Brand DNA extraction gives you a strong starting point. Editing transforms that starting point into a precision instrument for ad generation. Most users make 2-5 edits after their first extraction. The founders who take 10 minutes to refine their Brand DNA see measurably better ad output for months afterward.
The editing philosophy: refine, do not replace
The extraction captures what your brand IS based on your website. Editing lets you adjust what your brand SHOULD BE in advertising. These are sometimes the same thing, but not always. Your website might use formal language because it contains legal terms and technical documentation. Your ads should probably be more casual and direct. The tone edit addresses this gap.
The principle: use the extraction as the foundation and make targeted adjustments. Do not delete everything and start from scratch. The extraction captured real signals from your website. Even imperfect results contain useful data that you would not have identified on your own.
Editing tone descriptors
Tone is the highest-impact field to edit because it affects every word of generated copy. The extraction gives you 3-5 adjectives. Review each one and ask: "If someone read our ads and had to describe the voice, would they use this word?"
Common edits: replacing formal descriptors with casual ones ("professional" to "founder-voiced"), adding specificity ("friendly" to "friendly-irreverent"), or removing mismatches ("corporate" if your ads should feel personal). The order matters. The first descriptor has the strongest influence on generation. Put your most important voice quality first.
Test your edits: after changing tone descriptors, generate a test batch of 5 ads. Read them aloud. Do they sound like your brand? If not, adjust the descriptors and test again. Two or three iterations usually dial in the perfect tone.
Editing the audience description
The audience description is the second highest-impact field. It determines who your ads speak to and what pain points they address. A specific audience produces tight, relevant copy. A vague audience produces generic copy that resonates with nobody.
The gold standard for an audience description includes: demographics (age, location, income), psychographics (values, beliefs, behaviors), and pain points (what problem drives them to seek your product). Example: "Female DTC founders, 28-42, running beauty brands on Shopify, spending $5K-$20K/month on Meta ads, frustrated by the cost of freelance designers and the inconsistency of template-based tools."
That description is specific enough to generate copy that references Shopify, Meta ads, freelance designer costs, and the template-consistency problem. Each detail in the audience description becomes a potential messaging angle in generated ads.
Editing the color palette
Palette edits are the most straightforward because colors are objective. Click any swatch to open the color picker. Paste hex codes from your brand guide or Figma file. Ensure all 5 roles (primary, secondary, accent, text, background) are filled.
Common issues: the scanner sometimes picks up background grays or border colors as palette entries. Replace these with your actual brand accent colors. The palette drives the visual appearance of generated ads, so accuracy here directly affects output quality.
If you do not have a formal brand guide, use a browser extension to pick colors from your website. The 5 most prominent colors on your homepage are your palette. Do not overthink this. The palette from your existing website is the right palette for your ads because it is already associated with your brand in your audience's memory.
Editing products
Add any products the scanner missed. For each product, include: name (as it appears on your site), a one-sentence description, the price (if publicly listed), and optionally the product page URL. Products referenced in your Brand DNA get mentioned by name in generated ads.
For brands with large catalogs (50+ SKUs), focus on your 10-15 hero products. Not every product needs to be in the Brand DNA. The products listed are the ones Mani will feature most prominently in generated ads. Choose the products you most want to promote.
Editing keywords
Keywords influence the vocabulary of generated copy. Add industry terms, branded phrases, and specific language your audience uses. Remove generic terms that do not distinguish your brand.
A practical approach: read your 5 best-performing social media posts. Note the specific words and phrases that appear. These are your keywords. They represent the language your audience already responds to, which makes them the best vocabulary for generated ads.
Editing visual style
The visual style description guides creative direction. Be specific: "Clean product photography on white backgrounds with soft shadows" is actionable. "Modern design" is not. Describe: the lighting (studio, natural, dramatic), the composition (product-centered, lifestyle, flat-lay), the styling (minimal, maximalist, editorial), and the color treatment (saturated, muted, monochromatic).
When to re-extract vs edit
Re-extract when your website has changed significantly: a rebrand, a new product line, or a complete redesign. Manual edits are overwritten on re-extraction, so save any values you want to keep before clicking Re-scan.
Edit (do not re-extract) when you want to adjust how your brand sounds in ads without changing your website. Most brands edit 2-3 times in the first month, then settle into a stable Brand DNA that works well for ongoing generation.
The next module covers managing multiple brand profiles for founders who run more than one brand.
For more on Brand DNA extraction, see How Brand DNA extraction works in the help center.