To meter SMS usage per customer, record one usage row for every text your platform sends or receives for that customer, count it in segments (not messages), and key the row to your own tenant id. Then total the rows per billing period and apply your own price. That's the whole design, and nearly all the bugs come from counting the wrong unit or attributing a send to the wrong customer.

This post is for the product or engineering lead at a platform that texts on behalf of its customers and wants to bill them for it. We make Ready, so the cost facts are Ready's, but the metering design works over any provider. For the alerting side of usage billing, see overage alerts by SMS.

Meter the unit you're charged

Ready bills per segment, and a segment isn't a message. A single SMS segment holds 160 GSM-7 characters. Longer texts split into 153-character segments. One emoji or other non-GSM character drops the limit to 70 characters, 67 per segment on a longer text. Each segment is billed.

If you meter messages and your provider bills segments, your margin leaks on every long or emoji text. If you meter segments, your usage rows match the upstream bill line for line, and you can reconcile them.

TextCharactersEncodingSegments
"Your table is ready."20GSM-71
A 175-character promo175GSM-72
A 140-character text with one emoji140Unicode3
A 175-character promo with one emoji175Unicode3

Ready swaps curly quotes, long dashes and ellipses for plain ones before sending, so those never change the count on Ready. Your estimator should do the same, or it will overcount smart punctuation pasted from a word processor.

Also remember the footer. Unless the account turns it off, Ready adds "Reply STOP to opt out" and the business name to the first text to each contact, and again from time to time. A first text near a segment limit can tip into a second segment because of it. Leave a margin when you estimate.

What the upstream cost looks like

Per segment, the upstream cost is $0.0075 plus a $0.0045 carrier fee passed through at cost, so $0.012 all-in. That figure doesn't include screening: the Scrub lists and Number validation switches, each $0.003 extra per segment when switched on for an account. There's no volume discount and the rate doesn't drop as you grow. Accounts opened before November 12, 2026 keep their signup prices, shown under Billing. The pricing page has the full list.

Ready's all-in prices per segment include the carrier fee but not screening: Scrub lists and Number validation, the two switches in Settings under Compliance, each add $0.003 per segment when switched on, shown as screening on your bill, and both are off by default. Unless you turn the opt-out footer off, the first text to each contact, and one again from time to time, also gets "Reply STOP to opt out" and your business name if the message does not already have them, so a text close to the segment limit can bill one more segment.

Other costs are per customer, not per text:

  • Registration: $25 a month per customer, covering the brand and its first campaign. Each extra campaign is $20 a month.
  • Numbers: $2 a month per number.
  • Local presence, if a customer turns it on: $0.0015 for each segment of a delivered text, on top of the text itself.

Your plan can cover these in the base fee and put the segments in metered usage, or pass everything through.

Attribute every send to a tenant

Per-customer metering needs one rule: every send carries a tenant id from the moment it's created, not one inferred later. Two ways to get there, and they work best together.

Sub-accounts. Ready's sub-accounts (GET /v1/sub-accounts, POST /v1/sub-accounts) give each of your customers its own numbers and its own carrier registration. Sending from a customer's sub-account means the numbers, the registration and the traffic are separate by construction, which is also what the carriers expect. Multi-tenant SMS architecture covers the isolation.

Your own ledger. Don't rely on reading the provider's usage back to find out who sent what. Write a usage row in your own database in the same transaction that enqueues the send, with the tenant id, a direction (sent or received), the segment count you computed and the idempotency key. The key is what ties your row to exactly one send.

curl -X POST https://api.tryready.com/v1/messages \
  -H "X-API-Key: $READY_API_KEY" \
  -H "Idempotency-Key: 3f6a2b9c-5d41-4e7a-8c2d-1b0e9f7a6d55" \
  -H "Content-Type: application/json" \
  -d '{"to": "+12125550199", "body": "Your table is ready.", "consent_attested": true}'

The key above is a placeholder. In your system it's the one stored with the usage row. A retry reuses it, so a retried request neither sends twice nor needs a second usage row. Idempotency keys and duplicate texts explains the pattern.

The edge cases that break the invoice

Most disputes with a customer come from rows that look wrong. Decide each of these before you launch.

  1. Replies are usage. A text you receive pays the per-segment rate plus a $0.003 received carrier fee, lower than the send fee. If your plan says "texts sent", your margin on a chatty customer is negative. Meter received segments as their own direction.
  2. Refusals aren't charged. Ready refuses opted-out numbers, numbers on your Ban list and known landlines up front, with no charge. Don't bill your customer for a text that never went out. Write the usage row when the send is accepted, not when it's attempted.
  3. Held texts are charged when they send. An API text is held only if the quiet-hours window opens within 6 hours, and otherwise it isn't sent and the API answers 409 QUIET_HOURS_BLOCK. Nothing is charged for the held attempt. Meter on the send that goes out.
  4. Scheduled texts aren't charged until they send. A text set with POST /sms/schedule should create its usage row at send time, not at the time you scheduled it.
  5. Free allowances are yours to set. Ready charges every text from the first one, so if your customers get a free allowance, it comes out of your margin. Decide it on purpose and meter it like any other usage.
  6. Bulk calls. POST /v1/messages/bulk sends one text to up to 100 numbers, so write one usage row per recipient, with that recipient's segments, not one row per call.

A worked example, with assumptions

Suppose a customer sends 4,000 texts in a month. Assumptions for the example: the average text is 2 segments, the customer receives 600 replies of 1 segment each, and none use screening or local presence.

  • Sent: 4,000 × 2 = 8,000 segments
  • Received: 600 × 1 = 600 segments
  • Total metered: 8,600 segments

Your upstream cost for the month is 8,600 segments at $0.012 each, plus the customer's registration and number charges. Whatever you charge on top is your margin decision. The point is that your ledger and the upstream bill should agree on 8,600. If your ledger says 4,600 because it counted messages and ignored replies, you've given away roughly half the usage.

Reconcile every month

A meter you never check drifts. Once a month, compare your ledger totals per tenant against the upstream bill and look for the gap:

  • A ledger lower than the bill usually means received segments, long texts or emoji segments went uncounted.
  • A ledger higher than the bill usually means you wrote rows for texts that were refused or held.
  • A gap that's the same size every month points to a footer or smart-punctuation counting rule.

For the history side of reconciliation, you can pull sent and received texts into your own database with GET /v1/messages. Page through it with before_id and upsert on the message id. Then compare counts per tenant against your usage rows.

The practical takeaway

Meter segments, key every usage row to a tenant and an idempotency key, count replies as their own direction and don't bill refusals or unsent holds. Use sub-accounts so each customer's traffic and registration are separate from the start. Then reconcile monthly against the upstream bill and fix whichever rule the gap points to. When you want to test the numbers, sign up and compare a few sends against what your ledger recorded.