Creative experiments

The two images look identical. The ad names the product clearly, gives two specific feature claims (rounded base, SpoutGuard) with icons, and shows the size range. The hero shot features the real bottle with its whisk ball visible, so the product is the centre of the ad. The layout is clean and easy to read on a phone. It is a competent spec sheet rather than a creative idea.

Control check: The bottle matches reference image 15: green lid, frosted grey cup, navy flip cap and loop, vertical navy BlenderBottle logo, embossed lid wordmark, and a whisk ball visible inside. The cup reads slightly clearer than the reference, which is acceptable.; The icon next to 'Rounded base for improved mixing.' looks like a wire whisk spiral, not a rounded base. That could suggest the BlenderBall is the feature being described. This is minor and the copy is still supported by blenderbottle-fact-1.; The three size silhouettes are generic icons drawn to scale. They don't make any extra claim, and the size text matches blenderbottle-fact-3.; At phone size the size text and the feature text are both legible. There is no extra text beyond the supplied lines and the real product label.

Variation check: This image looks identical to blenderbottle-control, so the variant gives no creative contrast for the test.; Same notes as the control: the bottle matches reference image 15, and the 'rounded base' icon looks more like the whisk than a rounded base (minor). Text is legible and there is no extra text.

Primary source

The two images look identical. It sells one specific feature (recurring tasks) and shows it as the same repeat-icon task row on five devices, matching the cross-platform subline. The illustration is clean and on brand, and the message is quick to grasp. It is simple but somewhat generic, with plain placeholder task bars.

Control check: The device illustration is a generic workflow: a checkbox circle, a placeholder bar and a repeat icon on a phone, tablet, watch, laptop and browser window. It doesn't pose as a real Todoist screenshot and shows no capability beyond recurring tasks (todoist-fact-2).; 'Desktop' and 'web' are represented by the laptop and the browser window. The browser window has no clear stand or monitor body, so 'desktop' as a separate device type is only loosely shown. This is minor.; 'Todoist' is set in red type as plain text, with no invented logo mark. All lines are legible at phone size.

Variation check: This image looks identical to todoist-control, so the variant gives no creative contrast.; Same notes as the control: the generic illustrative UI is acceptable, and 'desktop' is only loosely shown apart from 'web'.

Primary source

The two images look identical. The storm sky and lit house set a clear, timely mood. The support cables in the tree quietly show the service being sold, which ties the offer into the picture. The serif headline reads well against the dark sky. The subline is fairly small at phone size, and the brand name 'Davey' alone says little about what the company is.

Control check: The scene is a conceptual photo: a large tree near a house under storm clouds, with cables in the canopy that suggest a support system (davey-fact-2). It isn't presented as actual Davey work, which is acceptable.; At phone size the canopy cables are thin and could read as overhead power or utility lines instead of tree cabling. That weakens the 'support systems' idea and could raise a line-clearance safety concern.; 'Get your trees ready before the storm' is a call to action, not a guarantee against damage. It is within what the storm-prevention record supports. There is no locality, credential or guarantee.; The small subline in sans-serif over the dark sky is legible. 'Davey' appears as plain serif text, with no invented logo.

Variation check: This image looks identical to davey-control, so the variant gives no creative contrast.; Same notes as the control: at phone size the canopy support cables could be mistaken for power lines.

Primary source

Both use the same before/after slider idea, which fits a design-build service well. X runs the text over the photo, so the long subline is small and has to compete with the ceiling area. Its slider line runs horizontally across a vertical split, which looks wrong, and the 'after' side shows less of the finished kitchen. Y puts the headline and supporting copy on clean bands of their own, so it reads better at phone size. It frames the before/after as a tidy panel, and its 'after' kitchen is fuller and more aspirational, with pendants, an island and cabinetry. Y's slider device is also a bit odd, but overall it is the clearer and more polished ad.

Control check: The two halves aren't the same room: the window placement, ceiling and layout differ. So the 'before/after' framing is internally inconsistent as well as unsupported.; The slider graphic has handles at both outer edges and none at the split. This is a confusing, non-functional UI element.; The brand name is used as the headline with no invented logo. The copy matches the supplied lines (the third line is wrapped across two lines) and is legible at phone size.; Suggested fix: use one conceptual scene (for example a design meeting with material samples, or a single in-progress or finished space) instead of a before/after split, or clearly label the image as illustrative.; The before/after layout, with a bare-stud room on the left and a finished high-end kitchen on the right, implies a project outcome and a level of finish. The brand's limitations explicitly exclude 'luxury', 'flawless' and outcome promises.; The split before/after image with a slider line and handles is a format normally used to show real completed projects. Here it is generated and has no label saying it is illustrative, so viewers may read it as Turn Key Designs client work.

Variation check: The supplied lines 3 and 4 ('Design-build studio.' / 'Procurement, coordination & oversight included.') are small, dark text over the light ceiling area of the photo. They have low contrast and are only marginally legible at phone size. The control's solid band was clearer.; The kitchen run on the left half doesn't fully line up with the right half: the hood and cabinetry positions are inconsistent at the split.; Suggested fix: drop the before/after slider device or label the image as illustrative, and put the small copy on a solid band or enlarge it.; The finished half shows a high-end kitchen as the end state of a 'before/after'. This implies a quality and outcome promise, and the limitations exclude 'luxury' and outcome claims.; This is an even stronger before/after slider than the control: one continuous room split at the center, with a slider handle in the middle. It convincingly mimics a real project transformation and isn't labeled as illustrative, so it reads as Turn Key Designs client work.

Primary source

The two images look identical. The flat-lay of craft projects (typography sketch, architecture model, yarn, ink drawing, camera) backs up the 'real-world projects' message and the list of categories. The composition is rich and well lit. The phone is a small nod to online learning, but the course platform itself is barely shown. The category line is small at phone size.

Control check: The flat-lay props (camera, yarn and needle, botanical ink drawing, type-sketch notebook, architectural model, phone with abstract art) map to the creative subjects in domestika-fact-2. They are conceptual and aren't presented as specific student or course work.; The notebook shows large drawn letterforms ('Aa', 'B', 'g'). These are decorative type specimens, but they are readable glyphs beyond the supplied lines. This is minor; check it is acceptable under the no-extra-text rule.; The subject line 'Design · Illustration · Photography · Crafts' is the smallest text and sits close to busy props. It is legible at phone size but marginal.; The phone screen shows abstract art, not a Domestika app UI, so no app capability is invented. 'Domestika' is plain text with no logo.

Variation check: This image looks identical to domestika-control, so the variant gives no creative contrast.; Same notes as the control: the decorative 'Aa B g' letterforms are readable glyphs beyond the supplied lines, and the subject line is small.

Primary source

The two images look identical. The offer is very specific: a Registered Dietitian Nutritionist over Zoom, plus messaging and meal logging in the app. The scene shows each part, with a video call on the laptop, a chat and meal log on the phone, and a healthy bowl in front. The warm, cohesive style and clear type hierarchy work well. The illustrated faces on the call look slightly generic.

Control check: The video-call screen is a generic illustrated UI with two cartoon people. It has no Zoom logo and doesn't pose as a real Zoom or Nourish screenshot, or as an actual Nourish dietitian. That is acceptable as illustration.; The phone shows a generic chat and meal-log card, which matches nourish-fact-3. It doesn't add progress charts or outcome metrics.; The grain bowl suggests healthy eating, but there are no weight-loss, clinical or insurance claims. The 'Registered Dietitian Nutritionist' credential is supported by nourish-fact-1.; Faint lettering may appear on the laptop's bottom bezel. Confirm there is no third-party device brand text visible at full resolution.; All lines are legible at phone size. 'Nourish' is plain serif text with no invented logo.

Variation check: This image looks identical to nourish-control, so the variant gives no creative contrast.; Same notes as the control: the generic illustrated call UI is acceptable with no Zoom logo. Confirm there is no device brand text on the laptop bezel.

Primary source

The two images look identical. The idea is clear and specific: three equipment setups (dumbbells, rack, band), each with an app screen that picks the matching workout. This shows the product's personalization well. However, the headline has visible smearing on 'Fitbod', 'your' and 'goals', which hurts the craft. The bottom of the frame is also empty space.

Control check: There are visible text-rendering errors. The 'Fitbod' wordmark has smeared and blurred letter edges, especially 'tb' and 'd'. The headline glyphs in 'your', 'goals' and 'equipment' are ghosted or doubled where the lime diagonal stripe and its glow overlap. The subline 'personalized strength' also looks soft. At phone size this reads as a rendering defect, not a deliberate effect.; The three phone screens show generic equipment-selection cards (dumbbells, bench, pull-up bar, handles, band). They are illustrative and in line with fitbod-fact-2 (available equipment). Make sure they aren't presented as the real Fitbod UI.; There are no outcome or performance claims, and no platform or price is stated. The bottom part of the canvas is mostly empty.; Suggested fix: re-render the text cleanly, keeping the lime stripe away from the type.

Variation check: This image looks identical to fitbod-control, so the variant gives no creative contrast.; It has the same visible text-rendering defects: the 'Fitbod' wordmark and the words 'your', 'goals', 'equipment' and 'personalized strength' are smeared or ghosted where the lime stripe and its glow overlap. This is noticeable at phone size.; The equipment-card screens are illustrative only.

Primary source

X and Y look identical, so this is a tie. The product fills the frame in use, wrapped around a person at a sunset overlook, and its printed pattern echoes the landscape. That's a nice touch. 'The Original Puffy' and the subline name the category clearly. 'For out there' is a bit vague, and the warm lantern-and-sunset setup is a familiar outdoor-brand formula.

Control check: The blanket's print matches the reference: a misty grey/white top fading into dark pines, a bright green meadow in the centre and teal binding and backing. Draping the blanket makes the quilting lines look more horizontal than the reference's arched contour stitching. This is acceptable, but a close comparison would show the difference.; The real label (the Rumpl mark on the teal corner) is hidden in the drape. That's fine, and nothing was invented in its place.; The ad shows the Olympic Fog-style colorway, but the records don't confirm that colorway is still sold. The copy makes no stock or colorway claim, so this is a note, not a failure.; The sunset, lantern and model are a conceptual lifestyle scene. 'For out there' is a reasonable reading of the 'outdoor blankets' record (rumpl-fact-1). No warmth or performance claims.; All three lines are legible against the sky at phone size. The 'Rumpl' line is small but readable.; This ad looks the same as rumpl-variant, so the pair doesn't test any difference.

Variation check: This ad looks the same as rumpl-control, so the pair doesn't test any difference.; The print matches the reference product (misty pines, green meadow, teal binding). Draping makes the quilting lines look flatter and more horizontal than the reference's arched stitching. Acceptable.; The colorway's availability isn't established in the records. The copy doesn't claim it's available.; The lifestyle scene with lantern and sunset is conceptual. The copy is supported by rumpl-fact-1 and rumpl-fact-2.

Primary source

X and Y look identical, so this is a tie. It shows the feature well: a search bar and an answer card with numbered citations wired to recognizable apps. The diagram is clean and on-brand. But the placeholder grey bars and tiny citation chips don't say much at phone size, and a few connector lines end vaguely (for example, the doc icon's line). The large Notion wordmark anchors the layout.

Control check: The third-party brand logos make specific integration claims that aren't in Notion's records. Replace them with generic, unbranded app tiles, or cite a record that names each app.; Visual error: in the bottom-left, a connector line from citation marker 2 runs left and ends in empty space instead of reaching an app tile.; The search bar, answer card and numbered markers 1–3 work as an illustrative, wireframe-style workflow and don't pose as a real screenshot. That's acceptable.; The copy matches notion-fact-1 ('across all your apps') and notion-fact-2 ('with citations'). The 'Notion' wordmark is plain text, not a logo copy.; This ad looks the same as notion-variant.; The Google Drive, Gmail, Slack, Google Calendar and Dropbox logos imply that Notion connects to those specific apps. The records only support the general wording 'search across all your apps' (notion-fact-1) and never name an app.

Variation check: This ad looks the same as notion-control.; Replace the branded third-party logos with generic app tiles, or cite a record for each integration.; Visual error: a connector line from citation marker 2 runs left and ends in empty space.; The wireframe workflow reads as illustrative, which is acceptable. The copy matches notion-fact-1 and notion-fact-2.; The Google Drive, Gmail, Slack, Google Calendar and Dropbox logos imply specific integrations that aren't in Notion's records. notion-fact-1 only supports the general wording 'all your apps'.

Primary source

X and Y look identical, so this is a tie. It's an attractive, sunlit shot of a mature tree by a home, and the serif type is readable. But 'Tree services for your home.' is generic, and nothing shows the service itself: no arborist, no care, no specialist claim. The image is just a nice tree, so the brand's expertise isn't part of the idea.

Control check: 'Tree services for your home' is supported by the residential tree services record (bartlett-tree-experts-fact-1).; The large healthy tree in a front yard is a conceptual scene. No crew, before/after or result is shown, so it doesn't read as proof of Bartlett's work.; The dark serif headline over the bright, hazy sky is legible at phone size. The 'Bartlett Tree Experts' line is small but readable.; The offering is broad ('tree services') because the source only gives category labels. That's acceptable.; This ad looks the same as bartlett-tree-experts-variant.

Variation check: This ad looks the same as bartlett-tree-experts-control.; The copy is supported by bartlett-tree-experts-fact-1. The yard scene is conceptual and shows no work or result.; The text is legible against the bright background at phone size.

Primary source

X and Y look identical, so this is a tie. The kitchen is polished and aspirational, and the blueprints on the island quietly tie in the design-and-build service. 'Full-service kitchen remodels' states the offer. The headline 'Where design meets craftsmanship.' is a stock category line, and the image reads more like a finished showroom than a sign of what makes this brand different.

Control check: The headline matches 'combining design and craftsmanship' (dreammaker-bath-kitchen-fact-2). 'Full-service kitchen remodels' matches dreammaker-bath-kitchen-fact-1.; The finished kitchen with blueprints is a generated, conceptual scene and isn't labelled as a completed project. Risk to watch: a viewer could read it as a portfolio shot, so keep it out of placements that say 'our work'.; The brand name is set as a two-line serif block that looks a bit like a logo lockup. It isn't an official logo and is acceptable as plain text.; The ad doesn't mention the local-franchise delivery model, and no locality is claimed. That's fine.; This ad looks the same as dreammaker-bath-kitchen-variant.

Variation check: This ad looks the same as dreammaker-bath-kitchen-control.; The copy is supported by dreammaker-bath-kitchen-fact-1 and dreammaker-bath-kitchen-fact-2.; The kitchen is a conceptual render and must not be presented as a completed DreamMaker project.; The serif brand block looks a bit like a logo lockup but is plain text, so it's acceptable.

Primary source

X and Y look identical, so this is a tie. This is the strongest concept in the set. 'No signal? Keep learning.' pairs a specific feature (offline viewing) with a believable situation: a learner on a train sketching the landscape along with a class on a tablet. The product is central to the idea. The text is clear on a phone. The one gap is that the tablet doesn't visibly signal 'offline' or 'downloaded'.

Control check: 'Watch classes offline' is supported by skillshare-fact-3 ('offline viewing'). Nuance: the record describes offline viewing as a member feature, and the ad doesn't mention membership. That's acceptable for a short headline.; The tablet shows a generic video player with an illustration and no Skillshare interface, so it reads as illustrative, not as a screenshot.; The train scene with a blurred window shows travel without signal well, and the sketchbook suggests a creative class without claiming any outcome.; The white text is legible over the window and the dark interior at phone size.; This ad looks the same as skillshare-variant.

Variation check: This ad looks the same as skillshare-control.; Offline viewing is supported by skillshare-fact-3. The record ties it to membership, which the ad leaves out. Acceptable.; The tablet's video player is generic and illustrative, with no copied Skillshare interface. The text is legible.

Primary source

X and Y look identical, so this is a tie. The headline says plainly what you get: a session with a registered dietitian. The consult scene with a fruit bowl and notebook backs that up. The ad is generic and stock-like, though. 'Fay connects you.' is small and there's no logo, so the brand is easy to miss. The image shows the service, but nothing in it is specific to Fay as the matchmaker.

Control check: The copy matches fay-fact-1 ('connects members with registered dietitians'). The face-to-face scene fits 'Meet online or in-person' (fay-fact-2).; The smiling woman is a generated figure that viewers will read as a registered dietitian. That's acceptable as an illustration, but it isn't a real Fay clinician and shouldn't be presented as one.; There are no weight-loss, clinical, insurance or cost cues. The fruit bowl is neutral.; The brand name appears only inside 'Fay connects you.' and is fairly small. It's legible at phone size but carries little brand weight.; This ad looks the same as fay-variant.

Variation check: This ad looks the same as fay-control.; The copy is supported by fay-fact-1, and the in-person scene by fay-fact-2. The dietitian figure is generated and illustrative.; There are no clinical, weight-loss or cost cues. 'Fay' appears only inside the subhead, so brand presence is light.

Primary source

X and Y look identical, so this is a tie. The offer is very clear: one-on-one coaching by video. The laptop showing the coach in a home-workout setting turns the service itself into the visual idea. The lime-on-black type is bold and reads well on a phone, and the brand shows up twice. The scene is fairly standard, which keeps craft at 4.

Control check: Extra text: a standalone 'Caliber' title sits top-left. It isn't one of the intended lines ('Your coach. One-on-one.' / 'On video, with Caliber.'), and it looks like a made-up logo. Remove it, because the brand already appears in the second line.; 'One-on-one' and 'On video' are supported by caliber-fact-1.; The coach on the laptop screen is a generated person in an illustrative video call, not a real Caliber coach.; The dumbbell, mat and home setting don't promise any body change or measurable result. Acceptable.; This ad looks the same as caliber-variant.

Variation check: This ad looks the same as caliber-control.; Extra text: a top-left 'Caliber' title isn't one of the intended lines and looks like a made-up logo. Remove it, because the brand already appears in 'On video, with Caliber.'; The copy is supported by caliber-fact-1. The on-screen coach is a generated, illustrative person, and the ad promises no measurable or body-change result.

Primary source

Both put the product at the center. X shows hands opening the twist top to reload, but you can't actually see the 'bag stabilizer' it mentions. Y deals with a concrete purchase worry (fits any standard roll) and shows it right away with colorful rolls from different brands around the carrier. It's bright and easy to read on a phone. Y's benefit is a bit more specific and easier to see.

Control check: The carrier matches the terracotta reference (images 15 and 16): rounded cylinder, side window, seam near the base, elastic loop with a sleeve at the end, and a peach bag coming out of the window.; The 'Fits any standard-size poop bag roll' line is supported by wildone-fact-2 ('works with any standard size poop bag rolls').; 'Bring your usual rolls' quietly assumes the viewer's usual rolls are standard size. This is acceptable because the subhead states the limit.; The pastel bag rolls have no labels and no brand. They are not presented as Wild One products, which is fine because the bag roll offer is excluded from the records.; All text is legible at phone size, and there is no extra text.

Variation check: The ad invents how the carrier opens. The cap is shown lifting off near the top, with a central spindle or post going into the roll core. The reference images (15, 16, 17) show only one seam, near the base, and no internal mechanism. The ad keeps that base seam and adds a second split near the top, which contradicts the reference.; The 'built-in bag stabilizer' in the subhead is visually implied to be this invented post. The records say only that there is a stabilizer, not what it looks like.; The physical setup is confusing: the roll sticks up above the open rim while a bag is still coming out of the lower side window.; The cord is thicker than the elastic band in the reference and continues into a long rope loop across the background that looks like a leash. It is unclear what is part of the product.; The text matches wildone-fact-1 ('Twist-top & built-in bag stabilizer') and is legible at phone size.; Suggested revision: show the closed carrier from the reference with a hand twisting it, or show the reload without inventing internal parts.

Primary source

The images are identical. A bold play button with a cursor clicking it shows one-click tracking, but the ad never says what gets tracked (time), so first-time viewers may not get it. The graphic is clean and on-brand.

Control check: Both lines are supported by toggl-track-fact-1 ('Track time in the background or with a single click').; The play button and cursor are an abstract illustration, not a fake app screenshot. No capabilities are invented.; The offering is understandable mainly from the name 'Toggl Track' plus 'Tracked.' The art does not say it tracks time. This is acceptable but slightly abstract.; All text is legible at phone size, and there is no extra text.

Variation check: This image is pixel-identical to toggl-track-control (same lines, same art), so the pair cannot be compared. The variant needs to be regenerated if a real A/B test is intended.; The content itself is fine: both lines are supported by toggl-track-fact-1, the art is illustrative, and the text is legible.

Primary source

The images are identical. 'Too' plus 'From the tree care team' cleverly ties a holiday-lighting add-on to the core tree service. The lit tree and house scene is attractive, and the call to action is clear.

Control check: All three lines are supported: holiday lighting by savatree-fact-3, the tree care team by savatree-fact-1, and the consultation call to action by savatree-fact-2.; The house at night with a lit tree is photorealistic and has no label, so viewers could read it as a real SavATree project. It is not presented as client work, so it is acceptable as a concept image, but it is not evidence of the company's work.; The scene also shows a lit door wreath, lit potted topiaries and landscape uplights. The records say only 'custom holiday lighting services', so the exact scope (wreaths, shrubs) is unconfirmed. This is minor.; The button-style call to action and all text are legible at phone size. The headline sits on busy branches but has enough contrast.

Variation check: This image is identical to savatree-control, so it is not a real variant.; As with the control, the photorealistic lit-house scene has no label and could be read as real SavATree work. It is not presented as client work, so it is acceptable only as a concept image.; The lit wreath, topiaries and uplights go beyond the confirmed record wording ('custom holiday lighting services'). This is minor.; All lines are supported by savatree-fact-1, -2 and -3, and are legible at phone size.

Primary source

The images are identical. 'Tub out. Shower in.' plus a before-and-after line drawing with an arrow shows the conversion service right away. The style is clean and consistent.

Control check: The 'Tub-to-shower conversions' line is supported by re-bath-fact-1.; The before-and-after is clearly a line drawing, so it will not be mistaken for a real project photo. The specific shower features shown (glass door, rain head, hand shower, niche) are illustrative, not promised.; All text is legible at phone size, and there is no extra text.

Variation check: This image is identical to re-bath-control, so it is not a real variant.; The content is fine: the claim is supported by re-bath-fact-1, the line-art drawing is clearly conceptual, and the text is legible.

Primary source

The images are identical. 'Write it. Run it.' with an editor and run-button mockup clearly says you code in the browser. The code shown is generic placeholder bars, which makes it feel less specific to the product.

Control check: The claim is supported by codecademy-fact-1 ('write and run code right in your browser').; The editor mockup uses placeholder bars, not real code or a Codecademy screenshot, so it reads as an illustrative workflow.; The offering is only implied as 'learning'. The ad shows coding in the browser but does not say it is a learning platform; only the brand name hints at that. This is acceptable.; All text is legible at phone size, and there is no extra text.

Variation check: This image is identical to codecademy-control, so it is not a real variant.; The content is fine: the claim is supported by codecademy-fact-1, the editor mockup uses placeholders, and the text is legible.

Primary source

X has more facts (1:1 video visits, registered dietitians) and strong serif type, but the overlapping circles are an abstract, familiar metaphor that doesn't show the service. Y's picture acts out its headline: you scroll through dietitian cards and tap one, which shows the 'choose a specialist' benefit as an action. Y's faceless flat avatars are a bit template-like, but the product is more central to the idea.

Control check: '1:1 video visits · registered dietitians' is supported by berry-street-fact-1.; 'Just you and your dietitian' suggests one-on-one sessions, which fact-1 supports. It makes no outcome, privacy-guarantee or weight-loss claims.; The overlapping circles are purely abstract. The ad does not show what the service is beyond the subhead, but the subhead makes it clear enough.; The serif 'J' in 'Just' has an unusual descender but is still readable. All text is legible at phone size.

Variation check: The claim is supported by berry-street-fact-2 ('Pick a registered dietitian who specializes in what you need').; The carousel of faceless profile cards with a tapping hand is an illustrative picker. It is not a real Berry Street screen, and it does not claim how many dietitians there are or which specialties are offered. Acceptable.; The headline does not say 'registered' dietitian. That is not misleading, just less specific than the record.; The word spacing in the headline ('Pick the', 'who fits') is tight but readable at phone size. There is no extra text.

Primary source

X's split frame (warm living room vs. dark gym, with the same app on a phone in each) shows 'Home today. Gym tomorrow.' immediately, and the app links the two halves. Y's 'built around your gear' is a specific benefit, but the photo just shows gear next to a phone, so the personalization isn't shown. Both are well made. X's concept reads faster at phone size.

Control check: There is extra visible text: a standalone 'Centr' wordmark at the top that is not in intended_lines (only 'Home today. Gym tomorrow.' and 'Train in either with Centr Coach' were supplied).; That wordmark is set like a masthead logo and could be read as Centr's official logo, which was not supplied.; The claim is supported by centr-fact-2 ('Train at home or in the gym').; The phone screens show a generic workout-player interface (figure, progress bar, play button). This is illustrative, not a real screenshot. Acceptable, but these are not real Centr Coach screens.; The split-scene photography is coherent. The subhead is legible at phone size but sits on a busy split background; contrast is adequate.

Variation check: All lines are supported by centr-fact-1 ('Personalized workouts built around your goals and your equipment').; The dumbbells, kettlebell and mat are generic. Centr also sells equipment, and no Centr equipment reference was supplied, so viewers might read this gear as Centr products. It is not labeled as Centr equipment, so this is minor.; The dumbbell heads show small, illegible embossed marks, which is stray pseudo-text. They are not readable at phone size.; The phone screen shows a generic workout-player interface. This is illustrative, not a real Centr Coach screenshot.; The text sits on a bright window background with adequate contrast and is legible at phone size.

Primary source

Both use nearly the same towel photo on stone with the same specific claims (770 GSM Aegean Turkish cotton, zero-twist). X leads with the product name 'Super-Plush Bath Towels' in large type and shows the towels bigger, so the terry texture reads well. Y leads with the brand and centers the text. Its bolder body copy and terracotta accent on the GSM line read a little better on a phone. Each has a slight edge somewhere, so overall it's even.

Control check: The towels match the reference: a light grey pair, folded and stacked, with the flat woven hem band. All four lines match the records (brooklinen-fact-1, brooklinen-fact-2) and are readable at phone size.; Minor: there is an extra white folded cloth in a bowl at the bottom left. It looks like a prop, but it could be read as another product. Consider cropping it out.

Variation check: The towels match the reference (grey, two stacked, hem band). The centered serif and sans lines match the records and are readable. The brown accent line '770 GSM Aegean Turkish cotton' has slightly lower contrast but is still readable.; Minor: the soap bar in a dish is a neutral prop and doesn't suggest a product.

Primary source

X has a strong headline ('Set your hours. Get paid up front.') and clear supporting lines. The phone confirmation screen and laptop calendar show the payments-and-booking idea on both devices it mentions. Y has the same copy and visuals, but the subhead runs into the laptop, and the comma after 'packages' is clipped. Its blue pill also has the broken rectangle. These visible flaws lower clarity and craft.

Control check: The laptop calendar and the phone check-mark screen are generic mockups with no invented features or numbers. They read as illustrations, not real screenshots.; Minor nuance: 'On mobile or in your browser' slightly widens the record's 'mobile app or browser extension' (calendly-fact-3). It's acceptable, but 'browser extension' would be the literal wording.; The light grey 'On mobile or in your browser' line has low contrast on the off-white background at phone size. It's still readable.

Variation check: Visible layout error: the line 'Charge for meetings, sell packages,' runs into the laptop screen's top-left bezel. The trailing comma sits on or against the screen edge, and 'send invoices' wraps awkwardly underneath.; Visible rendering artifact: the blue 'On mobile or in your browser' pill has a square-cornered rectangular tab sticking out of its right side, so the shape looks broken.; The same nuance as the control applies: 'in your browser' slightly widens 'browser extension' (calendly-fact-3).

Primary source

Y uses a two-tone 'Monster / Tree Service' headline and three services with matching icons over a dramatic low-angle shot of a healthy tree. It's clean and easy to scan. X boxes the text in heavy dark-green panels, with lime bars under each service that look glitchy and cluttered and block much of the photo. Neither shows the service being done, but Y presents the offer much more cleanly.

Control check: The backyard tree scene is conceptual. It shows no crew, equipment or before-and-after, so it doesn't suggest client work. The three service lines match the records (fact-1 to fact-3), and the icons (shears, leaf, tree) fit them.; The text sits on a soft dark gradient and is readable at phone size.

Variation check: Visible layout artifacts in the services panel: the thick lime bars under each item look like misplaced highlight or progress bars, not deliberate dividers. They don't line up with the text, and the last bar is thinner and cropped.; The service labels have almost no padding: 'Pruning & Trimming' presses against the top edge of the panel. The header box, by contrast, has a lot of empty space on the right, so the two panels look mismatched.; The scene is conceptual with no crew or work shown, which is fine.

Primary source

Both are simple brand-plus-descriptor ads built on an upscale modern house at dusk. The house is the product proof. X shows the whole house, with lake, mountains and sunset framed below a centered serif title and a warm gold descriptor, which is a polished and balanced layout. Y's larger left-aligned type reads well, but the house is cropped at the right edge and partly hidden, which weakens the showcase of what they build.

Control check: The modern lakeside house is generated. It shows no address, project name or 'our work' wording, so it reads as mood imagery and not as a portfolio piece. Keep it unlabeled as an actual build.; The ad gives no location or way to get in touch, which matches the records' limitations. The offering ('Custom Home Builders', alair-homes-fact-1) is clear.

Variation check: The generated house is conceptual and carries no wording that presents it as a completed Alair project. Keep it that way.; The gold 'Custom Home Builders' text is fairly small against the blue sky but readable at phone size.

Primary source

Both show a moody evening learning scene with a tablet showing abstract art rather than real lesson content, so the leadership offer only comes from the copy. Y puts the MasterClass name and 'Leadership skills, in bite-sized lessons' in clean dark space at the top, and the hand-held tablet feels more like casual, bite-sized viewing. In X the text sits low over desk clutter and the sweater, which makes it busier. 'Online classes' is also small and floats far from the rest of the text.

Control check: The tablet shows abstract artwork, with no fake instructor, class title or interface. The lines match masterclass-fact-1 to fact-3.; The scene is very dark overall, but the white text has strong contrast and is readable at phone size.

Variation check: The headline sits on top of the desk, notebook and pen. Contrast is still enough on the dark background, but the busy area behind 'MasterClass' cuts clarity slightly at phone size.; The tablet shows abstract artwork and no invented class content.

Primary source

X has a clear hierarchy: a large 'Virtual nutrition counseling' headline, two icon-led benefits, and a clean CTA pill. Its laptop video call fits the one-on-one or group idea. Y uses the same message, but the type is stacked tightly in one green tone with weaker hierarchy. Its CTA button is broken, with a rectangle sticking out of the pill's right edge. That looks unfinished on a phone.

Control check: The video call with several participants is a fair illustration of the group/one-on-one option (culina-health-fact-1). The ad makes no weight-loss, body or clinical promises, and the fruit bowl is neutral.; The CTA 'Book your first session online' is supported by culina-health-fact-2. All lines are readable and the layout is clean.

Variation check: Visible layout error: 'Virtual nutrition counseling' runs right up to the hanging picture frame, and the final 'g' touches or overlaps the frame edge.; Visible rendering artifact: the green CTA pill 'Book your first session online' has a square rectangular tab sticking out of its right side.; Minor: the laptop bezel shows garbled pseudo-text below the screen. A four-person video grid is a fine illustration of group counseling, and the ad makes no health-outcome claims.

Primary source

In X, 'Know today's workout' is big and clear, and two icon-led benefits match the laptop, which shows a calendar with highlighted days. That makes the interactive-calendar idea visible. Y's two-tone headline is fine, but the laptop calendar is small. The orange 'At-home workout videos' pill has the same broken rectangle on its right edge, which hurts finish.

Control check: The calendar on the laptop is an abstract grid with no invented program names or numbers, so it reads as an illustrative workflow. The ad makes no fitness or body-result promises.; 'Know today's workout' is a reasonable restatement of the day-by-day calendar guidance (fitness-blender-fact-2). The lines are readable at phone size.

Variation check: Visible rendering artifact: the orange 'At-home workout videos' pill has the same square rectangular tab sticking out on its right side.; Minor: putting a feature line in a button-shaped pill makes it look like a tappable CTA, though it describes the offering and isn't an action.; The calendar on the laptop is abstract and illustrative, and the ad makes no body or performance promises.

Primary source

X's transit-map idea, 'Same glass. Every stop.,' turns the fridge, freezer, microwave and oven uses into a route the same container travels. It's clean, clever and instant to read on a phone, and the product is the idea itself. Y is a well-lit lifestyle product shot with icons and more detail (Air Release Technology, magnetic organizers). It is more informative but has no concept, and the icons repeat the text. X is the stronger, more memorable ad.

Control check: Containers match the reference: cream bodies, grey glass-window lids and the small 'Caraway' lid label. Sizes and stacking are plausible.; The sage-green organizer holding the lids is not in the product reference. Its color and form can't be checked. The magnetic feature is only stated in text, not shown.; Fact caraway-fact-1 says the containers are oven-safe. The appliance icons and headline sit above lidded containers, which could suggest the lids are oven-safe too. That is unknown from the records, so consider showing the oven use without lids or keeping the claim tied to the containers.; All four lines are legible at phone size. The appliance icons only repeat the headline.

Variation check: The four identical small containers match the reference: cream body, grey lid window and 'Caraway' lid label.; 'Same glass. Every stop.' paired with the oven stop again shows lidded containers. caraway-fact-1 covers the containers being oven-safe, and lid oven safety is unknown. Consider showing the oven-stop container without its lid.; The transit-line metaphor is clear, and all text is legible at phone size.

Primary source

X has a real idea: a clipped stack of printed screenshots next to a laptop recording screen and camera. 'Skip the screenshot stack' frames the problem the product solves. Y is a clean feature card with a flat illustration and a trim/stitch timeline. It lists more features (trimming and stitching clips) but has no concept, and the vector style looks like a template. Both name the platforms clearly. X is more memorable and builds the product into the comparison.

Control check: The flat-vector laptop, phone and clip timeline read as an illustrative workflow, not a real Loom screenshot. That's fine.; The scissors and dashed cut line between two clips stand for trim/stitch (loom-fact-3) without inventing features.; All lines are legible. The smallest line (Chrome extension, desktop or mobile app) is still readable at phone size.

Variation check: The laptop recording UI (camera bubble plus pause/record/stop controls) is an illustrative workflow, not a real Loom screenshot. The photoreal face in the bubble is a generated person, not a user or testimonial.; 'Skip the screenshot stack' is creative framing based on a hypothesized pain point. It makes no quantified productivity claim, which is correct.; Showing the 'screenshot stack' as a clipped pile of printed paper is a bit literal but readable. All lines are legible.

Primary source

Both show the tree and its roots in cross-section. X puts the headline over busy foliage, and the glowing blue water lines look generic and fantasy-like. Y actually shows the service: a probe going into the soil and spreading water and air through the roots. That makes 'root aeration with water injection' easy to see. Y's headline sits on clear sky, and its backyard setting feels local and residential. Y is more specific and better made.

Control check: The cutaway of roots with glowing blue water lines is clearly a conceptual illustration, not a record of TreeNewal's work.; 'Tree care that starts at the roots' is a creative framing. The records support root aeration but not a claim that all of their tree care starts at the roots. The risk is low because it doesn't promise any tree-health outcome.; The lush, healthy canopy is just illustration, but it should not be read as a promised result.

Variation check: The hose-fed probe injecting water near the roots is a conceptual illustration of 'root aeration with water injection' (treenewal-fact-2). It doesn't show TreeNewal's actual equipment or a real job.; There is no tree-health outcome claim. The brand name and service lines sit on dark soil and are legible.

Primary source

X is a direct-response layout: a question headline, two service icons, and a moldy-to-clean wall split. It is very clear but generic, and the blue banner feels heavy. Y's 'A bucket won't clean this up' shows a ceiling leak dripping into a bucket in a nice living room. It's a relatable moment that makes the case for calling a professional. The service list in the footer stays readable. Y has the stronger idea and polish with no loss of clarity.

Control check: The wall and floor change from moldy and water-stained on the left to spotless on the right. That reads as an implied before/after restoration result. It's generated and not labeled as client work, but it pushes toward outcome imagery. Consider keeping only the damage state or making the transition clearly symbolic.; The text uses '•' between services where the intended line has '·'. That's trivial.; Locations are independently owned franchises, so availability varies. The ad makes no locality or response-time claims, which is correct.

Variation check: The ceiling leak dripping into a bucket on a wet floor is a conceptual damage scene. It only shows the problem, with no implied before/after, which is good.; 'Water damage cleanup and repair' combines paul-davis-fact-1 and paul-davis-fact-2. 'Homes and businesses' is a fair paraphrase of 'residential and commercial properties'.; No response-time or locality promises, which is correct since service varies by franchise location. All text is legible on the cream band.

Primary source

X explains the product well: the subjects, 'Short videos. Then you code,' and laptop and phone mockups. But the illustration is generic. Y puts the app on a phone during a train commute, next to a crossword the rider has put aside. It suggests that learning can fill the time a puzzle usually does, and it stresses the mobile app. The subject icons (Python, R, statistics) and the brand are clear at the bottom. The phone screen is a bit abstract, but Y has the better idea and photography.

Control check: The laptop and phone UIs (video player beside an input field) are a generic illustrative workflow, not a real DataCamp screenshot.; 'Short videos. Then you code.' is supported by datacamp-fact-1 (coding challenges) and datacamp-fact-2 (videos broken up by exercises).; The bottom line and phone icon are legible.

Variation check: The Python icon looks like the Python Software Foundation's trademarked two-snake logo, which is third-party branding. A neutral glyph would avoid implying endorsement.; The phone UI (video plus colored code-like bars) is illustrative, not a real app screenshot. The instructor in the video is generated, not a real DataCamp instructor.; The crossword prop casts the app as a commute pastime. That's a hypothesis framing, not a claim. Text is legible.

Primary source

Both say the same thing: video calls with registered dietitians. X uses a stock-style illustration with a salad bowl. Its headline, 'Nutrition help, made personal,' is generic, and the two-line subhead has an odd capital 'From'. Y's 'Dietitian time, from your kitchen' is more specific, and the photo shows exactly that: a phone propped on a kitchen counter. The blurred dietitian on Y's screen is soft, but the warm photography and clear type order (headline first, brand at the bottom) make Y better made. Y wins on specificity.

Control check: The video-call screen with an illustrated dietitian is conceptual. Records say 'telehealth' but don't specify video calls. Showing a video call is a reasonable reading but not confirmed.; The salad bowl is styling only. There is no health or weight claim, which is correct.; The third line, 'From a national network', is a slightly dangling fragment on its own, but it's accurate.

Variation check: The blurred clinician on a video call is conceptual. Records say telehealth but not specifically video calls, so 'from your kitchen' is a reasonable reading of remote telehealth, not a sourced feature.; No health-outcome claims. The fridge checklist has only squiggles, with no invented text.; The lower text block is legible against the warm wood background.

Primary source

X makes the claim directly with 'All you need is an iPhone.' Its three filter pills (trainer, music, meditation) match the app screen on the phone, so the feature is shown and not just stated. Y's 'Pick your workout like a record' has nice mood and an original metaphor. But the sleeves are blank colored squares, the phone is small, and the record idea only fits the music filter, not trainer or meditation. Y is the more distinctive idea but less clear. X ties the product to the message better.

Control check: The required qualifier ('Subscription and compatible hardware and software required.') is tiny fine print at the bottom edge. It would be about 6px tall on a phone, so it's effectively unreadable. That matters because the headline 'All you need is an iPhone.' depends on it. Enlarge it or raise its contrast and position.; The phone UI (yoga thumbnail plus trainer, music and meditation filter rows) is illustrative, not a genuine Fitness+ screenshot. The rows have icons but no labels, so the filter idea depends on the copy.; The yoga mat as a prop slightly undercuts 'All you need is an iPhone', but it isn't a claim. There is no Apple logo, which is correct.

Variation check: The required qualifier ('Requires a subscription and compatible hardware and software.') is small grey text on a dark background, about 7px on a phone. It's hard to read even though it qualifies 'All you need is an iPhone.' Increase its size and contrast.; Nothing in the image shows a workout. The record-crate metaphor and the colored tiles on the phone rely entirely on the copy to explain the offering. The text is clear enough, but the visual link is weak.; The phone's colored tiles are an abstract UI, not a genuine Fitness+ screen. There is no Apple logo, which is correct.

Primary source

Both are warm, believable lifestyle shots of the blocks being played with. X only says 'Solid wood blocks', and the brand and product name sit small at the bottom. Y names the product up top and adds who it's for plus 'Stage-based activities included', which tells you more about what you're buying. Y's stacked tower is a little implausible (a pentagon balanced on a triangle point), but it doesn't hurt how the ad reads.

Control check: The blocks mostly match the reference set: the orange and teal arches, yellow pentagon, orange trapezoid, teal triangle, green disc and teal cylinder all appear in it. Small drift: the solid blue rectangular block and the thick purple bar don't clearly match any piece in the reference. The magenta triangle is larger and a different shape than the set's small pink right triangle. This is minor and doesn't misrepresent the product.; All visible text comes from the supplied lines and is backed by lovevery-fact-1, 2 and 3. Nothing makes a developmental claim.; Both sub-lines are legible at phone size.

Variation check: Product drift: the centerpiece is an unpainted natural-wood arch on a natural-wood cube. The real Block Set's arches are painted orange and teal, and the reference set has no unpainted pieces.; The green cylinder lying on its side, the yellow rectangular block and the magenta cube don't match the reference pieces. The real set has a green flat disc and painted cubes in a set palette.; Because the ad names 'The Block Set', the arrangement should use pieces that are actually in the set.; The text is accurate to lovevery-fact-1 and legible.

Primary source

These are very close. Both show input turning into a diagram. X has the punchier headline ('Prompt in. Diagram out.'), uses sticky notes that feel like Miro, and adds the agent-on-canvas point. Y is cleaner, and the dot-grid canvas is a nice touch, but its headline reads more like a formula and it says a bit less. X wins slightly on specificity.

Control check: The sticky notes, document, arrow and diagram are a clear illustration of the workflow, not a fake screenshot. That's acceptable.; 'Any agent can work on the canvas' (miro-fact-3) isn't shown in the image. The cursor reads as a human user. This is a missed chance, not an error.

Variation check: The text-block-to-flowchart graphic is a clearly illustrative workflow, not a screenshot.; Only a document-style input is shown, though the headline lists prompt, doc or data. That's acceptable.; The claims are backed by miro-fact-1 and 2. All text is legible.

Primary source

X's magnifying glass on bark is a simple, punchy picture of 'diagnose', and the headline is tight. Y is more specific: it names stress, disease and pests, and adds shareable assessments and documentation. Its yellowing leaves show what 'stressed' looks like, and the clipboard suggests a professional assessment. Y gives a homeowner more reason to call.

Control check: The gloved hand and blank clipboard are a conceptual service scene. Nothing presents it as real client work, and the blank clipboard avoids a made-up report.; The claims are backed by treecarela-fact-2 and 3. No licensing or certification is implied.; The two body lines are fairly small over the light sky but still legible at phone size.

Variation check: The magnifying glass on bark is a conceptual diagnostic image, not real client work. That's acceptable.; The tree and background look healthy, so the image doesn't show 'stressed'. The headline carries that idea alone.; The white text over a busy green bokeh background is mostly legible. The smaller sub-line has the least contrast.

Primary source

X shows the finished result: framing on one side, a polished kitchen on the other. It also spells out the four service steps and lists the project types (kitchens, baths, ADUs), so the offer is concrete. The line-art drawn over the photo is a bit cluttered. Y is more elegant, with one continuous line tying the stages together as a 'one team' metaphor, but it never says what they build and shows no outcome.

Control check: Bare framing on the left next to a finished kitchen on the right reads like a before/after. Viewers could take it as a real Neil Kelly project result. The two halves are clearly different rooms and no caption calls it real work, so it stays conceptual, but the split invites that reading.; In the four-step row, the '·' separators are replaced by icons and lines. The words still match.; The service claims are backed by neil-kelly-fact-1, 2 and 3. The ad names no location, which avoids the Eugene office not taking new projects.

Variation check: The offering is unclear. 'One team' plus a process chain and a timber-framed house icon never says home remodeling, kitchens, baths, additions or ADUs. A viewer could read Neil Kelly as a new-home builder or a general contractor. Add a short line naming the project types from neil-kelly-fact-1 or 3.; 'Permits' and 'Build' are fair short forms of 'permitting' and 'construction' (neil-kelly-fact-2).; The line art is clean and legible at phone size.

Primary source

X is just the logo and four category cards, with no benefit or hook. It's clean but generic. Y gives a reason to care ('Learn at the level you need'), backs it with blocks rising in steps, and adds the Coursera Plus subscription detail. Y's small text line will be tight on a phone and the laptop screen is blurred, but it is clearly the stronger concept.

Control check: The headline 'Learn at the level you need' is a creative reading. No record mentions difficulty levels, and the rising blocks reinforce a progression idea. It only holds up as a nod to the four program types in coursera-fact-1. Keep it as a soft hook and don't extend it to skill-level or career claims.; 'Coursera Plus: one annual subscription' doesn't say what the subscription gives you. It's accurate but not very informative.; The program-type line is the smallest text. It's tight but readable at phone size.; The blurred laptop screen doesn't show any invented interface. That's acceptable.

Variation check: The card labels match the supplied lines and coursera-fact-1, and they're spelled correctly.; The tilted card text, especially 'Online courses' and 'Degrees', is slightly harder to read at phone size but still legible.; There's no headline or benefit line. The ad is a brand name plus program types, which is clear but thin.

Primary source

X shows the actual service: a video call with a dietitian on the laptop and a chat on the phone. Its three icon bullets cover the location, chatting between sessions and how often you meet. Y's headline ('Your dietitian. Your cadence.') is neat, but the laptop screen is blank, so the service never shows up. That makes it feel like a stock desk photo. X's last bullet runs close to the right edge but stays readable.

Control check: The blurred person on the video call is a generic stand-in, not a named or credentialed dietitian. That's acceptable as an illustrative scene.; 'anytime' was dropped from the chat line, which is a conservative paraphrase of top-nutrition-coaching-fact-2.; The fruit bowl doesn't suggest weight loss or a health outcome. The ad makes no outcome or insurance claims.

Variation check: 'Your cadence' with Weekly / Biweekly / Monthly never says these are session frequencies. The blank laptop is the only hint that the service is virtual, so the offering is understandable but only weakly signaled.; The pills look like interactive UI buttons, but they're only shown in the ad. That's acceptable.; The ad makes no outcome, insurance or credential claims beyond 'registered dietitian' (top-nutrition-coaching-fact-1).

Primary source

Y gets the multi-device idea across instantly: the same workout appears on a phone, tablet, laptop and TV, scaled up across the row. The visual is the claim, and it's crisp at phone size. X has more facts (offline downloads, workout types), but it's busy. The foreground woman is blurry and the TV content is smeared, so the scene is less clear even though it looks real.

Control check: 'Download workouts for offline use' sits right under the device list, so it could suggest downloads on computer or TV too. The record (les-mills-fact-2) only supports downloads to Android or iOS devices. Consider tightening the line or adding 'on mobile'.; The workout footage on the TV and phone is illustrative, not real class footage. It's fine as long as it's not presented as actual content.; The category line is small. It's legible at phone size but on the edge.; The ad shows no body-transformation or results message.

Variation check: The flat cartoon figure on each device is illustrative app content, not real LES MILLS+ footage or a screenshot. That's acceptable.; Neither the word 'workouts' nor the subscription nature appears in the text. The workout figure is the only thing telling viewers it's a fitness app.; The ad has no results or body-image messaging. The text is large and legible.

Primary source

The two images look identical. The product tube and blush/bronzer swatches show the texture and both shades, and the visual matches the copy closely. Both have the same capitalisation error: 'In blush' should start with a lowercase 'i'. The body copy is a bit small on a phone.

Control check: The tube has no printed text. The real product has 'cloud paint / seamless cheek color / blush aérien' on the body and the shade name 'BEAM' on the coral band. Leaving the label off isn't invented text, but the tube can only be identified from the headline.; Otherwise the tube matches the reference: white crimped tube, coral bottom band, coral hex cap. The pink and bronze swatches fit 'blush and bronzer shades' (glossier-fact-2), and 'blends' is supported by 'blendable' (glossier-fact-1).; The lowercase line 'In blush and bronzer shades' starts with a capital 'In', which matches the intended line exactly. It reads like a sentence fragment, but this is a style point, not an error.

Variation check: This looks the same as glossier-control. The tube has no label, though the real product has 'cloud paint / seamless cheek color / blush aérien' and the shade 'BEAM'. Its shape and colours match the reference.; The copy is supported by glossier-fact-1 and glossier-fact-2.

Primary source

Same copy and style. Y's two-way arrows between the action and the table directly show 'Zaps can read and update Tables', so the picture matches the claim. X's loop from the table back to the trigger suggests a cycle instead of read/write, which is less clear. Both are clean, on-brand and readable on a phone. Y leaves a bit more empty space at the bottom.

Control check: The workflow diagram is clearly an illustration, not a screenshot, which is fine.; The flow only shows one direction (gear to table). 'Read and update' would be shown better with a two-way link between the Zap step and the table, as in the variant.; 'Zapier' is plain orange text, not the official logo. That's acceptable.; An arrow loops from the Tables icon back to the trigger icon, suggesting Tables can start a Zap. The records (zapier-fact-3) only support Zaps reading and updating Tables. Nothing supports Tables acting as a trigger.

Variation check: The two-way arrows between the gear (Zap step) and the table fit 'read and update Tables' (zapier-fact-3). It's an illustration, not a real screenshot.; No problems found.

Primary source

The two images look identical. The patch of ground-down stump mulch flush with the lawn directly shows the 'ground below ground level' claim, which is a smart, specific visual. The headline type is strong, and the photo is warm and believable.

Control check: The lawn with a flush patch of wood chips is a conceptual picture of a stump ground below ground level (heartwood-tree-service-fact-2). Nothing labels it as a real client job, which is fine.; The lake in the background could suggest Lake Norman, but no place is claimed in the text, so this is not a problem.; The text is dark on a bright sky but still legible at phone size.

Variation check: This looks the same as the Heartwood control. The flush patch of wood chips is a conceptual picture of a ground-down stump and isn't presented as client work.; No problems found.

Primary source

The two images look identical. Colour-coded callouts map windows, siding, roofing and doors onto one house, so the service is clearly part of the idea. The type is clear at phone size. The energy-efficiency line is small and not shown visually.

Control check: The house is a generated concept image with coloured outlines and labels for windows, siding, roofing and doors. It isn't presented as a finished Power project, which is fine.; 'Built on energy efficiency' rewords 'focused on energy efficiency' (power-home-remodeling-fact-1) without adding numbers or savings claims. That's acceptable.; The third line is fairly small at phone size but still readable.

Variation check: This looks the same as the Power control. It's a concept house with labelled callouts, not a real project, and makes no numbers or savings claims.; The third line is small but readable.

Primary source

The two images look identical. Both clearly say Udemy offers certification prep with courses and practice tests. The laptop mockup with a video and quiz screen ties into the offer. The desk scene is clean but generic, and the keyboard keys are garbled when you look closely.

Control check: The laptop screen shows a generic video-lesson layout with no Udemy content or features, so it reads as an illustration, which is fine.; Some keyboard key labels are garbled up close, but they can't be seen at phone size.; There is no exam pass promise or voucher offer, which is correct.

Variation check: This looks the same as the Udemy control. The screen is a generic lesson layout, and the garbled key labels can't be seen at phone size.; No problems found.

Primary source

The two images look identical. The video call with a dietitian on the laptop and the three format icons explain the service well. Both have a glitched 's' in 'self-guided', the route icon is vague, and the healthy-kitchen scene is generic.

Control check: Visible text error: the 's' in 'self-guided' is broken, with what looks like a strike or bar through it. The rendered text doesn't match the intended line.; The woman in a white coat on the video call is generated and could be read as a real OnPoint dietitian. She isn't named, so this is acceptable as an illustration, but it is a light risk.; The healthy food bowl makes no weight-loss or health claim, which is fine.; 'Online coaching with a registered dietitian' sits directly above 'One-on-one, group or self-guided', so it reads as if every format, self-guided included, comes with a dietitian. The records support dietitian support (onpoint-nutrition-fact-2) and the three formats (onpoint-nutrition-fact-1) separately, not together.

Variation check: The same text error appears as in the control: the 's' in 'self-guided' is broken, with what looks like a strike or bar through it.; The generated person in a white coat on the call could be read as a real OnPoint dietitian. She isn't named, so the risk is light.; 'Online coaching with a registered dietitian' above 'One-on-one, group or self-guided' suggests a dietitian is part of self-guided coaching too. The records don't connect the two.

Primary source

The copy and layout are the same: one app working across a bike, treadmill and rower. Y makes the screens bigger and brighter, so the app classes are easier to read on a phone and the product is more central to the idea. In X, the screens are small and the rower's screen and seat placement look awkward. Both have some AI-generated equipment oddities, but they're less noticeable in Y.

Control check: The rower looks like it has no handle and no foot rests, just a seat on a rail next to the flywheel. That's a visible mistake on a machine the copy names.; All three machines are dark, premium models with mounted screens, and the bike has a red resistance knob like Peloton's. They read as Peloton's own Bike, Tread and Row, which confuses the app offer with hardware and works against 'any bike, treadmill or rower' (peloton-app-fact-2).; The screens show silhouetted figures, with no class counts or instructor claims, which is fine.

Variation check: The machines are drawn correctly: this rower has a handle and foot rests.; As in the control, the dark bike, treadmill and rower with screens look like Peloton's own hardware. That blurs the app-only offer and the 'any bike, treadmill or rower' message (peloton-app-fact-2). Using clearly generic, mixed-brand equipment or a phone or TV would fit the copy better.

Primary source

Same composition, copy and art direction. Only the headphone colour changes: tan in X, off-white in Y. Both stand out well against the orange. In both, the folded headphones sitting in a backpack show 'folds flat, goes anywhere' directly, so the product is the idea.

Control check: The headphones are a close match for the beige reference (image 15). The shape, hinge, folded-in earcup and three-button layout (seen on the white reference) are consistent with the real product.; The skull mark on the headband is a darker grey than the product's tonal emboss in the reference. It's minor, but it stands out a bit more than the real one.; 'Goes anywhere' is broad, but it comes straight from the source quote 'Flat-folding, collapsible design goes anywhere' (skullcandy-fact-2). The 'Folds flat' claim is consistent with the collapsed headphones shown in the bag.; All lines are legible at phone size. There is no extra text, price or battery claim.

Variation check: The white headphones match the white reference (image 16) well. The headband hinge, three earcup buttons and tonal skull mark are all accurate, and the headphones are shown folded into the bag.; 'Goes anywhere' comes from the source quote (skullcandy-fact-2).; All four lines match the intended words and are legible. There is no extra text.

Primary source

These are nearly the same image. The only differences are small icon details: X's document icon has lines and Y's AI sparkle is pink. Both show data flowing into Airtable's grid and then into automated steps, which makes the product part of the idea. Neither is noticeably better at phone size.

Control check: The flow diagram and the laptop/phone mockup are generic, so they read as an illustrative workflow rather than a real Airtable screenshot. That is acceptable.; The green grid icon looks a little like the Google Sheets icon, which might suggest a Sheets integration. The records only name Slack, Google Drive and Salesforce. It's a low risk.; The brand name is set as plain text, not a copy of the official logo. The text is legible at phone size.

Variation check: The diagram and device mockups are generic and illustrative. They are not shown as a real Airtable screenshot.; The purple document icon with lines and the green grid look a little like Google Docs and Sheets icons. That's a small risk of implying integrations beyond the three named, but no third-party logos are shown.; The copy matches airtable-fact-1, -2 and -3 exactly.

Primary source

Same copy and lighting. X includes patio furniture and a seating area, so 'Reclaim your yard' reads as a yard you can actually use. Y is just lawn and trees, which links less clearly to the outcome. Neither ad shows tree work actually being done.

Control check: The neat yard with its fresh mulch circle is a generated, conceptual scene. Next to 'Reclaim your yard' and 'Stump grinding', the mulch patch could be read as a finished stump-grinding job. It is not presented as client work, but don't caption it as such.; The brand name sits in a dark green pill that looks like a CTA button but isn't one. That's a minor point of confusion.; There are no locations, certifications, guarantees or reviews. The headline 'Reclaim your yard' is supported by arbor-masters-fact-3.

Variation check: This is a conceptual generated yard scene, now with patio furniture. The mulch bed could still be read as a stump-grinding result. It is not presented as client work.; The brand name is in a pill shape that looks like a button. That's minor.; There are no locations, credentials, guarantees or reviews. Every line is supported.

Primary source

Same copy and the same cross-section showing bath, kitchen, addition and whole home, which lines up neatly with the headline. X adds a soft blue wash on the trees. Y adds clouds and slightly crisper linework. The differences are minor and don't change how well either ad works.

Control check: The line-drawn cross-section of bath, kitchen, dining/addition and two-storey home is clearly conceptual. It isn't shown as a finished project.; 'Neighborhood design studios across DC, VA & MD' combines case-architects-remodelers-fact-2 and fact-3, and it matches the source wording. The studio count is correctly left out.; There are no award, guarantee or year claims. The text is legible at phone size.

Variation check: The architectural line drawing, with clouds added, is clearly conceptual and not shown as a completed Case project.; The copy matches case-architects-remodelers-fact-1, -2 and -3. There are no award or guarantee claims, and the text is legible.

Primary source

The two layouts and all the copy are the same. In X, the over-ear headphones fill the middle panel, so 'Hear it' is obvious at phone size. In Y, the earbuds are tiny and the mug takes over, so the audio idea is weaker. X's watch/hear/read panels line up more clearly.

Control check: The earbuds in the 'Hear it' panel look like Apple AirPods Pro: white stem, case-style tips and the same shape. That is identifiable third-party trade dress and could imply an Apple tie-in. Replace them with generic headphones, as the critic edit did.; The video player, text tablet and code editor are generic and read as illustrations. That is acceptable.; All copy matches linkedin-learning-fact-1, -2 and -3. The brand name is plain text, not the official 'in' logo.

Variation check: The AirPods-like earbuds have been replaced with generic navy over-ear headphones, which removes the third-party trade dress problem.; The video, text and code-editor visuals are generic illustrations. The copy matches all three records and is legible at phone size.

Primary source

Same copy, call to action and set of props: a healthy bowl, a laptop for virtual sessions and a notebook. Y has warmer wood tones and a slightly bigger bowl. X has a cooler stone surface with the tea placed near the top. Both are well made and equally clear about the counselling offer.

Control check: The food bowl and desk setup are a conceptual lifestyle scene. There is no weight-loss, body or clinical promise, and 'your goal' stays neutral.; The CTA and 'Virtual, or in person at select locations' match jm-nutrition-fact-1 and -3. The Canadian spelling 'counselling' matches the source.; The laptop keyboard has garbled key glyphs, but they can't be read at phone size. It's a minor issue.

Variation check: This is a conceptual desk-and-meal scene with no weight-loss, medical or outcome claims.; The keyboard glyphs are garbled, but they can't be read at phone size.; The copy and CTA match jm-nutrition-fact-1, -2 and -3.

Primary source

Same layout, copy and moody gym scene. X shows a man squatting on the phone screen and includes a shaker bottle. Y shows a woman doing a lunge. Both show the in-app video demo the copy describes, and the moody lighting looks equally good in each.

Control check: The phone shows a squat silhouette as an illustrative app screen. It isn't a real Future UI and has no video controls, which is acceptable as a concept.; The '1' in 'Custom 1-on-1' uses Space Grotesk's long-flag glyph. At small sizes it can look like '/1' or 'A'. Check that it reads clearly.; The silhouette is lean and athletic, but there are no before/after, weight-loss or quantified performance claims. Copy matches future-fact-1, -2 and -3.

Variation check: The lunge silhouette on the phone is an illustrative app screen, not a real Future UI. There are no results, body-transformation or quantified claims.; The '1' glyph in 'Custom 1-on-1' (Space Grotesk) has a long flag that may read oddly at small sizes. Check that it's legible.; The copy matches future-fact-1, -2 and -3.

Primary source