Brand DNA extraction works best on content-rich websites with clear branding. Sometimes it misses, especially for new sites, single-page apps, or highly visual brands with minimal text. Here is how to identify and fix extraction gaps.
Common extraction gaps
- Generic audience: "Modern professionals" or "businesses" instead of your actual target. This is the most common gap. Edit the audience field with specific demographics, psychographics, and pain points. For example: "Female DTC founders, 28-42, running beauty brands on Shopify, spending $5K-$20K/month on Meta ads."
- Missing products: If your products are behind a login, loaded dynamically via JavaScript, or listed on a separate subdomain, the scraper may not find them. Add products manually with names, descriptions, and prices.
- Wrong tone: If your website copy is formal (corporate pages, legal-adjacent language) but your ads should be casual and founder-voiced, override the tone descriptors. The website tone and ad tone do not need to match.
- Incomplete palette: The scanner pulls colors from CSS stylesheets. If your brand colors are primarily embedded in images or SVGs, the palette may be incomplete. Add hex codes manually from your brand guide or Figma file.
- Missing keywords: Industry jargon, competitor names, or product-specific terms that do not appear on your homepage may be absent. Add them to the keywords field.
Manual override workflow
- Run the initial extraction to get a baseline. Even a partial extraction saves time versus building everything from scratch.
- Review each section of your Brand DNA card. Check tone, audience, palette, products, keywords, and visual style.
- Click the edit icon on any section that needs correction.
- Make your changes and save.
- Generate a test batch of 5 ads to verify the output matches your expectations. If not, iterate on the Brand DNA fields and regenerate.
Requesting a re-extraction
If your website changed significantly (rebrand, new product line, redesign), click "Re-scan" to run a fresh extraction. Manual overrides will be replaced with new extracted values. Save any manual values you want to keep before re-scanning.