Do custom domains raise conversion? What the data actually says

The claim people want to cite is "giving customers their own domain raises conversion by some percentage". We looked for a study behind it: a randomized comparison of the same product with and without custom domains, measuring sign-ups or purchases. We did not find one. Every figure we found measures something narrower, and was published by a company that sells branded links.

That does not mean custom domains do nothing. It means the honest statement is smaller: branded links have been reported to earn more clicks in some settings, and customers pay for custom domains. Whether the feature raises your conversion is a question your own data can answer, and the section on testing it below shows how.

Branch reports the only experiment of its own among these sources. It ran two Facebook ads that differed only in the link:

We ran two identical Facebook ads — same audience, same budget, same content.

When the ad included a preview image, both links performed similarly. But when we removed the preview image, the ad with the branded link saw double the CTR.

The page does not publish impressions, clicks or how long the ads ran, so the size of the effect cannot be checked. It is also one pair of ads, and in the version with a preview image, the one most ads use, the branded link made no difference.

Rebrandly, in its guide to branded short domains, says that sharing branded links on a custom domain:

It helps build trust, enhances brand recognition, and increases click-through rates by 40% or more.

The guide does not describe a study, a sample or a method behind the figure.

PIMMS goes a step further and talks about conversion, not clicks:

The data is clear: branded short links boost conversion rates by 39% compared to generic links, but the real business impact goes far beyond click-through rates.

Elsewhere on the same page the 39% is a click-through figure ("With a 39% higher click-through rate"), footnoted to a LinkDrip blog post and a Rebrandly help-center article. The page also cites Branch's ad test and Rebrandly's guide above. We found no experiment of PIMMS's own on the page, so we do not repeat its other percentages here.

Three caveats apply to all of them:

• They sell branded links. Each publisher is a link-management vendor, writing about its own product category.
• A short link is not a product hostname. A branded link in an ad is seen for a second before a click. A custom domain on a SaaS product is where a customer's visitors, or the customer themselves, spend their time.
• Clicks are not conversion. A higher click-through rate is an early step in a funnel. None of these sources follows the visitors who clicked to a purchase.

Some of the best-known products that serve customers' sites charge for custom domains. Webflow's free Starter site plan publishes to a webflow.io subdomain, and its paid site plans add a custom domain. The pricing page describes that feature, "Publish to a custom domain", as:

Connect a domain you own and publish directly to it.

Notion sells it per domain, on top of the plan:

Connect a custom domain to your Site and remove Notion branding for $8/month/domain paid annually, or $10/month per domain paid monthly.

That is evidence that customers value the feature enough to pay for it. It is not evidence that it lifts conversion for anyone: a product can charge for something because it matters to buyers without that thing changing what their visitors do. Gating it behind a paid plan also means these vendors cannot easily tell whether it would have converted free users, which is the test the last section describes.

DoDomain's data is about the step before any conversion: getting a customer's domain connected at all. The State of Domain Connect page probes the Domain Connect APIs of the major DNS providers every night. As of 2026-11-04, 8 of 14 watched providers serve every one of DoDomain's one-click templates. A real one-click apply, with the record written and verified on the authoritative nameservers, has been proven end to end on Glauca, Cloudflare, Vercel, GoDaddy only.

Everywhere else, the customer copies records into their DNS provider by hand. That is where a custom-domain feature loses people: at a DNS editor they rarely use, sometimes at the wrong company (see registrar vs DNS host). Setup completion and support load are the effects of a custom-domain feature you can measure with certainty, and they are worth measuring before any conversion claim.

Offer it during the trial. If custom domains unlock only after upgrading, no trial user ever has one, and you cannot see whether having one changes whether they upgrade. For the test, make the custom-domain step available to trial accounts.

Randomize the offer, not the outcome. The tempting comparison, accounts that connected a domain against accounts that did not, measures selection: a team willing to edit DNS in week one was probably going to buy anyway. Instead, show the custom-domain step to a random half of new trials, and compare the two halves by what they were assigned, whether or not each account connected a domain. Report the connected-versus-not split as a description, never as the result.

Decide what you will measure first. Write down the primary metric and the smallest effect worth acting on before the test starts, so the result cannot be chosen after the fact. Useful measures:

• Trial-to-paid conversion, by assignment: the primary metric.
• Retention of the accounts that did convert, a billing cycle or two later.
• Time from sign-up to a verified domain. With DoDomain, the signed connection.verified webhook carries the moment it went live.
• Abandonment inside the connect flow: sessions started against domains verified, and the step where the rest stopped.
• Support tickets about DNS or domains, per hundred trials, in each arm. This is the support cost the feature adds, and the one a one-click setup can remove.
• Domains that break after going live. DoDomain reports those as connection.failed, and each is a customer whose branded page went dark.
• If your customers share links to what they build with you, the click-through rate on those links, by whether the link is on their domain or yours. That is the one place the branded-link studies above are directly relevant.

Run it for a full billing cycle. Trial-to-paid only resolves once the trial ends, and retention needs at least one renewal after that. A test stopped the week it looks good is measuring noise.

Custom domains are a feature customers demonstrably pay for, and link vendors report click-through gains for branded links that their pages do not show the method for. Neither is a measured lift in SaaS conversion. If you want that number, it is one randomized trial offer and one billing cycle away, in your own product.

To build the feature itself, the subdomain vs custom domain entry covers the trade-offs, and the guides walk through it in Next.js, Rails, Laravel and Django.

No card, no sales call: create an app, mint a session, and watch the first domain verify.

All external pages retrieved 2026-09-23. Quotes are verbatim.

1. Branch, Do Branded Links Matter? Benefits & Best Practices. Its own two-ad Facebook test of a branded versus a generic link.
2. Rebrandly, The Complete Guide to Branded Short Domains. Its claim that branded links increase click-through rates by 40% or more.
3. PIMMS, Branded vs Generic Links: Conversion Rate Impact 2025. Its 39% conversion claim, and the pages it footnotes for it.
4. Webflow, Pricing. Which site plans publish to a webflow.io subdomain and which to a custom domain.
5. Notion, Pricing. The per-domain add-on price for connecting a custom domain to a Notion Site.
6. DoDomain, State of Domain Connect. Which DNS providers serve one-click setup templates, measured nightly, and where a real apply is proven.