Every vendor in this category leads with a number. We link to those numbers on the provider pages and label them as vendor claims, because that is what they are. Here is how to read them without either dismissing them or believing them.
“98% of texts are opened”
Directionally true and close to useless. A text arriving on a lock screen is 'opened' in the sense that the words entered someone's field of vision. It does not mean it was read, understood, or acted on. The figure is also decades old and gets recycled without a source.
What you should take from it: messaging genuinely does get seen far more reliably than email. What you should not take from it: that seen equals effective. Measure the action, not the impression — measuring ROI.
“3x more replies than SMS”
Plausible, and the mechanism is real: blue bubbles carry trust signals that green bubbles do not, and threads with unfurled link previews from a recognised domain look less like scams. But almost every published version of this figure comes from a vendor measuring their own customers, with no control for message quality, list quality or industry.
Selection bias is doing most of the work
The businesses that adopt an iMessage API are businesses that already care about customer communication. Comparing their results to an industry-average SMS blast measures the difference between two kinds of company, not two protocols.
“5–10% reply rates”
This one is more honest than most, because reply rate is a real behaviour rather than an impression. It is also highly dependent on what you send. A transactional reminder asking for a one-letter confirmation will beat 10% comfortably. A cold-ish marketing message to a lapsed list will not come close.
What numbers to actually expect
Rather than a benchmark, use the shape. Confirmation and reminder messages that ask for a single-character reply get very high response. Anything asking someone to leave the thread — tap a link, fill a form, book a slot — drops sharply. Anything marketing drops again.
- Ask for a reply, not a click. Every step out of the Messages app costs you response.
- One action per message. Two options roughly halves the response to each.
- Transactional beats promotional by a wide margin, because the recipient was expecting it.
- Your own baseline is the only benchmark that matters. Measure four weeks, then improve on it.
Run a holdout and stop arguing about benchmarks
Withhold the message from a random 10% and compare completion rates. That single number tells you what the channel did for your business, which no industry average can. How to set one up.
Next step
Generate a tagged link for whatever you send next with the UTM builder, see what this looks like in your industry, or compare the services that can send it on the providers page.