Should you really replace underscores with dashes in your URLs?

Why is there a distinction between underscores and dashes?

Historically, Google treats dashes as word separators in URLs, whereas underscores are considered linking characters. Specifically, "my-site.com/running-shoes" is interpreted as two distinct words ("shoes" and "running"), while "my-site.com/running_shoes" can be read as a single term.

This nuance has fueled a debate about URL optimization for years. SEO purists have long recommended consistently favoring dashes to maximize semantic clarity.

What is the actual benefit of migrating from underscore to dash?

Mueller is clear: the positive impact is minimal. We are talking about such a marginal effect that it does not justify a dedicated migration operation. Modern crawlers understand a page's context far beyond just the structure of the URL.

The engine now employs hundreds of signals — semantic content, recognized entities, co-occurrences, internal linking — that greatly overshadow the microscopic weight of a character in a URL. The underscore is no longer a significant technical handicap.

What are the concrete risks of a massive URL revamp?

Any URL modification leads to a re-indexation phase by Googlebot. Even with perfect 301 redirects, you expose your site to temporary volatility in rankings while Google recalculates the authority and relevance of the new addresses.

Worse still: poor management of redirects — 301 chains, loops, unmonitored 404 errors — can break your link architecture and cause drastic traffic drops. The effort is rarely worth the candle for such a tiny gain.

• Dashes are preferred for new URLs, but not to the point of justifying a project on the existing ones
• A URL migration always carries risk: temporary loss of positions, dilution of PageRank, redirect bugs
• The ROI of an underscore → dash operation is almost zero compared to the engineering time and risks incurred
• Focus your efforts on structural optimizations that are truly impactful: speed, linking, content quality

Is this statement consistent with real-world observations?

Absolutely. Audits of sites with a mix of underscores/dashes rarely show a correlation between separator type and organic performance. The real ranking levers are elsewhere: content depth, thematic authority, user experience.

Let's be honest: if your site is stagnating on page 2, it's not because your URLs contain underscores. That's a false problem that some junior SEOs continue to raise reflexively, without measuring the actual delta. [To be verified] if you observe an inverse correlation, but fifteen years of field experience have never confirmed it.

In what cases can a migration still be justified?

If you're carrying out a comprehensive technical overhaul — changing CMS, complete restructuring of the hierarchy, migrating to a new domain — then yes, take the opportunity to standardize on dashes. The marginal cost of harmonization becomes negligible in this context.

Similarly, if your current URLs are illegible or filled with unnecessary parameters ("?id=12345&cat=abc_def_ghi"), a redesign focused on readability and semantic structure makes sense. But in this case, the problem is not the underscore; it's the overall shaky architecture.

What critical mistakes must be absolutely avoided?

The worst decision is to launch a URL migration without a comprehensive redirection plan. Each old URL must point to its new version via a clean 301 — no chains, no detours through the homepage. Map everything out before switching.

Another classic trap: not monitoring the Search Console post-migration. 404 errors, soft-404s, redirect loops can take weeks to surface if you're not actively tracking them. A daily follow-up is non-negotiable for the first 30 days.

What should you actually do with your existing URLs?

Don't touch anything if your current URLs contain underscores and the rest of your SEO is healthy. Instead, invest this time in content optimization, cleaning up internal linking, or improving Core Web Vitals.

For new sections or pages to create, systematically adopt the dash as a separator. This is a best practice that incurs no cost and avoids any future ambiguity. Configure your CMS and URL generators accordingly.

How to quickly audit the real impact on your site?

Extract through Screaming Frog or your favorite crawler the complete list of indexed URLs. Segment between dashes, underscores, and mixed. Cross-reference with GSC data (impressions, clicks, CTR) to identify if a pattern emerges.

Spoiler: you probably won't see any significant difference. If your underscore URLs perform just as well — or even better — than those with dashes, you have your answer. Move on.

When should you consider a migration anyway?

Only if it fits within a larger technical project that is already budgeted and planned. CMS overhaul, late HTTPS migration, multi-domain consolidation, restructuring of hierarchy — these are all occasions when harmonizing to dashes costs almost nothing more.

In that case, rigorously document the old → new mapping, test your redirects in a staging environment, and deploy in waves if possible to limit exposure to risk. A phased deployment allows for correcting the course before an error spreads.

• Keep your current underscore URLs unless a comprehensive redesign is justified
• Adopt dashes for all new URLs created from now on
• Never migrate without a comprehensive plan for 301 redirects that is tested in staging
• Monitor Search Console daily for 30 days post-migration if you proceed
• Prioritize optimizations with proven ROI: content, linking, speed, user experience
• Document every URL change in a centralized tracking file (mapping, dates, reasons)