Speaker 0: Hey, everyone. Thanks for joining us on day two of. Uh, for those of y’all that are joining us for the first time today, welcome. We’re glad to have you. So, like, we’re waiting for our speaker, Alvar, to join on stage here. One minute for Alvar to join on here.

Speaker 1: Let’s see.

Speaker 0: There we go. Here he is. Hello. Alright. Again, welcome, everyone. Uh, this is Alvar Diaz with s four g consulting. He’s here to talk to you on the marketing cloud track today around that. So go ahead and take it away if you have your presentation ready.

Speaker 1: Hello, everyone. Uh, so first of all, thank you for the opportunity to share, um, my knowledge about anything, about marketing cloud, uh, about Salesforce. Uh, thank you for the audience to to come and to be interested in in this topic. Okay? Uh, we will speak about how we can use Einstein, uh, you know, the IAI of Salesforce to try to improve, uh, or go further or communications, uh, in in marketing cloud, uh, and most precisely in in the newsletter. Okay? Oh, do you see my screen? I hope so. We will start, like, uh, speaking a little bit about Einstein. Uh, we will be focused on the three most interesting tools of Einstein in marketing cloud. That is STO, frequency and and scoring because they are quite easy to use, and they can we can see quite a huge change, uh, in a short period of time. And then we will see how to use them in a case study, a real case study, a newsletter with our customer. Okay? Uh, so feel free to, uh, have any answer, you know, and I hope that you enjoy. Uh, I don’t know if I present myself, but, well, I am Albert Viet. I am from s four g, uh, McKinsey company, And I am not able to and as long as I can I can share my the information to to you and to try to have a good time and let’s, uh, learn altogether, it will be a a good presentation?

Speaker 1: So first of all, uh, what is Einstein? Uh, probably most of you know what is Einstein, but just to clarify, uh, that Einstein is the artificial intelligence of Salesforce. Uh, it has it is in in in every cloud. And in marketing cloud, uh, it’s very useful because it is help us to to have insights of our data, to know better our clients, and also to give them the correct content at a good time. Uh, and we will speak about this, uh, in in the next slides. But it is very interesting to to have a good engagement and to personalize our communications. So, uh, with this kind of tool, we can work less and have more results.

Speaker 1: As I told you before, I will be focused in the three main tools. Uh, one of them is instant STO. STO comes from a send time optimization, and it help us to predict what is the best time to send a communication to our clients. Uh, when I say that, I mean, our clients, but every client. Okay? Uh, in fact, it, uh, gather data for more than 20 factors and then, uh, using machine learning, Einstein assigns, uh, the wave of each factor according to the influence of predict the future behavior. So, uh, it needs, uh, like, around two or three months to have data. But then, uh, Einstein will help us to know what is the best time for every con for every contact, uh, according for a period of time. Uh, we can use Einstein in in two different ways. Uh, as you can see in the in the in the right is the dashboard. Okay? We will have, like, a quick view of, uh, how is our our database. We can know, uh, what is the best day of the of the week to send or the best, uh, hours. But this is the general information. Okay? Uh, every contact has his own information, and Einstein will send the communication to each contact in the best time. Uh, as I told you before, we need a period of time and sometimes we need to give, uh, time to do the AI to to know what is the best information. Then move to the left. We can see how to implement it. Uh, we can use it in Jammy builder. Okay. We select the activity, and then we can select the period of time that we want to to send it. Uh, as you can see, this is from four hours to almost seven days. And if we select for some of 24, it’s one day. Uh, in this period of time, uh, a sting will, uh, send the communication to each contact. Okay? Not it is individual, uh, measure. If we don’t have enough data about the client, it will be sending, uh, an hour. So it will gather more data, and then we’ll we’ll know what is the best, uh, moment to to send the communication.

Speaker 1: We have also instant frequency. Instant frequency is a powerful tool to know how saturated is your database. Okay? Uh, it evaluates what is the optimal number to of of, uh, messages to read to our data, and then it calculates if it’s unsaturated, on target and saturated. Okay? Uh, as you can see there, uh, it has metrics. It, uh, have metric of every, we will say, every month, every twenty four twenty eight twenty eight days. And then, uh, with the machine learning, we’re able to know if our database is, uh, saturated or not. Okay? Uh, as you can see on the top of the of the picture, uh, is the dashboard. Uh, we will be able to see, uh, like, we will say general information, uh, if our data is saturated, on target, or unsaturated. But, uh, this data is not so, so, so real. Okay? Uh, we have many clients and we have tasted, and it’s okay. But if you want to to know what’s the real data, we will need to use journey builder, um, and to gather information about that. As you can see in the bottom, uh, we can, uh, active this tool in dummy builder, uh, using activity. And we will select what path we want to to create. There is no four there is no three like in the dashboard. There is four. Okay? Because we have one more that is on target. On target is the people that now is we will say, uh, in a good target to to be sent. But if we will send more message to them, we will saturate it. We can, uh, select if we want one, three, or five message to be saturated. Uh, normally, we will put three test the the normal configuration. Uh, and it’s, uh, quite a strange, but this, uh, this kind of audience, uh, target on target and almost saturated normally, uh, have, uh, the best options. Saturated has also a good engagement, but, uh, hit all, they also has they also have a a high percentage of a subscribe. So we will need to know how is our database and to send to saturated only when we really want it with, uh, communication that are quite, uh, interesting. Okay?

Speaker 1: And let’s move to the last one. Last but not least, we have extended scoring. Probably, this is my favorite, uh, tool of Einstein in Marketing Cloud because it’s so easy to to to configure. But we have all we have a lot of information, uh, in words. Okay? Instant scoring have, uh, many ways to to play the consumer engagement. As you can see in the in the left on the top, uh, this is how we can configure it, uh, in marketing cloud. In Jeremy builder, we select the activity, and we can select many kind of file. We can select for open. That is not a good device because, as you know, open is no sour puree nowadays, clicks, unsubscribes, or even, uh, the conversion in the world. Uh, but the most interesting part of there is persona. Persona is, we would say, like, a person that try to do an engagement and the it select I mean, segment the the people in four different groups that you can see there. Loyalist, that is the people that love us. Uh, it, uh, click on our message. It opened them. Uh, so they are the people that we need to care. Uh, we have, uh, win back on the bottom. That is the people that don’t open our message, uh, don’t click on our message. Uh, they’re almost, uh, sleep. Okay? And we have in the middle two kind of groups that is selected, uh, subscribers. Uh, that is the subscribers that, uh, normally, they open our emails, but, uh, they’re not clicking on them. And Windows Server, that are the people that don’t open our emails, but when they open our emails, they don’t, uh, they have to reach. I don’t know if I missed it, but, uh, one of them is they don’t open, but they click it. And another one is they click uh, they don’t open, but when they open, uh, they click it. Okay? Um, so there are four different kind of roofs. Normally, when we try to when we have tested in our clients, we see that between selective and Windows Server, there’s not too much different. It’s quite close to the to the average, uh, of the of the database. So sometimes, uh, we we gather both groups in just one. But on the other hand, loyalist and Uber are groups that are very strong, very, uh, different from the other ones, and very useful. Uh, as you can see on the bottom, okay, we have, uh, information that we can see in the dashboard, and we can see how is our database. Normally, of course, we have more people that than the loyalists, but it’s good to know how it’s moving. Okay?

Speaker 1: So I think that, uh, in this five, ten minutes, we have enough information to know what is Einstein. Einstein, uh, scoring, Einstein frequency, and Einstein STO. Now we know what is the theory, but let’s go to see what is the the real world. Okay? So I would like you to to send a case study, a real one, uh, that we have. It was a client. It was a retail company, quite huge, quite large, uh, amount of people, but they have a very low engagement. We see, uh, how was the communication, how was the the data, and we see that they have huge saturation of the people because they have some departments. And every department, uh, sends, uh, his own communications. Uh, so, uh, they don’t care about the other departments, and the user experience was awful because the people received three, four, five messages in one day with different promotions, and they didn’t know what was the important one. So, uh, for example, on this point, we think that it was interesting to use Einstein, uh, to try to help us. They have also problems with the personalization. They use the same generic email for everyone, and they only use a email. Despite they have an app, they only use email to to send communications. Uh, they also, uh, don’t coordinate. They they don’t have a a global vision about what was the communication. Uh, so the result was that the people, the database was born, and they don’t have quite, uh, they don’t have engagement. So they need something. Okay?

Speaker 1: Um, on this point, we we say, okay. It’s going to so use Einstein to try to improve this point. Okay? And we just Einstein. Uh, as I told you before, uh, we think that, uh, we thought that the instant frequency, extended scoring, and extended still, uh, will will help us to to try to to improve this this lack of engagement. Uh, first of all, the most important thing was to let them know that they have a saturated database and try to, uh, make conscious that, uh, they have to change how they were doing the the communications. So, uh, we established a communication governance to to try to send communications, uh, each department one day and depends of the saturation of the contacts. Uh, we will send one or more communications. And in this point, for example, instant frequency was very useful to know how was the data and to to be able to to see to show, uh, to the client, uh, that they have a problem. Okay? Because many times, they didn’t know it. Um, it is interesting to to have the the analytics to to try to solve them. Uh, on the other hand, uh, we also, uh, try to don’t, uh, send to win back, uh, if it’s if it wasn’t necessary, uh, because it’s people that they will not care about you. So just go to don’t disturb it until it will be necessary. We also try to make new channels, try to, uh, personalize the communications, uh, using dynamic, uh, parameters. As I told you before, they don’t just stay up just for fortune. For, uh, we will say, ordinary communications, and we will use it, uh, for a newsletter. Uh, they’re so we’re quite good because the, uh, the contacts were not, uh, used to have this information, and we don’t disturb them. So they have a good engagement. And, uh, we thought that dynamic parameters will be so useful to to make communication more, uh, more powerful uh, for them. Uh, for example, if we have a communication that is connected, uh, with, uh, pets and I have a dog, if you put a picture of a dog in in the newsletter, I will be more happy that, uh, you put a cat or another animal that I don’t have in my house. Um, and we don’t only say what we should do. Okay? We have, um, metrics. We gather all the metrics, and we have done all these each way every every week, uh, to know how how was it going. Okay? Um, because sometimes you you think that the one solution will be great and then you realize that it wasn’t. And on the other hand, you have also the this this thing. You, for example, can think that by almost a rated people will not, uh, have a good engagement and sometimes it’s the best group. So on this point, it was very important just not to have a photo, if not a video of all the database and how was the adaptation, how was the the things that we were proving if they were right or not. In the right thing of the of the screen, you can see a picture of, uh, journey builder with all the all the information that I I shared with you. As you can see, there there are many instant activities, and every activity has one one ideal goal. Okay? For example, we use a stainless steel for the communications, and it help us to to increase to decimal points or or three decimal points the the open stack. It’s not too much, but it help us if we have million of of, uh, contacts.

Speaker 1: So, uh, with all of this information, I mean, and the first information that we have about the client and the solutions that we share with them, uh, we, uh, work with them for six months, and the results are not the ones. Uh, as you can see, uh, they begin to be conscious about what was their problem. They know that they have problem with saturation. So they start to, uh, sending less emails to the people that they don’t want it, and then start sending another information that it was really important. So, uh, the saturated contacts decreased from 20% to just 4.5. Uh, I think that it was a great success. Uh, as you can see in the picture of the bottom, uh, that was the the evolution. It’s always, uh, not ideal graphic, but it was, uh, the tendencies is quite clear. Uh, and the engagement was also clear because, uh, the opens, uh, increased also from 14 to forty twenty two, and the clicks, uh, was more than the double. Okay? We have 4.5 in the beginning, and we were able to go for 1.1. Uh, there’s no more information that was you are looking here. I mean, uh, we were able to make interesting communication to the the contacts, sending only the communication, the relevant content to them, and we have a best better engagement. Uh, we don’t have all of us only, um, I mean, we’ll say good analytics. I mean, the the the team was real, and they will they they changed their their point of view, uh, and they grew up the the engagement of of the data, uh, because they they practice the governance of communications and the personalization and the channel. And they were able to see, as you can see in the screen, how was the the real the real stage of of the data. Um, and just to sum up, uh, normally, we we propose, uh, using newsletter for for instant because newsletter is the wide communication. And with the wide communication, you can know more about your data. If we use just a little once, a small one, we only have, like, a little part of our database. Using instant, we recommend to to to use a wide data to know more about them and to be able to see how they are changing and how, uh, can instant help. Uh, finally, just to say thank you for for your attention. Um, as I told you before, uh, machine learning is very important. It help us, uh, but we need also people behind that to to think and to discuss what we can use it and how can use Einstein to to get better results. Uh, I think we have three, four, five minutes. So just to say thank you for for all of you. You have my contact there. And if you have any questions or any information that you want to share, I’m more than open to to listen to them.

Speaker 0: Thanks. Thanks. It’s Alvaro. We we had, um, one question in the chat from Donna. She was asking, is Einstein data at the BU level at the business unit there?

Speaker 1: Sorry. Can you repeat it?

Speaker 0: Um, from Donna, is the Einstein data at the BU level?

Speaker 1: Uh, Yeah. Yeah. It’s it’s, uh, it’s at the BU level. You can, well, you can configure it in in the setup. And then it needs, like, two, three months to have, uh, we will say, accurate results.

Speaker 0: Awesome. Alright. Another question from Tim. Um, if a subscriber does not have enough data yet for STO, will they receive the email as part of the journey, or will they not receive it due to lack of engagement data?

Speaker 1: Uh, they will receive the email, but, Einstein, we’ll send it randomly. Okay? Okay. And same yeah. Same, for example, uh, scoring. Normally, when you oh, no. Scrolling. Sorry. Frequency. When you don’t have information, uh, instance in the dashboard, they say that is, uh, under saturated. And, normally, in the Jammy builder, they they they make the segmentation in a random way, uh, that you don’t know, and maybe, uh, they go for a random path. But, uh, yeah, STO is is random. And you are also able to to make a trick that is, uh, you can send step by step every hour. You can make a small group of of, uh, contacts, and they can send it, like, in in this period of time, uh, all the contacts.

Speaker 0: Okay. Um, one question from Fernando. We have about a minute left. Um, how does Einstein make the assumptions? Is that related to a hidden pixel?

Speaker 1: Yeah. Um, the problem of testing is that we don’t I would I mean, it’s so easy to to use, but we are not able to, uh, control all the faster stat, uh, they use. Okay? Uh, so sometimes, uh, we are able to configure two or three things, but the other way is that, uh, hide them by by servers. K.

Speaker 0: And I think Andrew’s question here was similar. Can we configure what’s the positive versus a negative influence on the Einstein score?

Speaker 1: Yeah. I’m I’m looking the the questions. Uh, Einstein hasn’t got any any hidden pixel. It’s like a, uh, artificial intelligence in in the platform. So they have information about all the faster stat has in the in the platform. And we are not able I mean, I know that this question is is is close because, uh, for example, we we can use it with, uh, social media, but, uh, no. We are not able to see if, uh, it’s negative or or positive, the difference of instance call. We can I mean, instance call is is so useful because we can say, uh, what is the group? We are not able to say if it’s negative or or positive.

Speaker 0: Awesome. Well, thank you so much, Alvar. We’re out of time now, but anyone, uh, everyone who’s attended, thank you so much for joining the session. Please feel free to reach out to Alvar in the meeting chat or on LinkedIn. Really appreciate all of you. Shout out to our sponsors, and then we also have, um, marketing cloud product road map coming up in a few minutes. So thank you, everyone.

Speaker 1: Thank you, Paul. Bye. Bye, guys. For for your time. Thank you for for select my my my information. Thank you for giving me the opportunity. Yeah. It it was a good, uh, I will I will say I would, uh, yeah, a good, um, point. And, um, well, I I hope that it was good. I it’s good work. I don’t see many.