Jason Brenier, Director of Strategy at Georgian Partners Impact discusses conversation, NLP and AI

Jason is a linguist that works at Georgian Partners, a growth-stage investment firm.  Georgian Partners focuses on conversational business as an investment thesis. They’ve written a great deal of best practices and think pieces on where the space is heading. 

As part of the Impact team, Jason helps Georgian Partners companies use NLP and advanced linguistic tactics to improve products and customer outcomes. Jason has more than 15 years of experience applying advanced analytics, computational linguistics and machine learning to business problems in the financial services, pharmaceutical and legal domains.

Our conversation covered many topics with a focus on how language, linguistics and NLP work for technology companies. 

You can read more about Georgian Partners, their investment thesis and any of their whitepapers at https://georgianpartners.com

The benefits of SMS as a marketing & communications channel

Why companies and organizations should do SMS.

The solo series has been about explaining messaging channels. We’ve been concentrating on SMS and there is a lot of background information and terms that we’ve discussed over the last two episodes. It’s been a lot of detail, this episode is the payoff.

We’re going to talk through from an organization or marketers point of view, what SMS is good for and how an organization should use the channel. These are not the only answers, but they are pretty comprehensive and we’re talking at a high level, so it covers a lot.

I’d love to discuss all of this or answer questions. So if you disagree, or are thinking about a use case that I don’t mention, please connect on Messenger so we can discuss. CTA.

Let’s talk about the cold hard facts. First and foremost, SMS is the most ubiquitous channel of communication on the planet. I’d imagine there are more phones on the planet than any other type of device and every single phone has SMS. Like every one. Almost everyone on the planet has one.

I’ve helped to launch campaigns that were focused on homeless populations and SMS was the obvious and only channel. You might be thinking that you’re a marketer and the homeless is not in your target demographic, but young people it’s their only channel in a different way. They just don’t use email even if they do have to have an email address for some reason. I don’t think anyone really uses email until they have a job that requires it.

Anyway, I’m getting ahead of myself. First factor is the ubiquity of SMS – absolutely everywhere.

The second factor is the importance of the channel. Text messaging is the way that people talk to their parents, children, best friend and significant other. It’s simply the most important channel out there.

Now this is still true, but I feel it’s slipping a bit. I have some friends that moved out of the country and we use Facebook Messenger. I have one friend in the UK and we use Whatsapp. Most things that are group messages we’re doing on Messenger. And of course I have friends that work in the messaging space, so we use Allo ironically. Although I do like it a lot.

For now, SMS is still king. When anyone here’s that sound they automatically & instinctively look at their phone. It simply can’t be helped. In fact I’d argue that the messaging ping, SMS or your messenger of choice, is the number #1 trigger in the world. If you’re a marketer and want to get someone’s attention there is nothing that grabs attention like the message ping.

What the ubiquity and importance of the SMS channel means is that when an organization sends a message, it’s going to be read and it will be read quickly. That’s something easy to say and easy to hear, but the numbers that are floating around are that messages have a 99 percent open rate and the average test is read in 90 seconds. This is just simply insanity. I bet for the average link on the internet, 99% of the people that click, don’t actually make it to the page. That’s how crazy a 99% open rate is.

NOW, there is a chance these numbers aren’t exactly true. It’s really hard to tell. It’s just way too much of an echo chamber out there. If you google for it, there are a million hits, all the numbers are impressive, but they are all companies sell or promote SMS campaigns or news stories talking about SMS campaigns. It’s unclear what the source for this data is and/or who a credible source would actually be.

BUT even with that uncertainty, SMS is ubiquitous and an important channel for almost everyone. Because of those facts open rates are very high and opens happen very quickly.

What is SMS good for from a marketing/communication channel view? First, let’s review what it’s NOT good for. Remember, SMS requires an opt in, so it’s not an advertising channel. By advertising channel I mean that a brand can use the channel to find new people. They can pay to get their message in front of people. This is not possible with SMS. You cannot just send a user a text, they would have had to message in first. And if they messaged in first, that means the brand reached them somewhere else and asked them to message in.

This is actually a common question. It’s not possible to buy a list of numbers and send them a text. That’s a big reason why you, as a text message user do not get spam texts often.

SMS is not an advertising channel, but it’s great as a response channel to advertising. This is actually the first big super-power of SMS. If you do a tv commercial and the call to action is to join, call 1800 or go to www.com or text join to 12345, the text message call to action will beat the 800 number and website. From a call to action someone is more likely to text in rather than call in or visit a site.

Let’s even take this a step further. Once someone texts in they are more likely to provide their email address over SMS compared to the phone or URL response channel. The first superpower of SMS is as a response channel (response to a brands ads and media) and that means that the brand gets more people, more leads from all of their advertising efforts.

Here’s a theoretical example. We’ve all seen heard and read a one hundred Geico commercials. Whether it’s cavemen or the lizard, they all basically say that I can save 15% or more on car insurance by getting a quote from Geico.com. From what I’ve seen with clients if Geico added or text the word GEICO to 12345 each commercial would drive more people to get a quote.

I’ve seen many different tests and they usually work something like this. Let’s stick with the Geico example. In order to give a quote, Geico needs to understand zip code, make & model of the car, age of the driver and miles driven each year in order to give a quote. So sending the user to the website is just the first step of this process. Same with the user texting in, it’s just the first step of the data collection process.
With test I’ve run in the past the SMS channel will drive about 3 times more leads than the URL call to action from a TV commercial. That’s huge. This ratio will change depending on the advertising medium. For a TV commercial, viewers will probably be near a computer or they’ll have a phone with a browser. If someone were to stand up at an event and do the same call to action… a person would be much more likely to text in rather than go to a URL.

Basically, the further away a viewer/listener is from a computer, the bigger the ratio will be.

Think about a scenario where Bruce Springsteen wants to help a charity…

Another big advantage for SMS as a response channel is that it’s an instant subscription. If a user messages in, but doesn’t provide the information asked for, the organization can follow up. If a user goes to the URL and just leaves, that user is gone.

Again, the best way to describe this value is that SMS works as a response channel and will drive MORE PEOPLE – whether that means leads, donors, customers etc.

Super power number 2 for SMS has to do with activation. When a message is sent, it’s going to be seen and it’s going to be seen quickly. The best use for messages is to drive action from the recipients.

In the Geico scenario, it would be better to send an activation message such as “remember that you need to add a credit card to the account before you’re covered” rather than an information message like, “Geico was rated #1 by JD Power and Associates.”

Texts are great at getting people to do things. Sometimes that means replying with data, sometimes it means a real world action and sometimes it can lead to long term behavior change. The one thing that I’ve seen over and over again is that the mobile list will take more action compared to any other list a marketer can access. Most of the time we’re comparing the mobile list to email.

I’ve seen campaigns where people on the mobile list are more likely to vote, donate or even quit smoking. This activation idea is a big deal for brands – Geico wants people to buy their insurance, Amazon wants people to sign up for Prime and a presidential campaign wants people to donate and vote. Everything comes down to activation at some level.

Here’s where it gets interesting. People will probably not finalize their insurance on the phone most of the time. They probably don’t sign up for Prime on the phone and they definitely can’t vote on the phone. But the people on the mobile list that receive a text will be more likely to take the action even if the action isn’t on text or even the phone.

An example that I used all the time. It’s a trick, but just think through it with me. Let’s say that you are the marketer for a CRM company like Salesforce. People come to a webinar and give their email address to attend the webinar. After the webinar there are automated follow ups over email. The whole point is that we want to nurture the leads that attended the webinar and get them to signup for salesforce give a credit card etc. So let’s say 100 people attend the webinar and Salesforce has email addresses for all of them. How many people signup for Salesforce through that email???

It’s a trick question. Zero people sign up through the email. It’s actually impossible to do that. People signup on a webpage. The email will link people to the webpage or simply remind them to go to the page and signup. SMS works the same way.

It’s slight of hand because email is definitely a direct route the signup page. With SMS it’s tough to link out and get the person to signup on their phone. With SMS it’s also probably that people get the text, but then signup on the desktop computer. And it’s also common that the text would actually capture the person’s attention, increase the chances that they open the email and then convert.

Whatever the route, the data is pretty solid. We would regularly see the mobile list convert 3-5x better than the non-mobile list. Just to get really specific, the text campaigns are usually working in the same department as the email campaigns or CRM. So the cohorts that are being compared are the email + SMS list vs the email – only list. In that scenario, the mobile+email list will usually do 3 times better.

It’s not surprising, right. The mobile phone number is a sign of interest. I mean when you’re single, you’re more likely to go out with someone if you get their number. This is a little like the business equivalent.

There’s one more thing to talk about quickly. I sold SMS capabilities for almost 8 years. When an organization would reach out they would always have one of these benefits in mind. So I’d get a call saying something like – we have this opportunity to partner with a rock back. We think it makes sense for the band to tell their fans to text in.

Or they would be thinking about the second superpower – we want to text people and get them to do X.

In order the campaign and organization to be successful both benefits, both superpowers need to be used. If you’re thinking of a great way to use SMS as a response to media, you still need a long term plan to activate the people that respond to the media.

If you have a great idea to text people and get them to purchase, or show up for an event, you also need an idea of how the person will join the list in the first place. That opt in is a requirement.

It’s really similar to the approach of email – an organization needs a way to grow their list and something to say to the list they’ve grown.

Lastly, I boil down the benefits of SMS to More People and More Action. In terms of direct marketing, not brand marketing, these two benefits are about all there is.

Elliott Golden founder of Civic Shout talks advocacy over Messenger

Elliott Golden joins the pod to talk about using Facebook Messenger to help people contact Congress. Civic Shout walks the user through a quick conversation and then sends a fax to the user’s legislators. 

Elliott created most if not all the technology and he helps clients promote their campaigns. So the conversation focuses on the details of building the interaction and strategies for promotion.

Hope you enjoy.

SMS terms you need to know before launch

We’re going to stick with SMS campaigns on today’s show and talk through the different terms, definitions and service providers in the space. This is a follow-on to the last episode where we talked through the different types of SMS campaigns. Now, we’re going deeper with broadcast messaging and mobile donations.

If you’re thinking about doing an SMS campaign or are new to the space, this background will be helpful. It’s a little dry if you’re just listening for fast-paced, thrilling messaging chat (that’s a joke). When I was selling SMS campaigns, these questions came up a lot. I hope it can be a reference to some people looking for answers.

Let’s jump in.

Long Codes and Short Codes: We touched on this last episode. If you’re doing SMS campaigns in any manner, you need a phone number to send and receive messages. As an organization or business you can choose between a long code and a short code.

A short code is a 5-6 or 6 digit number, that was created specifically to do high volume SMS campaigns. The carriers created this entire system to license short codes, approve campaigns and keep an eye on everything so their customers are not getting spammed. This is the way the MAN wants you to do it and most legitimate companies doing SMS will use a short code.

They are expensive and they take a little while to set up. The big advantages is that you can get more data around deliverability, send messages a lot faster and the organization has some protections about not being shut down.

There is one vendor for short codes in the US – CSCA – http://usshortcodes.com. You can go there to see if your number is available and reserve it or lease it. You can only lease a short code you can’t own it.

Withing short codes there are vanity numbers and random numbers.

Vanity means that the organization chooses the number and the lease is $1,000 per month. Random means that a number is randomly assigned and the cost is $500 per month to lease. There will usually be a setup fee and to launch a new short code it will take 8-12 weeks for the carriers to turn it on. I’m also going to bring up a shared short code. We’ll talk about this more when we get into the vendor breakdown. For most organizations the shared short code is the way to go. There is no cost and no wait time, but other organizations are using the short code.

A long code is another option for SMS campaigns. Long codes are meant more for the 1 on 1 campaigns or small businesses like a reminder from the dentist. Long codes are just phone numbers and then connect to APIs to send or receive SMS.
Long codes are cheap and fast. The downside is sending messages through them is really slow. With a short code you could send a million messages in an hour or two. With a long code it would just be impossible. So they are good for 1 on 1 messaging type stuff. So when your uber is arriving they would send a text from a longcode. They actually probably have thousands of long codes in different area codes so it seems, like it’s the driver texting.

On an upcoming show I’ll give all my opinions, but a big question from companies that want to do SMS – should we do a long code or short code? How about random or vanity? The short answer.. Big brands should get their own vanity code. If you have a special use case you might want your own random code. For instance clinical trials might require their own number, but they don’t care what it is. Most companies should pick a vendor that provides shared short code – my opinion.

If you’re a small business or doing 1 on 1 messaging, it’s OK to start with a longcode, but as you prove out the channel, think about leveling up to a short code.

Finally on this, if you’re going to do real world calls to action, get a short code. Billboards, flyers, events you don’t want to be telling people to message in, to a 10 digit number.

Let’s talk about the different players in the space. Obviously, we have the phone companies – Verizon ATT etc. They deliver the messages directly to the user. Below them there is this layer called an aggregator. We’ll dig in a bit there, but for an overview, they can be thought of like a hosting company for SMS. So they way you need to host a website, you can host your short code with the aggregator. Horrible terms, I know. Below them on the stack you’ll probably have a platform. This would be where you loging to manage campaigns. Finally there is the company that is creating the messages.

We don’t need to talk about the carriers other than there is an interesting insight that the short code system is completely separate from the P2P texting system. That’s a little surprising. They aren’t even related from the way that I understand it.

The aggregators are this middle-layer business. The carriers want to let businesses do SMS, but they are huge companies and networks – they don’t want to just open an API so that anyone can connect to them. They built this layer of businesses called aggregators. The aggregators connect to the carriers and then platforms or businesses connect to the aggregators. It’s actually 1000x more complicated than this.

No aggregator connects to all of the carriers directly, but they connect to at least 1 or more than one. For the carriers they don’t connect to, they have deals to trade traffic with the other aggregators. OK who cares.

The aggregators host the shortcodes. So if you lease a short code, you’ll want to provision it with the aggregator so that you can start sending messages. This is the setup that takes the 2-3 month process. In reality, you’ll only work directly with an aggregator if the messaging volume is large. An organization should be doing at least a half a million messages per month before an aggregator will even return your calls. If you do start a relationship with an aggregator, you’ll sign an agreement that includes a short code hosting cost, minimum payment per month that includes a specific message amount and a cost per message on top of that.

I wasn’t in the business then, but aggregators feel very 2005-ish. You need to be a big company to work with them and they are a part of the infrastructure. Once aggregators became established, people started building platforms to actually execute the campaigns.

A really dumbed down version of a platform or ASP would be something like mailchimp for SMS. A business could license mailchimp or they could build their own software and connect to their own email server. It’s the same with SMS, you could build your own software and connect to an aggregator, or you can sign up with a platform.

In my opinion the platform route has a lot of advantages. They’ll offer a lot of flexibility with short code options. They will help you get your own short code if you choose or they may offer a shared short code. A platform will have multiple customers and the platform can pay for the short code once, and then include short code use to their customers. It’s definitely a benefit.

Platforms range in price from almost free – $10 or $50 per month all the way to thousands or tens of thousands per month. There are some vertical specific platforms and more agnostic ones. Basically if you’re searching for a platform, you should set a budget and talk to a number of vendors, watch demo’s etc. I would recommend looking for

Shared Short code option
Setup, training and strategy help
Some level of connection or integration with other systems – like CRM or email
Keywords, webform and upload capabilities.

There are many more things that will become clearer as we discuss messaging strategy soon.

Just a few more types of vendors in this space. There is a company called Twilio. The best way to think of them might be a platform. Twilio is basically an aggregator, but they are very flexible and modern, where the aggregators are older feeling businesses. Twilio is a product created for developers to build apps for phone and SMS. They do short codes, but one of their major contributions has been introducing long codes as a flexible option.

The aggregators are for big businesses, and contracts. The platforms completely abstract the infrastructure. Twilio is for companies that want to build their own logic or platforms that connect to or send SMS. Again to use Uber as an example, they build into their system that when the driver is two minutes away, trigger a text. They could never use an ASP because it’s uber’s logic, they can’t code that into another system. They use the Twilio API to simply trigger a message.

Last week we talk about text donations. All that is really happening there is that there are just a few aggregators that can set up short codes in a certain way – to be premium SMS channels. The biggest such aggregator is the Mobile Giving Foundation. They have shared short codes or you can get your own. The MGF requires that organizations connect with a special type of ASP to help launch and manage mobile donation campaigns.

There are a lot of intricacies, but to sum everything up. If you are in product or operations you’ll probably want to check out Twilio to incorporate texting into your product. If you’re in marketing or a comparable field you’ll want to find a platform that can do what you want it to do.

Let’s talk about some terms that you’ll come across.

We talked about short codes and long codes. If you know what they are, you are way ahead of the game. Keywords are also important. A big value of SMS is that you can tell people to text in and they will. As an organization you have this one short code and you want to be able to source people as they are texting in. In a call to action you’ll say text KEYWORD to SHORTCODE – so text JOIN to 12345. This allows the ASP Application to know where the person heard to message in, and most ASPs can respond with a different message based on a different keyword.

Also, a keyword is needed to make the call to action make sense. Maybe it just caught on and it sounds right, but there should be something that we’re asking the user to send in. It’s weird to simply say text in to 12345 – telling people what to text where, just feels better.

MO – mobile originated. This is a message that is coming from a mobile phone.
MT – mobile terminated. This is a message being sent from a mobile phone.

Opt-in – there is a formally defined opt in process, but essentially opt in means the user is subscribed to messaging from this organization on this short code. A user has to give the opt in, in writing so that can be a text message in, or a webform – or I guess a paper form. It might be a violation, but it’s a regular practice that if someone says they want to opt in, that is fine too. Really they just need to agree to receive texts before they are texted.

Opt-out – There are some magic words required for SMS campaigns to unsubscribe people from the list. These words are STOP, CANCEL, END, UNSUBSCRIBE. The carriers require that when someone opts out, the organization needs to send a message confirming the opt out. This was actually part of a funny lawsuit a few years ago. I think someone sued Facebook because they opted out, but then Facebook sent a text confirming the opt out. Even though that’s what the carriers require. Facebook won the lawsuit.

Double Opt-in: When someone opts in, it’s the business’s responsibility to confirm it’s the person that owns the handset that is opting in. If the user texts in, that’s fine. We know that the text was sent from the phone. But if the number is entered into a website there is no connection to the phone. I could be entering an enemy’s phone number in every webform that I find. The double opt in serves as a check, to make sure that a confirmation is received from the phone that is opting in.

Generally, when the first opt in happens, the system will trigger a double opt in message that asks the user to reply YES to confirm the opt in. The user is in a pending state, not fully opted in until they confirm.

This double opt in can really hurt the opt in rates, so it’s in a gray area. Some people skip it, and are just sure to make it really clear that the user can opt out in the first messages that are sent to them.

Character Limit – texts are limited to 160 character and that’s including spaces. This can sometimes be confusing because tweets are 140 characters. The limit can be an issue. Messages have to be concise and texts cost money, so the answer isn’t always as easy as just sending 2 messages.

OK, that’s a lot of detail. Just thinking through all of this for the first time in a while I could probably add 100 things to the list, but this is enough info to be dangerous.

We’ll get more interesting on the next episode and I’m going to talk about the two biggest benefits to adding the SMS channel to an organization.