Short links in SMS and RCS: the constraints that matter
Character counts, carrier filtering and link previews behave differently in messaging than anywhere else you publish.
By ShortFreeURL Team · 28 August 2026
Segments, not characters
An SMS is billed and delivered in segments. A plain message using the basic alphabet fits 160 characters in a single segment; adding any character outside that set switches the whole message to a wider encoding and drops the limit to 70. Longer messages are split and reassembled, and the per-segment budget shrinks slightly to make room for the reassembly headers.
This is why link length is not a cosmetic concern in SMS. A twelve-character difference in a URL can be the difference between one segment and two, which changes cost and, at volume, changes it substantially. It can also be the difference between a message that displays cleanly and one that arrives split on older handsets.
Avoid characters that silently change the encoding
A typographic apostrophe, an em dash, a curly quotation mark or an emoji will push the entire message into the wider encoding. Copy written in a word processor is full of these, inserted automatically without the writer noticing. The visible symptom is a message that suddenly costs twice as much and truncates earlier than expected.
Compose SMS copy in a plain text editor, or run it through a checker that reports the encoding and segment count before sending. Apply the same care to the link: a slug containing anything outside plain letters and digits is a risk not worth taking here.
Carrier filtering treats unknown short domains with suspicion
Messaging networks filter aggressively for fraud, and shortened links are a standard signal in that filtering. Generic public shortener domains are heavily used by phishing campaigns, which means messages containing them are more likely to be blocked or silently dropped, and you often receive no useful error explaining why.
A branded short domain that consistently sends the same kind of content from the same sender identity is treated far more favourably over time. Use one domain for messaging, keep it separate from any domain used for user-generated or untrusted links, and do not rotate domains to evade filtering, which is exactly the behaviour filtering is designed to catch.
Register your sender and follow the local rules
Most markets now require some form of registration linking a sender identity to a business and, in several, to the specific message templates being sent. Unregistered traffic is throttled or blocked. The rules differ by country and change, and the practical effect is that the same message can deliver perfectly in one market and vanish in another.
Where template registration applies, the link is usually part of the registered template, which means changing the domain or the shape of the URL requires re-approval. Plan link structure before registration rather than discovering the constraint after launch.
RCS changes the surface but not the caution
RCS gives you richer messages: images, buttons, branded sender identity and read receipts where supported. A link can become a button with a label rather than a naked URL, which is better for both clarity and trust. But delivery is not universal, and a fallback to plain SMS has to be authored, which means writing the message twice.
Write the SMS fallback first and make it complete on its own. A fallback that reads like a caption for a missing image is worse than a plain message that was designed to be plain.
Give messaging its own links and its own parameters
Never reuse a link across SMS and another channel. Messaging traffic behaves distinctly, delivery is uncertain in ways other channels are not, and you need to be able to see the click rate for a specific send in isolation. One link per send, tagged with the channel and the send identifier, is the minimum.
This also gives you a fast diagnostic. If a send shows delivery receipts but almost no clicks, you are probably being filtered somewhere between the network and the handset, and the click count is the only place you will see it.
Check the link renders as a link on the handset
Messaging clients decide for themselves what to turn into a tappable link, and they do not all decide the same way. A URL written without a scheme is usually detected, but not always; a URL immediately followed by a full stop is sometimes detected with the full stop included, producing a broken tap; a URL wrapped in brackets or quotation marks can be mangled entirely.
The safe pattern is to put the link on its own with a space before it and nothing but a space or a line break after it, and to end the sentence before the link rather than after. Send a real message to real handsets on more than one operating system before a campaign goes out, and tap the link rather than reading it.
Make the destination trustworthy in one glance
A person reading a text message has almost no context and a well-founded fear of fraud. The destination page should identify your brand immediately, above the fold, without requiring scrolling or a cookie banner to be dismissed first. If the page is slow or looks generic, the recipient reasonably assumes the message was a scam.
Including your brand name in the message text next to the link helps, as does the branded domain itself. Consistency across sends is what builds the recognition, so do not vary the format for the sake of variety.
