Designer and developer portfolios are packed with clues about the candidate's identity. Domain names, client names, project details, and even file metadata all risk exposing who the applicant is. For recruiters running blind hiring processes, this leaks bias signals that blind hiring is meant to remove.
Here's a quick comparison of anonymisation methods for portfolios, so you can pick what fits your workflow:
| Method | Time per portfolio | Accuracy | Skill required | Notes |
|---|---|---|---|---|
| Manual | 10–30 minutes | Varies | Basic editing skills | Slow, error-prone |
| Scripted/Regex | 1–5 minutes | High | Regex knowledge | Needs maintenance |
| AI/ML | <1 minute | Moderate | No coding | Can miss niche or complex data |
| Distill one-click | <1 minute | High | None | Tailored for recruitment use |
Why anonymising portfolios matters in blind hiring
Blind hiring aims to reduce unconscious bias by hiding information that could influence decisions unfairly. Portfolios often reveal:
- Client or company names that suggest a candidate's background or network
- Domain names or URLs linked to personal websites or companies
- Project descriptions that mention locations or industries
- File metadata containing author names or creation dates
Under the Equality Act 2010, employers are generally required to avoid discrimination based on protected characteristics. While portfolios aren't explicitly covered, recruiters risk bias creeping in if identifying details remain visible. Removing these clues helps focus hiring decisions on skills and outputs, not personal history.
Method 1: Manual anonymisation
Open the portfolio file (PDF, Word, or web link) and comb through each page. Look for:
- Company and client names
- URLs and domain names
- Location or contact details embedded in project descriptions
- Author names in file properties or watermarks
- Photos or personal branding
Replace or redact these with generic terms like "Client A" or "Project X". Save a copy and check again for missed details.
Pros:
- Complete control
- Can handle tricky context
Cons:
- Very time-consuming (10–30 minutes per portfolio)
- Easy to miss subtle identifiers
- Not scalable for high volume
Method 2: Scripted or regex-based anonymisation
Write or use existing scripts to identify and remove patterns in text or file metadata:
- Regex can catch URLs (e.g.,
https?://[^\s]+) and email addresses - Lists of client names can be replaced via string matching
- File metadata stripped via command-line tools (ExifTool for PDFs/images)
This method requires building and maintaining pattern lists and scripts adapted per client or project type.
Pros:
- Fast once set up (1–5 minutes per portfolio)
- Reproducible and consistent
Cons:
- Requires regex and scripting knowledge
- Can fail on unexpected or new data formats
- Needs ongoing updates for new clients or domains
Method 3: AI and machine learning anonymisation
Some tools use AI to scan documents and flag personal or identifying information automatically. They can parse context better than regex, spotting indirect identifiers like "the London-based fintech client".
Pros:
- Very fast (<1 minute)
- Minimal user effort
Cons:
- Accuracy varies — AI can miss subtle or industry-specific details
- May flag false positives or miss complex identifiers
- Usually a paid or subscription service
If you try AI tools, always review outputs carefully before submission.
Method 4: Distill’s one-click anonymisation
Distill offers a tailored tool that strips identifying signals from portfolios automatically. It removes:
- Names, emails, phone numbers
- Client and company names based on your custom lists
- URLs and domain names
- Graduation years and other date identifiers
- Photos embedded in documents
Distill works with common CV and portfolio formats, producing a clean document ready for blind review in seconds.
Pros:
- Instant and consistent
- No technical skill required
- Easily integrates into recruitment workflows
Cons:
- Custom client lists need initial setup
- May not catch very rare or bespoke identifiers without updating lists
Edge cases and pitfalls to watch out for
- Embedded images: Screenshots or graphics may contain logos or client names. OCR or manual review may be needed.
- Metadata: Some file types store author or device info invisible in the document body. Strip metadata with dedicated tools.
- Linked files or pages: Web-based portfolios might link to personal social media or company sites. Check all links.
- Project descriptions: Avoid vague "anonymisation" that removes too much detail and undermines candidate evaluation. Keep skill-relevant info that doesn't identify.
- File formats: Some formats (e.g., Photoshop files) are harder to anonymise automatically and require manual handling.
FAQ
Can I fully anonymise a portfolio without losing important details?
Yes, but it takes care. Replace identifiers with generic terms rather than deleting whole sections. Keep project context and skills clear.
Is anonymising portfolios legally required?
Not explicitly, but it supports fair hiring practices under the Equality Act 2010 by reducing bias risks.
How often should I update client and domain lists for regex or Distill?
Update lists whenever you take on new clients or see new domain patterns in portfolios. Quarterly reviews are a good starting point.
Can AI tools handle non-English portfolios?
Some can, but accuracy depends on the language support of the AI model. Always verify outputs.
What if candidates submit portfolios as websites?
Download or screenshot key pages for anonymisation, or ask candidates to provide anonymised files.
If you send 20 or more portfolios a week for blind hiring, automating anonymisation saves hours and reduces errors. Distill strips names, emails, client details, URLs, and photos from portfolios with one click—saving time and cutting bias risk.
Try Distill free and see how quick anonymising portfolios can be.