/dp/B00CCPNDWI

Data sources used in this audit:

  • Live PDP scrape (2026-05-22) — 6 gallery images + A+ Content text, scraped via Claude in Chrome.
  • SQPR Q1 2026 (kw-source-sqpr.csv) — cluster ↔ slot coverage mapping.
  • Sellersprite reverse export (kw-source-sellersprite.xlsx) — ABA rank validation for sub-niche opportunity.
  • Listing-rewrite v3 (listing-rewrite.html) — buyer avatar, top objections, banned-claim findings.
  • Prior slot-2-only audit (2026-05-21, legacy folder) — baseline mechanism analysis; superseded by this run.
  • Rules canon: Amazon Search Algorithm & Listing Optimization.pdf (Mar 2026) + Amazon_Banned_Keywords_Guide.pdf (Apr 2026).

Note on image rendering: the slot thumbnails in this report are loaded directly from Amazon's CDN (m.media-amazon.com). The HTML file renders fully when opened with internet access. For an offline / fully portable copy, the images can be base64-embedded on request — say the word.