Loading tutorials…
Loading tutorials…
A social manager who spends 8 hours a week sourcing stock photos can shift to 2 hours of Midjourney generation that produces on-brand, unique imagery instead. Here is the workflow.
Who this is forSocial media managers, content creators, and solo founders producing 5-15 posts per week per channel. If you are spending more than 4 hours a week on visual sourcing, this saves you most of that.
What you'll need
Step 1
Most social accounts have 3-5 content pillars (e.g., educational, behind-the-scenes, product, lifestyle). Each pillar gets a defined visual mode.
List your content pillars from your social strategy. If you do not have explicit pillars, list 4-5 categories your last 30 posts naturally fell into.
For each pillar, define the visual mode: subject type, mood, color palette. Document the SREF code/URL that produces that mode.
Example for a SaaS social account: Educational (clean illustrated diagrams, brand colors) | Behind-the-scenes (warm candid photography) | Product (clean studio product shots) | Lifestyle (target customer in their environment).
This is your "social visual playbook." Refer to it every weekly production session.
Step 2
Different platforms need different aspect ratios. Build a template per platform/format so you never forget to generate the right size.
Instagram feed: --ar 1:1 (square) or --ar 4:5 (portrait, recommended for more screen real estate)
Instagram stories / Reels covers / TikTok: --ar 9:16
LinkedIn feed: --ar 1.91:1 (landscape) for link posts; --ar 1:1 for image posts
Twitter/X: --ar 16:9 for landscape; --ar 1:1 for square
YouTube thumbnails: --ar 16:9
Pinterest: --ar 2:3 (tall portrait)
Save these as quick-paste snippets in your notes app. Wrong aspect ratio = re-generate = wasted credits.
Step 3
Block 2 hours weekly. Generate all the visuals for next week's content in one focused session. Avoid daily one-off Midjourney trips.
Open your content calendar for the next 7 days. Tag each post with its pillar.
For each post, write the prompt fragment: subject + setting + pillar-mode SREF + aspect ratio.
Run all prompts in sequence. Each batch of 4 takes ~50 seconds; 10 posts = ~10 minutes of pure generation.
Cull as a batch at the end — 1 finalist per post (with 1 alternate kept just in case).
Upscale all finalists. Drop into your social scheduler (Buffer, Later, Hootsuite) with captions.
Total session time: 90-120 minutes for a full week of content.
Step 4
Quote cards and educational carousels need text overlays. Generate the background in Midjourney; add text in Canva or Figma.
Midjourney is terrible at rendering text. Stop fighting it.
Generate textless backgrounds with --sref locked to your brand. Mood: depends on the quote/topic.
In Canva: set up a master quote-card template with your brand fonts, color, logo placement.
Workflow: Midjourney background → drop into Canva template → swap quote/text → export.
Per post: 30 seconds in Midjourney + 60 seconds in Canva. Faster than searching Unsplash.
For multi-slide carousels: generate the cover in Midjourney, build the slide-content slides in Canva using your brand template.
Step 5
Short-form video covers (first frame thumbnails) get massive scroll-stop attention. Treat them as a dedicated visual category, not an afterthought.
Aspect ratio: --ar 9:16. Crop-safe area is the middle 60% — assume Instagram/TikTok will crop the top and bottom.
Prompt the subject occupying the middle band of the frame, not the edges.
Use --stylize 50-100 for short-form covers — too artistic and the cover looks like art instead of a stop-scroll punch.
Generate 3-5 cover options per reel. Pick the one with strongest face/gesture/contrast.
Add text overlay in your reel editing app (CapCut, InShot) or use a Canva cover template.
Step 6
Some post types repeat weekly (e.g., "Monday motivation," "Friday wrap-up"). Save the master prompts so you never rewrite.
Document a library: Post type | Prompt template | SREF | --ar | Notes.
Example: "Monday motivation = --sref [brand code] --sw 150 a sunrise over a calm lake with someone walking toward it, warm golden light, optimistic mood --ar 4:5 --v 7 --stylize 100"
Each week, copy the template and tweak one element (different setting, different time of day) to keep the post fresh while staying on-brand.
This is how 5+ posts/week becomes a 2-hour task instead of a 6-hour task.
Step 7
Tag posts by visual mode in your analytics. After 30 days, see which modes drive the most saves, shares, and reach.
In your scheduler or analytics tool, add a tag like "visual-mode=clean-illustrated" or "visual-mode=warm-candid."
After 30 days, compare engagement rate by visual mode.
Double down on the top 1-2 modes. Reduce or retire the bottom mode.
This turns Midjourney from "free art" into a creative testing system that compounds over time.
Common mistakes
Generating posts daily instead of in a weekly batch
What goes wrong: Daily Midjourney sessions consume 20-30 minutes each time for context-switching. Total weekly time hits 4-5 hours instead of 2. The "AI saves time" promise dies.
How to avoid: Block 2 hours one day a week. Generate the full week of visuals in one focused session.
Cropping instead of generating in target aspect ratio
What goes wrong: Generated 1:1 image cropped to 9:16 loses 60% of the visual. Quality drops, subject is awkwardly framed, post performs worse.
How to avoid: Always generate at target --ar. Each platform format = its own prompt run.
Asking Midjourney to render captions and quote text
What goes wrong: Quote cards come out with gibberish text. You spend 20 minutes per post fighting the model. Output is unusable.
How to avoid: Generate textless backgrounds. Add text in Canva using a master template. 60 seconds vs 20 minutes.
No brand SREF — every post a different look
What goes wrong: Your feed grid looks like 5 different brands. Brand recall on social tanks. Followers cannot tell your post from a random image at a glance.
How to avoid: Lock SREF per content pillar. Visual consistency makes posts recognizably yours at a thumbnail glance.
Not tracking which visual modes win
What goes wrong: You keep producing all 4 pillars equally even though 2 drive 80% of engagement. Wasted effort on modes nobody cares about.
How to avoid: Tag by visual mode in your analytics. Review monthly. Lean into winners.
Recap
Done — what's next
How to use Midjourney style references and SREF codes
Read the next tutorial
Hand it off
Midjourney is one piece of the social media job. Strategy, copy, community management, and analytics are the other pieces. If you want one person to own all of it — including the AI imagery workflow — EverestX matches you with a vetted social media manager in 48 hours. $14-16/hr; most engagements land at $400-1,000/mo.
See specialist rates
In a 2-hour weekly session: 10-15 polished posts across mixed formats (feed, stories, carousel covers, reel covers). After your prompt library is built, productivity compounds — at 6 months in, the same 2 hours can yield 20+ posts.
Depends on quality and execution. With good --sref, low stylize, and post-processing (slight grain, real text overlay), AI-generated imagery passes as professional photography or illustration for ~90% of viewers. The remaining 10% includes other designers and AI-savvy users.
Depends on your brand and audience. B2B SaaS: usually no disclosure needed. B2C lifestyle / influencer-adjacent: increasingly expected. Meta and TikTok have AI content labeling features that auto-tag some AI imagery — assume disclosure is becoming default.
Midjourney cannot accurately render YOUR specific product unless you train a custom model. For real product shots, use real product photography. Use Midjourney for: lifestyle scenes, abstract concept visuals, background/mood imagery, illustrated content.
Canva's AI is convenient and good for quick one-offs. Midjourney is significantly better at: photorealism, brand consistency via SREF, artistic style range, and pure quality. For volume production with brand consistency, Midjourney wins. For 1-2 quick images mid-design, Canva's built-in is enough.
Midjourney
--sref is the difference between "I got a great image once" and "I can ship 100 on-brand images per campaign." This is the workflow that production AI image teams actually use.
Canva
Most teams use Canva's default 1080×1080 template for everything and wonder why their Stories crop weird. Here's the per-format setup with current 2026 dimensions, safe zones, and template patterns that work.
Midjourney
Brand consistency in Midjourney is not luck. It is a system: SREF library, prompt templates, color/lighting rules, and a QA pass. This is the production system pros use.
Midjourney
DIY Midjourney is a great idea — until it is not. This is the honest framework: when the cost of self-managing your AI imagery workflow exceeds the cost of hiring help, and what to look for in a specialist.