A 2026 hiring scorecard for WordPress developers
Twelve questions to ask before hiring a freelance WP dev, the answers that matter, and the red flags AI portfolios are now manufacturing.
The essay
Twelve questions to ask before hiring a freelance WP dev, the answers that matter, and the red flags AI portfolios are now manufacturing.
Why the bar moved
AI-generated portfolios are now indistinguishable from real ones on visual inspection. A six-month-old “freelance WP developer” can ship a portfolio site with 12 client testimonials, three case studies, and an About page that reads like it was written by someone who actually built the work. None of it is real.
The 2026 hiring problem isn’t “find the best portfolio.” It’s “verify any portfolio.” The questions below are the ones AI portfolios cannot fake.
The twelve questions
Portfolio
- Show me the last three sites you shipped, by recency, with live URLs. AI portfolios use mockup or expired domains. Real freelance work is at real domains the buyer can visit and audit.
- What’s the Lighthouse score on the homepage of each? The candidate should know without looking. Off-by-five is fine; off-by-twenty is a tell.
- What’s the LCP at mobile p75? Same. PageSpeed Insights is public; running it during the call is healthy.
Process
- Show me a SOW you’ve used. If they don’t have one, hard pass. If the SOW lacks IP transfer language, negotiate before signing.
- How do you handle “while you’re in there” scope creep? Looking for a process, not just willingness — “2 hours/month included, separately priced beyond that.”
- What’s your milestone schedule? 40/40/20 is healthy. 50/50 concentrates risk at handover.
- Who owns the code at the end? “You do, on final payment.” Anything else is a one-way value transfer.
Communication
- What’s your communication cadence? Friday written status is the gold standard. Weekly Slack updates work. “I’ll email when there’s something to share” is a red flag.
- What’s your concurrent project count? Public count on the vendor’s site is best (mine is 2 / 3 at any time). Vague answer is a tell.
Verifiable evidence
- Can I talk to your last client? If no, hard pass. If yes, email them within 24 hours of the call.
- What’s your AI policy? Transparency, not absolutism. AI for scaffold + human review is healthy. AI for autonomous merge is not.
- Show me your GitHub. AI portfolios can’t fake 18 months of consistent commits at variable hours. Look for: long-tail contribution history, real open-source involvement, commits at human-plausible times.
What AI portfolios still can’t fake
- LinkedIn-verifiable testimonials with real names, real titles, real linked profiles.
- Live PageSpeed Insights URLs that you run during the call.
- A SOW with a real signature.
- A 30-minute conversation with specific timestamps, names, and decisions. “Tell me about the hardest call on the Cyberhive build” — real builders have stories; AI portfolios don’t.
- References who respond within two business days.
What I’d cut from the standard interview process
Coding tests. They optimise for take-home polish, not real-world decision-making. Replace them with: “show me the last bug you investigated; talk me through the diagnosis.” Real builders have specific stories; AI portfolios stall.
Generic “tell me about a difficult project” questions. Replace with specific-to-the-portfolio probes: “the Cyberhive case study mentions 38 plugins down to 11 — which three were the hardest to remove and why?” Real builders go deep on a specific decision; AI portfolios go wide.