¿Puedes enviar alertas de emergencia en español ?

How’s your Spanish?  Ours just got MUCH better.

Hyper-Reach now offers automated translation of English emergency alerts into Spanish.  The feature is available both for text messages and also does text-to-speech, in order to deliver the message as a voice call.

When an agency uses the Spanish option, the recipient will see the text message in both English in Spanish.  For voice calls, the person being called is prompted to “press 1 for English or 2 for Spanish”.  Both text translation and text-to-voice are powered by Google Translate.

More than 20% of US households speak a language other than English and Spanish is the largest and fastest growing non-English linguistic group in the country.  So being able to deliver your message in Spanish could be very important. In some parts of the US, Spanish is the dominant language for more than half the population.

How good is the translation?  A recent head-to-head comparison (Google vs. a human translator) by the translation service Verbal Ink concluded that “It does a pretty good job of getting the gist of a text or recording, certainly enough for basic understanding.”   Other commenters say roughly the same thing, and make it clear that it’s not good for complex documents, nuance or specialized language.  But emergency alerts are short and should be simple and specific in any case.  And it’s worth noting that more than half a billion people, and over half a million websites use Google Translate,  including sites for the states of California, Texas, New York, Washington and Indiana.  It’s even used on USA.Gov!

Can You Read This Message?

A recent post by Tom Phelan got us thinking about the readability of emergency messages.  Dr. Phelan’s point was that many emergency messages are written at too high a literacy level and won’t be correctly understood by many people.

So we decided to test this though by looking at some messages.  The messages are from three communities in different parts of the US using a popular mass ENS service.  The grade level score comes from readability-score.com.

Chance for light snow and patchy black ice this am please monitor your weather radio  **** Grade level: 9.2


19 mile fire in Jefferson County no homes or strctres affected in Butte Silver Bow MT2 frm Roosevelt Dr to MT41 JCT closed  **** Grade level: 10.5

Crdt Scam in the Butte area autocaller from #000-000-0000 caling phones prpmts calrs to enter prsnl info dont answr  **** Grade level: 8.2

We have taken several reports of various scams. Be very careful with activity involving personalinfo or financial transactions.  ****  Grade level 14.2

WEATHER ALERT: ‘Drenching’ Rain, Hazardous Weather Outlook Issued For New Jersey  ****  Grade level 10.5

We think there are more issues with messages like these than just the grade level.  There are typos, abbreviations that might not make sense and a general lack of clarity in some messages.

Try this exercise on messages you’ve sent in the past.  And ask random people (maybe a spouse, child, neighbor or someone at your church) if they understand the messages you’ve sent.  Then think about how to make them clearer.

After all, if the message doesn’t communicate well, you’ve lost an important opportunity.  And if a good message can save lives, a bad message might mean that lives are lost.


Should Infectious Disease PSAP Protocols Cover Informing the Public?

Thanks to Urgent Communications for leading us to this flow diagram, produced by the CDC, which outlines steps to take when receiving a call from someone who thinks they have Ebola.

It looks very helpful, but – despite the use of the word “inform” – we don’t see much about what information, if any, to give to either the folks living in the area of the suspected patient or the general public.

Since PSAPs are often the primary users of emergency notification systems and since these systems are sometimes used to alert the public of infectious diseases, we’re wondering if it’s time to develop protocols for when and how to alert the public in the case of infectious diseases.

The question is much broader than Ebola. Here in Charlotte, NC, we’ve had two cases of food service workers with hepatitis that could have spread to the general public. And the recent outbreak of measles is another case. We’re pretty sure there are many more. And we’re also sure that the protocols to follow will vary with the nature of the disease and how it spreads.

Emergency alert systems such as Hyper-Reach now cover at least 80% of the US population, mostly at the county level. At the rate they are growing, it won’t be long before they reach almost the entirety of the US. But despite that coverage, there isn’t much of a literature on best practices or established protocols for their use. We think it’s time to develop one.

How Not to Use Web Links in Emergency Alerts

Last week, we got this SMS/text message on our Samsung Galaxy 4.  So we clicked on the link.  You can see the page in the next screenshot (actual size).

There are any number of problems with the usefulness of the linked page.  Let’s list some of them:

  1. It’s unreadable on a smartphone.  While the phone let’s me expand the view, that’s not a great solution;
  2. There’s too much going on in this page.  As the receiver of the text, I’m interested in the “missing vulnerable adult”, not general press releases or other alerts;
  3. The actual alert I’m responding to is not on this page.  And worse…
  4. The alert I’m looking for is not on the next page (I clicked “Click here to view more”).

Now a link to a page with a picture of the “missing vulnerable adult” or a white Ford Winstar (my wife can’t tell one car model from another) might have been very helpful.  And after going back through other alerts, at least some of the prior links went to a page that was specific to the alert, so maybe this was just a simple mistake.  But this case illustrates a few principles that alert authors should keep in mind:

  • Text alerts are going to a phone, so format your linked pages to be readable on a smartphone screen;
  • Email alerts may also be going to a phone or a tablet.  Recent consumer surveys show that more than half of Internet access is on smartphones.  And – according to emailmonday.com – more than half of all email opens are on a mobile device.  So even subscribers who get their alerts by email may be reading the alerts on a small screen.
  • Make the link specific to the alert.  Consumer research shows that click-through rates for most short messages (text and tweets) are quite low – often in the range of 1% – 2%.  So the odds of getting a recipient to click twice are very low.
  • Make the content obvious.  Don’t make your recipient search for the information they are looking for.  You want them to see it immediately.


NYS screenshot 1


nys screenshot 2NYS screenshot 3

Loss of Volunteers Threatens Rural Emergency Response

This article suggests that rural areas in America (the article focuses on western and midwestern states) are losing EMS services as volunteer EMTs are retiring.

This is a problem that goes deeper than EMS. Volunteer firefighters are also short in supply, says this story. And rural hospitals are in trouble across the country, especially in states that are not expanding Medicaid.

All of which means that rural America, which supplies much of the food and fuel that the country depends on, may be headed for a state of emergency.

We sell emergency notification services, and we think that’s important. But we also think it’s important to support the folks who are there for victims when emergencies happen.

Formatting Content for the Mobile Web

On Monday, we got the following message via my county’s emergency messaging service:

Residents urged to prepare for hazardous weather conditions. Visit http://charmeck.org for the latest information.

When I clicked on the link, I got the image below (actual screenshot), which is mostly unreadable, because it’s formatted for a computer screen and not for my Samsung Galaxy 4 (which, BTW, has a bigger screen than most phones.)


Now the link came to me over my phone, delivered by text. So to read any of the information here, my options are these:

1) Use my fingers to make the screen bigger, which risks missing the important information (since I’ll only see a quarter or less of the page at any time), or

2) Go find my laptop and look for the link.

To make matters worse, the only substantive content was link to another page (also unreadable on my screen) anda video.

Most surveys on Internet usage are reporting that more than half of Americans’ Internet access is by smartphone, these days. And that share is going to be much higher when the stimulus to get on the web is delivered by phone. So why not format the content for the phone?

If it needs to be delivered by web – and needs to be read or watched – it has to be formatted for the phone, preferably automatically. The technology is there, so there’s no good excuse to do otherwise.

1% Clickthrough Means this: Focus On the Headline

Emergency managers who use Twitter should take this story to heart:  for the vast, vast majority (say, around, 99%), tweets may get read, but links on tweets don’t get clicked.

The story in the Atlantic is in a different context, of course. So maybe more than 1% of your Twitter readers will click through to your content.  But in the absence of data to prove that, we think it’s smarter to assume that they won’t.

And we don’t think this advice is limited to tweets.  Text messages with links may not get read, either.  For example, yesterday, because of an impending winter storm in Charlotte, I got this from the county’s emergency messaging system:

Residents urged to prepare for hazardous weather conditions.  Visit http://charmeck.org for the latest information.

Yes, I clicked on the link, but I’m in the business of emergency messaging, so that’s a poor sample. And I’ll bet that the vast majority of folks who got that text did not bother.

Whats the lesson here?  Simple.  Write your tweets so they have the information you want to convey.  And if 140 characters aren’t enough, write another tweet.  (We suggest the same thing for IPAWS WEA messages, BTW.)  Don’t assume that people will click on the link, because the odds are not in your favor.


To Tweet Emergency Alerts, Think Headline, Not Links

This article in the Atlantic makes a point that every emergency manager using Twitter needs to have burned in their brain: most people don’t click on links in Tweets to read content.

The article is long and the context is different, but the analysis shows only about 1% clicking a link in a tweet.

Now maybe your stats are better. But unless you’ve got proof, assuming that people will click on your links is just blind speculation.

Images are something else altogether. We’d guess that images get much higher click-through rates, which is useful if you’re using tweets to spread an Amber Alert.

And we think this lesson goes beyond Twitter. Monday, we got this from our county’s emergency alert system:

Residents urges to prepare for hazardous weather conditions. Visit http://charmeck.org for the latest information.

So get your point across in the tweet or text itself. And if you need more room than 140 characters, send two messages – or more. That’s our advice for IPAWS WEA messages as well.

After all, if it’s worth sending, it’s worth getting through.

Happy Birthday 911!

47 years ago, the first 911 call was placed in Haleyville, AL. On February 16, 1968, the first-ever 911 call was made by Rankin Fite, Alabama Speaker of the House. The call was placed from Haleyville City Hall to to the city’s police station. US Representative Tom Bevill reportedly answered the phone with “Hello.”

Why Alabama? According to one article, the president of Alabama Telephone Company read an article in the Wall Street Journal that AT&T was planning the use of 911 for emergency dialing for what was then called the Bell System. Alabama Telephone was an independent company, so they would not have been included. But Bob Gallagher, the president of ATC, was a competitive guy and wanted to beat AT&T to the punch.

The development of 911 took many years. The UK experimented with 999 in 1937. In the US, the idea of a single number was first proposed by the National Association of Fire Chiefs for reporting fires in 1957.

And after the Haleyville call, it took many years more for 911 to be adopted across the US and for all public safety emergencies (New York city expanded 911 to fire, police and EMS in 1973.)

Today, 911 covers more than 96% of the US.

IPAWS WEA messages should be longer – duh

Here’s a story we knew was coming. A research study at the University of Maryland has concluded that short messages – specifically, the 90 character messages that are standard in IPAWS wireless emergency alerts – are not as effective as longer messages.