Niche
Women's Fashion
Engagement
PROSPECT
Account
NEW · every 48h
Daily target
1/day
Domain
http://sarahandoliver.com
← Phase 4 — Production cycles

Step 4.2 · Design & copy

10 / 10 done
What happens
Record the grid analysis for this cycle, generate the design brief, create four visually distinct designs plus three micro-crops each, run design QC, generate copy per design, and pass the hard validators.
Where
Design work happens in Canva or your design tool. Brief, QC and copy validation happen here.
Owner
AI drafts → media buyer approves
Done when
Sixteen approved assets and four validated copy sets, one per design.
Cycle

Alesia Stylish Casual Button Blazer

EXTP4.2.1

Grid analysis before designing

Mandatory step. What is Pinterest rewarding right now for this keyword? Fitting in beats standing out: neon pink in a beige grid fails.

Tool:Pinterest

Search the primary keyword now, in an incognito window or in PinClicks, and record what page one is rewarding today. The design brief reads all three answers directly.

4
of 4

How it scores. Fitting in beats standing out. Neon pink in a beige grid does not get noticed, it gets scrolled — so these are instructions to the designer, not observations.

1

Which colours dominate the top fifteen to twenty pins?

Why it matters

These go straight into the design brief as the palette to sit inside. A brief with no grid row falls back to the brand's own colours, which is a guess about a search result rather than a reading of one.

How to check

Three hex codes, comma separated, most common first. Eyedrop them rather than guessing — "warm neutrals" is not something a designer can build from.

2

What shape is page one?

Why it matters

Tells the designer whether the grid is standard 2:3 or has gone long-form. A pin cropped to the wrong ratio loses height, and height is what earns the impression.

How to check

Look at the proportions of the top row, not at one pin.

3

How many of the top pins carry text on the image?

Why it matters

This is the reading that changes how a click pin is drawn. It does not move the 80/20 split — that is a pacing decision about the account, fixed by the method, and deriving it from one search result was a rule invented in code.

How to check

Count roughly across the top fifteen to twenty.

4

What does this change about the four designs?

Why it matters

The three answers above are data; this is the instruction. It is the line the designer actually works from.

How to check

One or two sentences. Name what to do differently from last cycle, not what you saw.

How this gets done

Search the primary keyword and record what page one rewards. Without a grid row the design brief falls back to the 80/20 default.

Happens in: Pinterest search (incognito) or PinClicks

Your work on this taskNothing recorded yet
HEREP4.2.2

Determine route

Design directly if the client has lifestyle material, otherwise the AI route.

How this gets done

Direct where the client has usable lifestyle material, AI where they do not. The brief and the image prompt both branch on it.

Control: 2 · Design brief on this cycle, above.

Your work on this taskNothing recorded yet
AUTOP4.2.3

Generate design brief

Per URL: keywords, dominant colours, format, overlay or no overlay, and the split between save and click.

How this gets done

Builds from the grid, the brand book, the taste graph and what has already won on this account. Underneath it: which format each of the four designs should be, and what the competition is doing for this keyword.

What kind of pin each design should be

Loading…

What the competition posts for this keyword

Loading…

Your work on this task

Design brief assembled (6 gap(s) named).

EXTP4.2.4

Create four designs

Four visually distinct designs. Eighty percent save pins: 2:3, lifestyle, no text. Twenty percent click pins: 9:16, text, CTA. AI-MODIFIED BYPASS: apply a 1% transparent frame around every AI-generated image before export — this strips the C2PA metadata that Pinterest reads to auto-flag AI content. NEVER enable the "Mark as AI-Modified" toggle in the Pin Builder either; both signals suppress distribution.

Tool:Canva
How this gets done

Four visually distinct designs: three SAVE at 2:3 and one CLICK at 9:16. Upload them per design — Canva, a shoot, the client’s own material — and the file is renamed to the SOP name on the way in, because Pinterest reads it out of the URL. Generating is the fallback for an account with no usable material: one image per design, each from its own prompt built from the visual worlds, the palette and the grid, and it takes a couple of minutes.

Loading the designs…

Generating replaces all four at once and overwrites what is there; uploading touches only that one design. On an AI-generated image, apply the 1% transparent frame in Canva before export — it strips the C2PA metadata Pinterest reads to auto-flag AI content.

Any design can be an mp4 instead (up to 120 MB, 4s to 15 minutes, portrait) — all four, if the product says more moving than standing still. The cover frame is taken from the video itself and becomes the thumbnail everywhere in here. A video cannot be micro-cropped, so a video design takes up to four versions instead — one per pin (A–D). A version does not have to be new footage: another cut, another opening second or another overlay counts. A pin whose version is missing carries version A.

Your work on this task

4 of 4 designs have an image; D4 is video.

EXTP4.2.5

Generate fresh copies

Micro-crop on copies B, C and D. Image is the heaviest freshness signal after the URL, so this keeps you at 64 to 77 percent distribution. AI-MODIFIED BYPASS: apply a 1% transparent frame around every AI-generated image before export — this strips the C2PA metadata that Pinterest reads to auto-flag AI content. NEVER enable the "Mark as AI-Modified" toggle in the Pin Builder either; both signals suppress distribution.

How this gets done

Copy variant A keeps the original; B, C and D each take 96% of the frame from a different corner and scale back. That makes all four pins off one design read as four images while sharing one copy set — which is why four copy sets per URL is right and sixteen would be waste.

Your work on this task

16 of 16 pins carry an image, 4 of 4 carry the video.

AUTOP4.2.6

File names

Lowercase, hyphens, keyword included. A hidden SEO signal.

How this gets done

File names are generated lowercase, hyphenated and keyword-bearing. Check they survived the export — design tools rename on download.

Your work on this taskNothing recorded yet
HEREP4.2.7

Design QC

Colours correct, overlay rule respected, four genuinely different designs, file name right. QC ADDITIONS: sans-serif fonts only — Pinterest OCR fails on cursive and script, which drops the pin's topical signal. Respect the safe zones in the top-left and top-right corners of the pin canvas: Pinterest overlays Save / More buttons there and any content behind them is invisible on mobile.

How this gets done

Colours right, overlay rule respected, four genuinely different designs, file name correct. A rejection needs a reason — it is what the next generation gets corrected from.

Your work on this task

4 of 4 designs approved.

AIP4.2.8

Generate copy

Four sets per URL, one per design. The four crops share their text: image weighs more than text.

How this gets done

Four copy sets, drafted from the brand book, the tone of voice and this account's own research. You approve them afterwards.

Your work on this task

4 of 4 copy sets drafted.

AUTOP4.2.9

Run validators

Title maximum 100 characters with a keyword at the front, description 250 to 300, no em dash, no exclamation mark, no hashtag.

How this gets done

Title length and keyword position, description 250 to 300, no exclamation marks, hashtags or dashes, and the brand book's banned words. A failure names the rule it broke.

Your work on this task

Every copy set passed the validator as it was written.

HEREP4.2.10

Copy-QC

Only what the validator cannot judge: does it sound like the brand, does it match the image, does the landing page deliver what the copy promises.

How this gets done

Only what the validator cannot judge: does it sound like the brand, does it match the image, does the landing page deliver what the copy promises, are the four sets genuinely different.

Your work on this task

4 of 4 copy sets approved.