## Why One Character Can Double the Cost of a Text Most SMS teams learn about encoding the expensive way. A campaign that looked like a single 150-character text goes out, the invoice arrives, and the message was billed as three parts. Nobody changed the offer or the list.
Somebody pasted the copy from a document that turned straight quotes into curly ones, or added a small emoji at the end, and every phone on the list received a multipart message instead of a single one. SMS length is not measured the way a word processor measures it. The number of characters you see on screen is only part of the story.
What matters to the network is which character set the message needs, how many bits each character takes, and whether the text fits in one message or has to be split into segments that the handset stitches back together. Get that right and you control cost, delivery, and how the text reads on the phone. Get it wrong and you pay more for a message that may look worse.
This guide explains the two encodings almost every bulk sender deals with, GSM-7 and UCS-2 (often just called Unicode), how concatenated messages are counted, which characters quietly force the expensive encoding, and a practical workflow for keeping campaigns inside the segment count you planned. ## The 140-Byte Container Behind Every Text A single SMS carries a payload of 140 bytes, or 1,120 bits. Everything about SMS length follows from that fixed container.
The question is simply how many bits each character needs. - **GSM-7** uses 7 bits per character. 1,120 bits divided by 7 is 160, which is where the familiar 160-character limit comes from. - **UCS-2** uses 16 bits per character. 1,120 bits divided by 16 is 70, so a Unicode message fits only 70 characters in a single part.
- **8-bit data** encoding exists as well, with 140 characters per part, but it is used for binary payloads and is rarely relevant for marketing text. That means the same 100-character message can be one SMS or two, depending only on whether every character fits in the GSM-7 alphabet. ## GSM-7: The Default Alphabet GSM-7 is the standard alphabet defined for SMS.
It covers the basic Latin letters, digits, common punctuation, and a selection of accented and Greek characters that were included for European languages. Characters such as `é`, `è`, `ü`, `ö`, `ñ`, `à`, `Ç`, `Ø`, `Å`, `ß`, `£`, `¥`, `§`, and `¿` are in the basic table, along with Greek capitals like `Δ`, `Φ`, `Γ`, `Λ`, `Ω`, `Π`, `Ψ`, `Σ`, `Θ`, and `Ξ`. Many characters that look just as ordinary are not in it.
The lowercase `á`, `í`, `ó`, and `ú`, for example, are missing from the basic table, as are `ç` in lowercase, curly quotes, long dashes, and every emoji. A single character outside the table means the whole message must be sent in UCS-2. ### The extension table: characters that count twice GSM-7 has a second, smaller table reached through an escape character.
Characters in this extension table are valid GSM-7, but each one costs two character positions: the escape plus the character itself. The extension characters are: - `^` `{` `}` `[` `]` `~` `\` `|` - `€` (the euro sign) - the form feed control character This catches teams out in two ways. A euro price such as `€19` uses four positions, not three.
And a template full of curly braces for placeholders, or square brackets around a call to action, can push a message just over 160 without anyone noticing, because each brace counts twice. ### National language shift tables The GSM standard also defines national language shift tables for languages such as Turkish, Spanish, and Portuguese, and for several South Asian languages. In theory these let a network carry more local characters in 7-bit encoding.
In practice, support depends on the provider, the route, and the receiving network. Unless your provider explicitly confirms support for the route you are using, plan as if a character outside the standard GSM-7 tables forces UCS-2. ## UCS-2: When the Message Becomes Unicode UCS-2 represents each character with 16 bits. It can carry Arabic, Hebrew, Cyrillic, Chinese, Japanese, Thai, Hindi, and the accented Latin characters GSM-7 lacks.
The trade-off is capacity: 70 characters in a single message instead of 160. There is no partial mode. If a 158-character message contains one curly apostrophe, the entire message switches to UCS-2, its capacity per part drops to 67 characters once it is split, and the text that would have been one SMS becomes three. ### Emoji count as two Most emoji sit outside the range UCS-2 can represent in a single 16-bit unit.
They are encoded as a pair of units, known as a surrogate pair, so a single emoji typically takes two of your 70 or 67 positions. Emoji built from several code points, such as skin-tone variants, flags, and family or profession combinations, can take four, six, or more positions while still looking like one picture on screen. That is why "just add a small heart" is rarely a small change.
It switches the encoding and spends several positions at once. ### Languages that are always Unicode If you send in Hebrew, Arabic, Russian, Greek with lowercase letters, Thai, or many other scripts, your messages will be UCS-2 no matter how carefully the copy is cleaned. That is normal.
The fix is not to fight the encoding but to plan for it: write to 70 characters for single-part messages, or budget for multipart sends and keep the copy tight. Our guide to SMS cost control covers how to forecast spend by country and segment count. ## How Multipart Messages Are Counted When text is longer than one message can carry, it is split into segments and sent as a concatenated message.
Each segment carries a small header, called the User Data Header, which tells the handset how to reassemble the parts in order. That header takes 6 bytes of the 140-byte payload in the common format, and in practice it reduces the room for text in every segment.
| Encoding | Single message | Per segment when split | |---|---|---| | GSM-7 | 160 characters | 153 characters | | UCS-2 | 70 characters | 67 characters | So the thresholds look like this: - **GSM-7:** up to 160 characters is 1 segment; 161 to 306 is 2; 307 to 459 is 3; 460 to 612 is 4.
- **UCS-2:** up to 70 characters is 1 segment; 71 to 134 is 2; 135 to 201 is 3; 202 to 268 is 4. Remember that GSM-7 extension characters count as two positions, and many emoji count as two in UCS-2. When a two-position character would straddle a segment boundary, encoders typically move the whole character to the next segment, which can trigger an extra segment a character or two earlier than simple arithmetic suggests.
### What the recipient sees On most modern handsets a concatenated message arrives as a single bubble, so the recipient never notices the segments. That is good for readability, but it hides the cost from the people writing copy. Older devices and some networks occasionally show parts separately or out of order, which is another reason to keep promotional texts short.
### Maximum length Concatenation can technically chain many segments, but providers and routes set their own limits on how long a message can be. Very long texts also tend to perform worse: they are harder to read on a lock screen and more likely to be truncated in notification previews. Check the limit on your provider route, and treat long SMS as an exception rather than a template.
## Billing Is Usually Per Segment Most SMS providers bill per segment, not per campaign message. A two-segment text to 50,000 recipients is typically charged as 100,000 SMS. The difference between 160 and 161 characters, or between a straight and a curly apostrophe, can therefore double the cost of an entire send. SESender lists its public per-message rates on the bulk SMS pricing page.
That table is a 23 Telecom list in EUR per SMS covering 8 countries and 91 network rows. The page notes that prices may vary by plan or package, so confirm the live rate in your account before a paid send. Multiply that rate by the number of segments, not just the number of recipients, when you estimate a campaign. Throughput behaves the same way.
Rate limits and pacing on many routes count segments, so a campaign that doubles in segments can also take longer to deliver. Our post on SMS throughput planning explains how caps and queues interact with send volume. ## The Characters That Quietly Force Unicode Most accidental UCS-2 sends are caused by a short list of characters that sneak in during copywriting and review. Learn to spot them.
### Smart quotes and apostrophes Word processors, notes apps, and many content tools turn `'` and `"` into curly versions. The straight versions are GSM-7. The curly versions are not. A single curly apostrophe in "you're" is enough to switch the message to UCS-2. ### Long dashes The en dash and em dash look like natural punctuation, but neither is in GSM-7. Replace them with a plain hyphen or rewrite the sentence.
### Ellipsis characters Some tools turn three periods into a single ellipsis char
Related Articles
SMS has no undo. Use this pre-send QA checklist to catch segment surprises, broken merge fields, bad links, wrong senders, and mistimed sends.
SMS works because it is personal and immediate. That same intimacy makes compliance non-negotiable.
Customers want relevant messages. They also want stronger privacy. That tension shapes SMS and email programs today.
Personalized texts break on empty fields and long values. Learn how to plan merge fields, fallbacks, length, spintax, and language for SMS at scale.
Discover how AI is transforming marketing message creation in 2026, enabling hyper-personalization, boosting efficiency, and driving engagement. Learn best practices for leveraging
STOP replies shape trust, compliance, and delivery. Learn how to handle SMS opt-out, HELP, and START keywords reliably across every number and route.
Learn how SMS quiet hours and send-time optimization protect opt-outs while lifting engagement — with timezone rules, tests, and a 30-day rollout.
Delivery rate, click rate, opt-out rate, and cost per click are easy to calculate and easy to get wrong. Learn the formulas, the denominators that change results, how segments
Explore SESender
SESender brings audience preparation, contact validation, sender and provider controls, scheduling, delivery tracking, and campaign reporting into one workspace. Review the current product and pricing information before deciding whether the platform fits your messaging workflow.
Explore the platform or review pricing.