Running 10 Fair Products on Mani: A Portfolio Case Study

I run Downshift, a portfolio of software products built on shared infrastructure. Fair alone has 40+ products spanning e-signatures, email marketing, CRM, scheduling, and more. Each product has its own brand identity. Each needs daily marketing content. Before mani, marketing 10+ brands meant either hiring 10 different freelancers or accepting that most brands would get inconsistent, infrequent content.

Mani was born out of this problem. This case study covers what happens when you run 10 real brands through the same AI creative pipeline simultaneously: what works, what breaks, and the numbers behind it.

The 10 brands

The Downshift portfolio includes products across multiple categories. For this case study, I will focus on the 10 brands that actively use mani for daily content generation. Each has a distinct visual identity, audience, and tone: some are developer-focused with dark-mode aesthetics. Some are consumer-facing with warm, approachable design. Some are B2B with professional, restrained branding. The diversity is deliberate. If mani works across these 10, it works for any brand.

Without naming every product (you can find the full portfolio at faircompany.ai), the spread includes: 3 developer tools, 2 consumer apps, 2 B2B SaaS products, 1 e-commerce platform, 1 creative tool, and 1 marketplace. Colors range from electric blue to forest green to warm amber. Tones range from "technical documentation" to "friendly neighborhood app." No two brands overlap visually.

The multi-brand workflow

Running 10 brands through mani requires a specific workflow that I have refined over 6 months. Here is what the week looks like:

Monday morning (45 minutes): Review the daily queue for all 10 brands. Each brand has 3-5 fresh ads waiting. Swipe through them: approve the ones that feel right, reject the ones that do not. This is the curation step. It takes about 4-5 minutes per brand. By 9am, I have 20-30 approved ads ready to schedule across all brands.

Monday afternoon (30 minutes): Export approved ads to each brand's scheduling tool (Later, Buffer, or direct to Meta Ads Manager depending on the brand). Set up the week's posting schedule. This is the distribution step.

Wednesday (15 minutes): Check which ads from Monday's batch are performing well. Star the top performers in mani so the learning loop biases future generations. Flag any ads that flopped (usually an audience mismatch or an off-brand color that slipped through).

Friday (30 minutes): Generate campaign-specific creative for any brand with a launch, sale, or event in the coming week. This is the ad-hoc generation step, separate from the daily queue.

Total weekly time across 10 brands: about 2 hours. Output: 30-50 approved ads per week. Per-brand cost: effectively $8/month per brand on the Pro plan (5 brands per subscription, 2 subscriptions for 10).

The numbers

Over the first 6 months of using mani for the portfolio, here are the aggregate numbers across all 10 brands:

Total ads generated: approximately 3,600 (60 per week * 60 weeks, but actual ramp was slower). Approval rate: 42% (1,512 ads approved out of ~3,600 generated). The approval rate increased from 28% in month 1 to 55% in month 6 as the learning loop improved. Publishing rate: 38% of approved ads were published (the rest were saved for future use or held as backlog). Total published: approximately 575 unique ads across 10 brands over 6 months.

Performance comparison: ads generated by mani performed within 15% of our best manually-designed ads on average CTR. On 3 specific brands, mani-generated ads outperformed historical manual creative by 20-30%, which I attribute to the higher volume of testing (more variants = more chances to find a winner).

Cost comparison: the alternative would have been 10 freelance designers at $500-1,000 each per month, or $5,000-10,000 monthly. Mani cost $158/month for 2 Pro subscriptions. The cost reduction is 97%. The time reduction is harder to quantify but I estimate we saved 80-100 hours per month across the portfolio compared to briefing, reviewing, and iterating with individual freelancers.

What works well across 10 brands

Brand consistency scales. The Brand DNA scan for each brand is done once. From that point, every ad for that brand uses the same color palette, tone, audience, and visual style. After 6 months and 360 ads per brand, the visual consistency holds. An observer looking at the last 50 ads from any single brand would see a coherent brand, not a random collection.

The daily queue creates a habit. Before mani, I marketed the portfolio brands in bursts: a flurry of content when a launch was coming, then silence for weeks. The daily queue eliminated this pattern. Every morning, fresh ads are ready. The consistency of output is the most valuable change, not any individual ad's quality.

Cross-brand learning. Interesting emergent behavior: when one brand's UGC-style ads outperform, I apply the same style bias to similar brands. The pattern recognition across 10 brands is richer than what any single brand would reveal. Running multiple brands is actually an advantage for ad quality because the data feedback loop is 10x wider.

What does not work (yet)

Seasonal creative still needs manual input. The daily queue generates general-purpose ads. For BFCM, Valentine's Day, product launches, and other time-specific moments, I still need to manually generate campaign-specific creative with specific briefs. The queue does not know my seasonal calendar (yet).

Some brands need more visual variety. After 100+ ads from the same Brand DNA, the visual compositions start to feel repetitive. The colors, tone, and products are correct, but the layouts settle into 3-4 patterns. I address this by periodically updating the brand's visual style descriptor or by manually requesting specific composition types (flat lay, lifestyle, close-up).

Cross-brand switching takes mental effort. Even with mani handling the creative production, reviewing 10 brands in sequence requires context-switching. Each brand has a different standard for "good." I have learned to review brands in clusters (all developer tools together, all consumer apps together) to maintain the right quality filter per context.

Advice for multi-brand operators

If you manage multiple brands (agency, portfolio, franchise), here is what I have learned from 6 months of multi-brand mani usage:

First, invest the time in getting each brand's DNA scan right. Spend 10 minutes reviewing and correcting the extracted tone, audience, and products for each brand. This upfront investment pays for itself across hundreds of generations.

Second, establish a weekly rhythm and stick to it. The portfolio marketing workflow only works if it is ritualized. Monday morning review, Wednesday performance check, Friday campaign prep. Without the rhythm, it degrades to "I will get to it when I have time" and that means never.

Third, build a swipe style guide for each brand: 5 approved ads that represent the ideal output. When you are unsure whether a generated ad is "on brand enough," compare it to the style guide. This makes the approval decision faster and more consistent, especially as your brand count grows.

If you run a multi-brand operation and you are managing creative production across brands, the math is simple. Try one brand on mani's free tier. Generate 20 ads. If the quality meets your standard, scale to the Pro tier and onboard all your brands. The per-brand economics at scale are unmatched by any other creative production method I have found.