- Platform dashboards often credit SMS with orders it did not cause, especially when a delivered text counts toward attribution. Check your settings before you trust the revenue number.
- Every text has an all-in cost: platform fees, message segments, carrier fees, MMS upgrades, discounts and the subscribers who opt out. Count all of it per message.
- The metric that decides whether SMS pays is incremental contribution margin per message, measured with a holdout. Everything else is a diagnostic.
Why do SMS reports often overstate what texting earns?
SMS feels like the best-performing channel in many dashboards. Texts are read quickly, and orders follow soon after. But “followed by an order” is not the same as “caused an order,” and the way your platform assigns credit decides how big the number looks.
Take Klaviyo as an example. Its help center says new accounts default to a last-touch model with a 5-day window for SMS clicks and a 12-hour window for SMS deliveries. Delivery-based attribution means a subscriber who received a text, never tapped it and bought through a Google search eight hours later can still be credited to SMS. Older accounts and other platforms use different settings, so open your attribution settings and write down what they are before you compare anything.
Last-touch credit also means SMS and email compete for the same orders. If a customer clicks an email on Monday and receives a text on Wednesday, the text can take the credit. Klaviyo’s help center advises considering email and SMS attribution settings together when you interpret results. Add the two channels’ attributed revenue and compare it with total store revenue from subscribers. If the sum looks too good, it probably is.
What does one text message really cost?
“Cost per message” is rarely one number. Prices vary by provider, plan and carrier, so we will not quote any here. What matters is that you add up every line.
| Cost line | What drives it | How to control it |
|---|---|---|
| Platform or credit cost | Your plan; many providers bill per segment or per credit | Know how your plan counts segments and MMS |
| Message length and encoding | Longer texts split into more segments | Write tight copy; check the segment count before sending |
| Emoji and special characters | Switch the message to a different encoding with fewer characters per segment | Use them only when they earn their cost |
| MMS (images, GIFs) | Usually priced higher than SMS | Test whether the image lifts profit, not just clicks |
| Carrier fees | Passed through by many providers; vary by message type and number type | Ask your provider how they are billed |
| Discount cost | Codes used by people who would have bought anyway | Hold offers for segments that need them |
| Opt-outs | Each unsubscribe removes a subscriber’s future value | Track opt-outs per send and per flow |
The encoding point surprises many teams. Klaviyo’s help center explains that a single segment holds up to 160 standard (GSM-7) characters, but a message with special characters or emoji uses UCS-2 encoding: 70 characters for a single segment and 67 per segment when the message runs longer. One emoji can turn a one-segment text into a two- or three-segment text, multiplying the send cost.
Which SMS marketing metrics actually matter?
Some metrics tell you whether the program is healthy. Only one tells you whether it pays.
| Metric | How to calculate it | What it tells you |
|---|---|---|
| Delivery rate | Delivered ÷ sent | List quality and carrier filtering |
| Click rate | Unique clicks ÷ delivered | Whether the message was relevant |
| Click-based conversion rate | Orders from clickers ÷ delivered | A cleaner view than delivery-based revenue |
| Opt-out rate per send | Opt-outs ÷ delivered | How fast you are wearing out the list |
| All-in cost per message | Total SMS cost ÷ messages delivered | What each text really costs |
| Revenue per recipient | Attributed revenue ÷ recipients | Useful for ranking messages, not for proving profit |
| Incremental margin per message | (Lift in revenue × contribution margin − cost) ÷ messages | Whether texting pays |
Opt-out cost belongs in the math
Every opt-out is a small loss. If an average SMS subscriber is worth something in future margin, a campaign that triggers a wave of opt-outs has a hidden cost. Estimate the value of a subscriber from your own data: average margin earned per SMS subscriber over the following months. Multiply by opt-outs and subtract it from the campaign result. A send that looks profitable on day one can be negative once you count the subscribers it lost.
Separate flows from campaigns
Automated flows such as abandoned cart and browse reminders reach people at a moment of intent, so they usually earn more per message than broadcast campaigns. Campaigns are where costs and opt-outs pile up. Report the two separately, and judge each SMS campaign type (launch, restock, sale, content) on its own. Our guide to SMS abandoned cart covers the highest-intent flow in detail.
How do you measure the real lift from SMS?
Attribution tells you which orders a text touched. A holdout group tells you how many orders it caused.
The method is simple. Before a campaign, randomly set aside a share of eligible subscribers (often 10% to 20%) who do not receive the text. Keep everything else the same, including email. After a fixed window, compare purchase rate and revenue per person between the two groups. The difference, multiplied by the size of the sent group, is the incremental revenue. The same logic underpins all incrementality testing: small groups give noisy answers, so keep the holdout large enough to detect the lift you care about.
Run holdouts on your biggest flows too. An abandoned cart text may look like it recovers thousands of dollars, but some of those shoppers would have returned through email or on their own.
Then convert incremental revenue into profit. Multiply by contribution margin, not gross revenue, and subtract the all-in cost of the messages and any discounts.
Worked example: one sale campaign, four answers (illustrative numbers)
All numbers below are illustrative.
A store has 44,444 eligible subscribers for a weekend sale. It texts 40,000 and holds out 10% (4,444) who get nothing. The message includes two emoji and runs two segments.
| View | Revenue credited to the text |
|---|---|
| Dashboard (delivery and click attribution) | $14,000 |
| Click-based only | $8,500 |
| Holdout lift (sent group vs. holdout, scaled to 40,000) | $5,000 |
Now the costs, using illustrative rates for this store’s plan:
- Messages: 40,000 × 2 segments × $0.015 per segment = $1,200.
- Carrier fees (illustrative): $200.
- Sale discount on incremental orders: already reflected in the lower order value.
- Opt-outs: 0.5%, or 200 subscribers. At an estimated $3 in future margin each, that is $600.
Incremental profit: $5,000 × 45% contribution margin = $2,250. Subtract $1,200, $200 and $600, and the campaign earned $250.
The dashboard said $14,000. The campaign actually made a small profit. Rewriting the text in one segment without emoji would have saved $600 in message cost, more than tripling the result. That is the kind of decision only an all-in view reveals.
How often should you text, and when?
Frequency is a cost question as much as a brand question. Each extra campaign adds message cost and opt-out risk, and its incremental lift usually shrinks as subscribers hear from you more. Track opt-out rate and holdout lift as frequency rises. When the extra send stops paying, stop sending it.
Timing has a legal side too. Recent class actions have targeted brands for promotional texts sent outside 8 a.m. to 9 p.m. in the recipient’s local time, citing the TCPA. Whether those rules apply to consented subscribers is disputed, so take advice from counsel and use your platform’s quiet hours settings. Our guide to SMS compliance covers consent and timing in more depth.
SMS metrics checklist
- [ ] Record your SMS attribution windows, including whether deliveries count.
- [ ] Compare SMS plus email attributed revenue with total subscriber revenue.
- [ ] Build an all-in cost per message: platform, segments, MMS, carrier fees and discounts.
- [ ] Check segment count and encoding before every send.
- [ ] Track opt-out rate per send and estimate the value of a lost subscriber.
- [ ] Report flows and campaigns separately.
- [ ] Use click-based conversion as a cleaner day-to-day signal.
- [ ] Run holdouts on major campaigns and your biggest flows.
- [ ] Convert lift to contribution margin before calling a send profitable.
- [ ] Review frequency when opt-outs rise or lift shrinks.
FAQ
What is a good click rate for SMS?
It depends on your list, offer and message type, and published benchmarks rarely match a given store. Compare against your own history by message type. A falling click rate with steady sends usually means the list is tiring or the messages are less relevant.
Should we turn off delivery-based attribution?
For decision-making, click-based attribution plus holdout tests gives a more honest view. If you keep delivery-based credit for reporting continuity, label it clearly, and never use it alone to justify more spend or more sends.
Is SMS worth it for a store doing under $5M a year?
It can be, when flows like abandoned cart and back-in-stock carry most of the volume and campaigns stay selective. Start with flows, measure them with a holdout, and add campaigns only when the all-in math works. Our guide to SMS marketing for eCommerce covers how to build the program.