SMS vs Email for Client Communication
Builds guide sites for a livingPublished July 24, 2026Last updated July 26, 202612 min read
- Emergency alerts are not text messages; many providers transmit them over a separate technology.
- Carrier participation in that system is voluntary, and non-participating carriers must notify customers.
- Route by whether a fact changed after booking, not by which channel feels more urgent.
- Cancellations go on both channels within a minute of each other, because both have conditions.
- The weather and water data behind urgent messages is published free and without a key.
Ask a guide why they text and the answer is that texts get read. It is probably true and it is doing more work in that sentence than it can carry. The country runs an emergency messaging system built for exactly one job, reaching phones when something is about to happen, and reading how it works is a useful corrective. Carrier participation in it is voluntary. It is not sent over text messaging at all. And it still needs the handset switched on, out of airplane mode, and in range of a tower belonging to a carrier that takes part. If the federally coordinated version carries that many conditions, the version you send from a phone on a boat ramp carries them too. Which is an argument for choosing by consequence rather than by habit. The rest of the client-communication material sits under the getting-booked hub.
Which message goes down which pipe
| Message | Channel | Why |
|---|---|---|
| Something changed today | Text | Reaches a pocket, not a desk |
| Meeting time and place | Both | Needed twice, at different distances |
| What to bring | Has to be re-findable | |
| Deposit and terms | A record either side may need | |
| Off-season news | Nobody wants it in a pocket |
Why is a text more likely to arrive?
It is not, exactly. It is more likely to be seen.
Those are different claims and conflating them is where the reasoning usually goes wrong.
A message on a phone competes with a handful of other notifications; a message in an inbox competes with everything a person receives that week.
That difference is about attention rather than delivery, and attention is genuinely where the advantage sits.
Delivery, meanwhile, has conditions on both sides, and the conditions on the phone side are the ones guides never examine.
The emergency alerting guidance lists them plainly for its own system: the device must be capable, switched on, off airplane mode, and receiving service from a participating carrier.
Every one of those applies to an ordinary message too, minus the participation question.
What arrives in an inbox has a different set of conditions, and they are the subject of the deliverability piece.
What does the emergency system tell us?
That reaching a phone reliably is hard enough to need a federal programme.
The system was established by legislation in 2008 and became operational in 2012, and the guidance describes it as a public and private partnership between the emergency management agency, the regulator and the wireless industry.
Since launch it has been used nearly 96,000 times, for dangerous weather, missing children and other critical situations.
Its messages appear on a handset as text-like messages with a distinct attention signal and vibration.
Text-like, and not text: the guidance is explicit that many providers transmit these using a technology separate and different from voice calls and text messages.
Which means the thing built to be unmissable deliberately avoided the channel guides assume is unmissable.
That is not an argument against texting clients. It is an argument against assuming the pipe is magic.
The consent rules governing what you may send by text are a separate matter, set out in the text marketing piece.
How precise is any of this geographically?
Precise now, and it took a decade.
The alerting programme's targeting history is a useful measure of how hard the problem is.
At launch, participating providers were generally required to send alerts to an area no larger than the affected county or counties.
From 2017 they had to transmit to an area that best approximated the affected area, even where that was smaller than a county.
And since December 2019 they must deliver to the area the originator specified with no more than a tenth of a mile of overshoot.
The trade association cited in the guidance estimated that about 83 percent of consumer smartphones supported that enhancement in 2023, up from about 60 percent in 2022, 34 percent in 2021 and 18 percent in 2020.
Handsets that do not support it still receive alerts on the older geographic basis, which is the shape of every phone-side improvement: real, gradual, and unevenly distributed across your client list.
The alerting material is published by the FCC in its consumer guide, reviewed in September 2025.
So which messages have to be texts?
The ones tied to a fact that changed after booking.
Almost everything a guide sends is known at the time of booking and can therefore be written once, in an email, weeks ahead.
What cannot be is the small set of facts that move: the weather, the water, the launch, the time.
A client who booked in March cannot be expected to check anything the night before, and the fact that moved is the whole content of the message.
That is the test. If the message exists because something changed, it is a text. If it exists because something needs to be known, it is an email.
Applied honestly, that routes perhaps three messages per trip to the phone and everything else to the inbox.
It also explains why guides who text everything find their texts stop being read: the channel's advantage is scarcity.
The sequence built on that principle is described in the pre-trip text piece.
Can you know in advance what changed?
For weather and water, yes, and for free.
The weather service publishes a public interface covering forecasts, alerts and observations, and states that all of the information presented through it is intended to be open data, free to use for any purpose.
It charges nothing, describing itself as a public service of the government, though it applies rate limits it describes as generous for typical use and does not publish the exact figure.
Exceeding a limit returns an error and may be retried after it clears, typically within five seconds, and requests through a proxy are more likely to hit it than direct ones.
It asks for one thing in return: a user agent identifying your application, with contact details if you are willing, so you can be reached about a security event.
None of that requires a key, an account or a card, which is unusual enough to be worth saying twice.
So the trigger for a weather message can be automatic even when the message itself is written by hand.
The interface is documented by the National Weather Service.
How many texts a trip actually justifies. Count the facts that can change between booking and launch: the forecast, the water, the meeting time and the launch itself. That is four candidates. Now count how many actually changed on your last ten trips, which for most operations is a handful across the whole ten rather than four per trip. Multiply your trips per season by that observed rate and you have the honest volume of texts your operation generates, which is almost always far below what a messaging tool's pricing tiers are built for. Everything else you were planning to send is an email that could have been written in March. The four candidate facts are a category count, not a measurement; the rate has to come from your own last ten trips.
What about the water itself?
Also published, also free, and station by station.
The oceanographic service runs a public interface for observations and predictions from its own stations, addressed by a seven-character station identifier.
Requests can be made for a date range, for a number of hours back from now, or with a keyword: today, latest, meaning the last data point available within the last eighteen minutes, or recent, meaning the last seventy-two hours.
Data length is capped by interval, with one-minute data limited to four days, six-minute data to a month, hourly and high-low data to a year, and daily means to ten years.
Water level data is verified monthly for the preceding month, so January's data is verified in February, which is worth knowing before quoting a figure from last week.
The service asks callers to space out successive requests and to download only what they need, which is a courtesy rather than a published limit.
For a guide, the practical use is a station near your launch and a rule about what reading changes a plan.
The interface is documented at the NOAA tides and currents service.
Does email have any advantage at all?
Two, and both are about time.
An email can be found again, which a text on a phone effectively cannot once a week of messages has passed over it.
Anything a client will need to refer back to, meaning what to bring, where to park, what the terms are and what happens if the weather turns, belongs where it can be searched.
The second advantage is that an email can be written long before it is needed and sent by machine, which is how a one-person operation gets consistent.
Texts written in advance read like texts written in advance, and clients notice.
There is a third, quieter advantage, which is that an email costs nothing per message and a text does not.
At a guide's volume that difference is small, and it stops being small the moment somebody sells you a campaign.
The reference email carrying all of that gets built in the trip reminder piece.
What about the cancellation message?
Both, in that order, within a minute of each other.
This is the one message where the cost of not arriving is a client driving three hours for nothing.
Send the text first, because it reaches a pocket, and then the same message by email, because it survives and can be forwarded to whoever else was coming.
Do not make the email an afterthought summary. Make it the same message with the detail the text could not carry.
Then call anybody who has not acknowledged either within an hour, which sounds excessive until the first time somebody drives.
The reason for the belt and braces is everything above: both channels have conditions, and this is the message where a condition failing costs a relationship.
Write it once, before the season, so the version you send at five in the morning is not composed at five in the morning.
The wording for that is in the cancellation scripts piece.
Which channel for reaching past clients?
Email, almost always.
Reaching somebody who fished with you two seasons ago is not urgent by definition, since nothing changed today that they need to know about.
It is also the message most likely to be unwelcome in a pocket, and the one most likely to be forwarded to a friend if it lands in an inbox.
The consent position differs between the two channels in ways that matter and are covered elsewhere, so treat the text version as the one needing more care rather than less.
There is a narrow exception, which is a client you know personally and have texted before, where a text is a message from a person rather than a campaign.
Scale that exception and it stops being the exception, which is how operations end up with a problem.
The practical answer is a seasonal email to everybody and a handful of individual texts to people you would text anyway.
What those individual messages say is in the past-client text piece.
How do you collect the two, and when?
The address at booking, the number at confirmation.
Almost every booking flow already captures an address, because it needs somewhere to send the confirmation.
The number is usually captured too and then never used deliberately, sitting in a booking system nobody sends from.
Getting both into one place you can actually send from is the unglamorous work that makes any of this possible.
Ask for the number as the one you will text on the morning, because that phrasing tells the client what it is for and sets the expectation you want.
Ask for the address as the one that will hold the details, which does the same job for the other channel.
Both framings also make the later message easier to write, since you already said what you would use it for.
The habit of gathering addresses on the water is described in the collecting emails piece.
Which habits cause the trouble?
Seven, beginning with sending everything by text.
Sending every update by text until the channel's scarcity, which was the whole advantage, is gone.
Putting the reference details in a text, where a client cannot find them the night before.
Treating a text as proof of delivery, when the conditions on the phone side are outside your control.
Sending the cancellation by only one channel, which is the single message that justifies both.
Composing the urgent message in the moment rather than before the season.
Assuming the consent position is the same for both channels, which it is not.
And never writing down which channel carries which message, so the decision gets made again every time under pressure.
What happens to the email side when nobody is watching is covered in the spam piece.
What surprises operators here?
That the emergency alerts are not text messages.
Most people assume the alert that buzzes their phone is a text, and the guidance says many providers use a separate and different technology.
The second surprise is that carrier participation in that system is voluntary, and non-participating carriers are required to tell their customers.
The third is that the geographic precision people take for granted only reached a tenth of a mile in December 2019.
The fourth is that the weather and water data behind most urgent guide messages is free, keyless and machine-readable.
The fifth is how few messages actually qualify as urgent once you apply the test honestly.
Taken together, the picture is less about channels and more about which facts move.
The registration a business needs before texting at any scale is explained in the registration piece.
What separates the operations that get this right?
One written page, decided in the off-season.
One page listing every message the operation sends, the channel it goes down, and who writes it.
That page is what stops the decision being remade at five in the morning, which is when it is made worst.
They also write the urgent messages in advance, in the off-season, when the phrasing can be considered.
They send the cancellation on both channels as a rule rather than a judgement call.
They keep the address list and the number list in the same place, which sounds trivial and is the reason most operations cannot send anything.
And they review the table once a year, usually after the season that produced the message they wished they had sent differently.
None of that requires a tool. It requires a decision made once instead of forty times.
The email half of that table starts with the welcome sequence piece.
The routing, in order
Test by change, then by whether it must be findable.
Ask first whether the message exists because a fact changed after booking. If it does, it is a text.
Ask second whether the client will need to refer back to it. If they will, it is an email regardless of the first answer.
Messages answering yes to both, which in practice means the cancellation and the meeting change, go on both channels within a minute of each other.
Write every message in that set before the season, so the urgent one is never composed under pressure.
Point the weather and water triggers at the free public interfaces rather than at a paid dashboard.
Keep the addresses and the numbers together, in something you can actually send from.
Treat the consent position for texting as the stricter one and act accordingly.
Then write the table on one page and leave it alone until next winter.
What a booking assistant should and should not answer is weighed in the booking chatbot piece.
Read rates, open rates and response rates appear nowhere above. Not for texting, not for email, not compared against each other, and not for any operation or market. Where a number of that kind would sit, nothing has been put in its place, because the widely repeated figures for both channels come from vendors selling one of them and none describes a guiding operation. What is quoted instead is what the emergency alerting guidance states about its own system and what two public data services state about their own terms. Read 26 July 2026: the alerting guide, then carrying a September 2025 review date, and both service documentation pages. Rate limits, thresholds and consent rules change. Consent obligations for text messaging get their own page here; none of this is legal advice.
How this was checked. The alerting material is quoted from Wireless Emergency Alerts, an FCC consumer guide at fcc.gov/consumers/guides/wireless-emergency-alerts, read on 26 July 2026 and carrying a last updated or reviewed date of Thursday, 25 September 2025. Taken from it: that the system has been used nearly 96,000 times since its 2012 launch to warn the public about dangerous weather, missing children and other critical situations; that the Warning, Alert and Response Network Act established it in 2008 and it became operational in 2012; that wireless companies volunteer to participate, in a public and private partnership between the Federal Emergency Management Agency, the FCC and the United States wireless industry; that an alert appears on the recipient's handset as a text-like message accompanied by a unique attention signal and vibration; that in answer to whether these are text messages, many providers have chosen to transmit them using a technology that is separate and different from voice calls and SMS text messages; that participation is widespread but voluntary, that some carriers may offer the service over only parts of their service areas or devices and others not at all, and that the FCC requires carriers that do not participate to notify customers; that to receive one, a handset must be capable, switched on, not in airplane mode, and in the vicinity of and receiving service from a tower of a participating carrier; and the targeting history, being an area no larger than the affected county or counties at launch, an area best approximating the affected area from 2017 even where smaller than a county, and from December 2019 delivery to the area specified by the originator with no more than a tenth of a mile overshoot, with the cited trade-association estimate that about 83 percent of consumers' smartphones supported that enhancement in 2023, up from about 60 percent in 2022, 34 percent in 2021 and 18 percent in 2020. The weather service terms are quoted from the API Web Service documentation at weather.gov/documentation/services-web-api, read the same day: that all information presented via the interface is intended to be open data, free to use for any purpose; that as a public service of the United States Government no fees are charged, although reasonable rate limits are in place, that the rate limit is not public information but allows a generous amount for typical use, that an exceeded limit returns an error and may be retried after the limit clears, typically within five seconds, and that proxies are more likely to reach the limit than direct requests; and that a user agent is required to identify your application, with contact information optional so the service can reach you if the string is associated with a security event. The water data terms are quoted from the CO-OPS Data Retrieval API documentation at api.tidesandcurrents.noaa.gov, also read the same day: the seven-character station identifier; the date options, including latest meaning the last data point available within the last eighteen minutes and recent meaning the last seventy-two hours; the data length limits of four days for one-minute interval data, one month for six-minute data, one year for hourly and for high and low data, and ten years for daily means and daily maximum or minimum data; that water level data is verified on a monthly basis for the past month, with January verified in February; and the request to provide spacing between successive calls and to download only the data you need. No figure was retrieved from either service for this page and none is asserted. The arithmetic panel counts categories and contains no measurement. Nothing here is legal advice.
If your booking calendar has more open weeks than you’d like, I’ll build you a free preview of your booking site before you pay a cent.
Get a free website previewThe routing table, message by message
Are texts really more reliable than email?
They are more likely to be seen, which is a different claim from more likely to arrive, and conflating the two is where the reasoning goes wrong. Both channels have delivery conditions. The FCC's guide to the emergency alert system lists the phone-side ones for its own service: to receive an alert, a handset must be capable, switched on, not in airplane mode, and in the vicinity of and receiving service from a tower belonging to a participating carrier. Every one of those applies to an ordinary message too. The advantage of a phone is attention, since a message there competes with a handful of notifications rather than a week of mail.
Are emergency alerts text messages?
No, and this surprises almost everyone. The FCC's guide answers the question directly: many providers have chosen to transmit these alerts using a technology that is separate and different from voice calls and SMS text messages. They appear on a handset as text-like messages with a unique attention signal and vibration, which is why people assume otherwise. The system was established by the Warning, Alert and Response Network Act in 2008, became operational in 2012, and has been used nearly 96,000 times since. That the service built to be unmissable deliberately avoided the text channel is worth knowing before assuming that channel is magic.
Which messages should be texts?
The ones that exist because a fact changed after booking. Almost everything a guide sends is known at booking and can be written once, in an email, weeks ahead. What cannot is the small set of facts that move: the weather, the water, the launch, the time. A client who booked in March cannot be expected to check anything the night before, and the fact that moved is the whole content of the message. Applied honestly, that routes about three messages per trip to the phone and everything else to the inbox. It also explains why guides who text everything find their texts stop being read.
Which messages should be email?
Anything the client will need to find again. A text is effectively unfindable once a week of messages has passed over it, so what to bring, where to park, the terms, and what happens if the weather turns all belong in an inbox. Email has a second advantage: it can be written long before it is needed and sent by machine, which is how a one-person operation stays consistent. Texts written in advance read like texts written in advance. There is a third, quieter advantage, which is that an email costs nothing per message. At a guide's volume that gap is small until somebody sells you a campaign.
What about the cancellation?
Both channels, text first, within a minute of each other. This is the one message where the cost of not arriving is a client driving three hours for nothing, and it is the message that justifies the belt and braces. Send the text because it reaches a pocket, then the same message by email because it survives and can be forwarded to whoever else was coming. Do not make the email an afterthought summary; make it the same message with the detail the text could not carry. Then call anybody who has not acknowledged either within an hour, which sounds excessive until the first time somebody drives.
Can the trigger be automated?
For weather and water, yes, for free and without a key. The National Weather Service documents a public interface for forecasts, alerts and observations, stating that all information presented through it is intended to be open data, free to use for any purpose, with no fees as a public service of the government. Rate limits exist but are not published, described as generous for typical use, and an exceeded limit returns an error that may be retried, typically within five seconds. It asks only for a user agent identifying your application. NOAA separately documents a water level and predictions interface addressed by a seven-character station identifier.
Which channel for past clients?
Email, almost always. Reaching somebody who fished with you two seasons ago is not urgent by definition, since nothing changed today that they need to know. It is also the message most likely to be unwelcome in a pocket and most likely to be forwarded to a friend from an inbox. The consent position differs between the channels in ways that matter, so treat the text version as needing more care rather than less. The narrow exception is a client you know personally and have texted before. Scale that exception and it stops being one, which is how operations end up with a problem.
Sources & methods
- Wireless Emergency Alerts (WEA), FCC consumer guide
- API Web Service documentation, National Weather Service
- CO-OPS Data Retrieval API documentation, NOAA Tides and Currents
Every figure here is traced to a named public source and checked against it. Licensing, tax, and fee rules change. Verify your state’s current rules with the agency directly before you count on any number here.
Evan Knox
I build booking websites and run the ads and search for owner-run fishing guides, one operation per stretch of water. My first guide client, Bowman Fly Fishing, grew its revenue 4x in a year from that work. Field Notes is where I put the straight numbers on the business of guiding.
More field notes
Texts for what changed. A site for everything else.
I'm Evan, and the guide whose clients always know where to be, and can always find the booking page, is the one who fills a calendar. I build booking sites and run the search and local SEO for owner-run guide operations, one operation per stretch of water. Text me at (470) 777-9686 and I'll build you a free preview of your site before you pay a thing.