Episode 15 | Transcript

Arthur Backouche: Artificial Intelligence, AMPScript, SSJS vs SQL

Anthony Lamot: Hi, and welcome to Heroes of Marketing Cloud. The show where we talk with Salesforce Marketing Cloud experts. My name is Anthony Lamot. I’m the CEO and co-founder of DESelect. And I’ll be your host for today. And today, we’ll be talking with Arthur Backouche. Archer is a senior Salesforce Marketing Cloud consultant at Skie and we have a very engaging conversation about AI and all things technical in Marketing Cloud: SQL, AMPScript and also SSGS as well as other trends going on in the marketing automation space. So, welcome to the show. Hi, Arthur, welcome to the show.

Arthur Backouche: Hi, Anthony. How are you?

AL: I’m doing very well, opposite sides of the world that I appreciate you making time on an early Saturday morning for you.

AB: Yeah, it’s great to meet you. Actually, your name seems to be French. I was wondering, are you from France and a lot of people from your team are in Switzerland? 

AL: Yeah, no, that’s a great question. So, I’m a, I’m not French although I speak some French or at least I like to think that I speak some French but I’m Belgian. 

AB: That’s nice, it must be great to live in Texas.

AL: Right. And I know you’re based in Sydney, but you don’t particularly have an Australian accent. 

AB: Yeah. I arrived in Sydney like five years ago now. Yeah, it’s really good. I think so far it’s like one of the best city I’ve lived in like you have like the outdoor activity and the lifestyle is really great, but I guess Texas must be nice as well. This is where there is a Tesla headquarters as well.

AL: Yeah, that’s right. Elon lives about 30 minutes from where I’m currently sitting.

AB: Really? Do, do you grab your coffee in the same cafe sometimes?

AL: Well, or we’re not grabbing drinks, we’re not on the first meeting basis just yet, but who knows, he likes to hang out with podcasters. 

AB: That’s that’s fine. Yeah, but yeah, no, winter is amazing here. I mean, Australia gets hot too, but right now, it’s been over 40 degrees Celsius over 100 Fahrenheit for like two months, but I love it. I’m glad to, you know, be in this hot climate. How did you come up to choose Texas?

AL: That’s that’s a long story… and I want to go into too much detail now because maybe, you know, some people hear more about Marketing Cloud. But the majority of our market is in the U.S.. We serve large enterprise companies, and mid-market, there’s a huge market for us in Europe. There’s also a huge market for us in Australia and APAC, the majority of big companies obviously are in the us. So we want to open office here and Austin is a very vibrant city with a lot of new tech talent. It’s centrally located and it has a, yeah, you know, it’s just a great city. So that’s why we chose Austin, Texas specifically.

AB: Okay. That’s nice. And, and you started your journey with Marketing Cloud. I hear that you went on holiday for a few weeks to learn how to code in order to build DESelect. Yeah. I mean, like maybe you read that, in a blog or something. That’s really a great story.

AL: I appreciate that. Yeah, kind of. So at some point I wanted to get better programming. So I took a sabbatical went to Bali lifter for three months while I rented my home back in Belgium out on Airbnb just to cover for costs and… I won’t say that lead directly, to say, but at least I had to technological know-how, to see a solution to think of a solution. So I don’t regret it. Yeah.

AB: Yeah. And were you a user of Marketing Cloud before?

AL: Yeah, I was already doing Marketing Cloud projects for. Yeah. And what about your background Arthur? Yeah. What is it like? You can see a bit about what your background is? And, and what is that you do these days?

AB: Yeah. So I used to, I used to study computer science. So I did a school called Epitech and the it’s like school where you only learn by practice. You know, there is no theory.

So like you have it’s like it’s exactly like working in like an agency. So you have always project to deliver with group of people, and you have no theory knowledge. So you no the record knowledge. So it’s like you deliver project. So it was great to learn that. But I didn’t really like language like, deep technology language. I was more interested in digital marketing or to drive customer to drive conversion to a website to optimize a website. So I started to do a lot of like in township like and work in digital marketing at first… and that, so I’ve learned like the digital marketing in general. And I worked like maybe five, six years as a digital marketer. And then when I came back to Australia, I realized that like working as a digital market, digital marketing is not great. But I don’t have like a good visa. For example, you don’t have like good salary as well. Like because there is no benchmark or digital marketer, you don’t know if like the, this guy is better than that guy. Like there is no one, to say like where with Salesforce certification, like you have an industry and you have like a benchmark between like the different consultants I guess. And then yeah, I started learning like about it about Salesforce more specifically because I already had like this it background. And yes. So I’ve learned about Salesforce. And then I started working as a consultant and it grew because Marketing Cloud specifically requires you to know about like it like you need to code in like SQL, JavaScript, and Script language. And also you need to have the knowledge into digital marketing such as like having conversations that make sense with the customer for example on what they can achieve, what are like the nurturing journey with at the segment. And I think it was the best of both worlds for me because like I had like there’s two… segments there’s two addons that I’ve learned over my life. And then I end up working as a self, as a Marketing Cloud consultant. 

AL: It sounds like a good, yeah, it sounds like a good marriage between your it background and your marketing agency background. I think just as an aside, I think it’s interesting what you said about how salaries for marketing clock consultants are better in, for agencies. I think scarcity in the job market is perhaps a part of it, but you’re probably right that certification helps because with an agency, it’s very hard to have objective measures, of the work, I suppose.

AB: Yeah. And, and I think from what I saw is like if you work as a generic digital marketer, like there is plenty of digital marketer in the world, I guess. And, and it’s more like the like if you work in marketing, you are the one searching for a job where if you are a consultant in Marketing Cloud, the company are searching for you as a skip. So there is like it’s totally different like, the balance between the supplier and the yeah.

AL: Exactly. So what’s the coolest marketing clap project you’ve done so far Arthur?

AB: I’ve worked on like plenty of interesting projects. I think like project that I like is where there is technical challenge with the platform. I think one that was quite interesting that we just delivered is an integration with hardware, for the BBC. So they created like a photo booth system. So, you know, like where people take photos and, we integrate it with Marketing Cloud. So for that, we are using like we are using different API, for example, content builder API to re, like, the files and the files were like MP four and like give you so, you know, it’s like more it’s quite interesting to learn how to encode like a video, for example, and pass it to content builder. And then we were sending like a journey, we were triggering a journey that sent an e-mail that includes this content to the user. So, so, so this one was maybe not the most complex one but as there is an integration to aware and to a product that is that you can touch or that you can interact with in the real life, I think it was pretty cool. I think.

AL: Yeah, no, for sure. I mean, I’ve had a number of great guests on the show and I think you’re the first person ever gave an example, of someone integrating between hardware and Marketing Cloud. Could you, maybe we have a mixed audience of technical and nontechnical users? So with that in mind, could you highlight some of, the challenges you ran into or how you went about that solution? Because I think it’s kind of interesting.

AB: Yes. So, so I think like… so there is a different type of challenge. There is a technical challenge and there is a process like, we had challenge with the process and I think something that we really improve is like the QA testing. So, so, so we work with someone of our team that runs the test properly, like, to see if there is any bug in the solution. And I think this, we add some challenges at the beginning where, we were delivering like the first… had some issue and someone of our team called Christian run like the QA test and find out there is some issue for example, with… you know, like when you deliver like promo code because when, we are sending the e-mail to people, we were giving some promo code like discount code, 20 percent of 100 percent of. And what we realize is that there is an issue when Johnny is delivering, the discount code. Sometimes I will deliver the same code, to different people even though like you have like a system that should make them deliver only one code per person.

AL: All right. So it sounds like it was like randomly generated but maybe not unique. Is that what you’re saying?

AB: No, they were stored into data extension, but still like I think like when the journey delivers the e-mail, it goes too fast, and sometimes like the journey doesn’t have the capability to say, okay, this concord has been already delivered now. I’m going to the next one now, I’m going to the next one. So we had to use a function for that. I don’t remember the name of the function, but one part of the function was raised, and it basically… stop not the system but switch to another contact in the system. If there is a, if this happened. So, so, so we find out the solution. Yeah.

AL: Yeah. So I just wanted to actually know a little bit about the race error thing and I know it’s going pretty technical here for the audience, but there’s a function call race error that if my understanding is correct, it can just cancel a specific e-mail being sent for whatever reason. It’s like a catch-all or if something goes wrong.

Don't miss an update

AB: Yeah. And you have two choices basically. So you have like either which I can sell the overall process. So no contact will receive any e-mail anymore or it’s just one contact that stopped receiving the e-mail and then they continue to switch to the other one in the process, right? Yeah. So this was one of the challenges that we discovered, we fixed it. So it was good… and we also improve a lot the way we work as a team with like the process like the testing part is very important, and I think it’s something that you learn and that is different with each team. So, so, so, you know, it’s not like if you are like a JavaScript developer and, you have like a proper framework that exists, I feel it’s up to each organization to build their own testing framework with Marketing Cloud. And, and I think we have learned a lot to build out in order to deliver products that are perfect for the customer.

AL: Cool. Yeah. I know as many people have actually commented on the show that quality assurance, and not having an actual sandbox is a typical challenge. So it sounds like an interesting way, to deal with that. Another functionality that’s been more discussed recently and obviously we run into as we have launched a new product call Engage is frequency capping. It was kinda curious if you’ve ever worked with Einstein frequency capping, if you’ve seen use cases, and how it turns out.

AB: Yes. So, I think frequency is very interesting in terms of what we call audience fatigue. So, it’s like you make sure you don’t deliver too much content to your contact in order to keep him like interested in your product, like you don’t struggle with like all your e-mail… if we use it. Yes, I think we use it, but differently, like for example, at this stage, we set up some fields… in Sales Cloud for example. And when we pull like the Salesforce data, we will do a filter on this particular filled in Sales Cloud just to make sure that we are not sending like… before 30 days, for example, like, the e-mail so, we haven’t used in a lot of projects like Einstein frequency cap. But I think I have a good understanding of what it is, and how we can… use it. You, from your point of view, why people are using Engage like your new product?

AL: Yeah, a great question. Thank you. Well, first off, there’s limitations with Einstein frequency capping. It’s a cool feature because it’s out of the box and you don’t need to do a lot to set up, but you need to be us, you need, to use the out of the box subscription model. And very few customers really use that because almost everyone ends up building a custom subscription center or something like that. And, and once you do that, you already can’t use it anymore. But I would say even, if you have that, there’s another issue with it, which is just a comfort that marketers have because it’s a black box. They don’t know, if Einstein says, you know, this person is saturated or not are not gonna send an e-mail or not. It doesn’t say why that is. So… as we spoke with our customers, and more than the context of DESelect Segment, it was require need a requirement for frequency capping, that is rule based, meaning marketers can define the rule. So I would say, for Engage to actually answer your questions. I think the main reasons why people get engaged are they need rule-based frequency campaign. They, they’re not only concerned about over-engagement but also under-engagement. So they also want to see, hey, who is my audience? Can I engage more with? And then, and this is also interesting. Actually, I didn’t think the product would turn out to be that way initially, but we ended up building this whole calendar where people can manage all their sense, all their campaigns if you prefer. Yeah. And then they can also manage priorities of campaigns in case our conflicts, like, if we can predict an audience is going to be over-saturated or, you know, they will hit their frequency cap.

AB: Yes. The way you describe Engage like your new product makes me think about the Einstein recipe in Interaction Studio. Have you heard about that? It’s like basically, it’s like, yeah, it’s like from what I understand is like Engage, you can have different ingredients like you can have different settings in order to build your own… what you define as a good frequency in a way?

AL: Right. But, okay. So you’re referring to Marketing Cloud Interaction Studio formerly known as Evergage.

AB: Yeah. And now it’s Personalization.

AL: Now, it’s Personalization, right? Yes. Thank you for reminding me. Oops, it’s different though, but here’s a problem with the constant renaming of the products. If, if we can be very candid and I feel on this show we can… the phrase “Marketing Cloud personalization” is a little bit misleading because you really have to ask what kind of personalization and ultimately Marketing Cloud personalization is really a digital solution, meaning it’s meant to personalize websites. Yes, there is a way to integrate with Marketing Cloud engagement. These a, so, the good all real Marketing Cloud. So I target and so, you could technically steer e-mail communications as well. But the driver of that is normally your digital interactions, the digital activities on your website, which there’s a whole subject in itself and effect is many people don’t have these two solutions. I mean, the companies or they’re not always completely integrated because they’re actually using different stacks. So… that’s not to say that there aren’t really interesting use cases in terms of driving e-mail personalization, by web interaction. I totally think there are it’s just that would not really be a solution for the problem we encounter with our customers who really just, yeah because they’re just our customers were just really worried about not being annoying and not spamming their customer.

AB: Yeah, yeah. I think it makes total sense. I think it’s a people need that and I think people want to see like a transparent solution where they understand each parameter of it like they want to know what will happen. And maybe that’s an issue that is happening with the current like feature of Marketing Cloud in frequency. And, and I think this can link us to the other question that is related to artificial intelligence, what’s the future of artificial intelligence with Marketing Cloud? And I think people really want control and transparency on what will happen. And I think this is a risk or like the part, but like it will be a challenge with generative artificial intelligence. I don’t think people want yet like a machine to write the e-mail copywriting that will be sent to the customer, they want to… they want to control that part mostly like the big organization because they are already like perfection is on the message that they deliver to their customer. And I think as well, what will be interesting with artificial intelligence? It will be on the insight and insights… and analytics. I think that’s the main area of development that I think will help. How, how about you, how do you see… artificial intelligence coming into Marketing Cloud?

AL: Yeah, great question. I mean, I, we’re gonna be able to talk a lot about this. I will take a step back first before even talking about Marketing Cloud specifically. We like we totally embraced it as a company. I personally use it. It’s like one of my new bookmarks. I have the paid version of GPT, and we actively encourage people in the company to use it for other lines of business. And we have a marketing team, right? So if I look at my own marketing team, we are already using it heavily for content creation. And if I work for instance, at a more strategic level, but whether it’s company’s strategy or… marketing strategy specifically, it’s a really interesting tool just for brainstorming. I find structure on your thoughts. I don’t know if you’ve tried that before.

AB: I do use it. I never use it to brainstorm… but yeah, I do use it like not every day, but, I think, yeah, like it saves me time, and I think he, you know, like I would be scared, if I was like a company such as Grammarly, you know, Grammarly, like the Chrome extension that fix your typo. I think like I tend to prefer using ChatGPT and like Grammarly for example.

AL: I don’t know. So this is, you say is because if I actually recently used ChatGPT, to help me write an article? Plus, I was plus I was doing, I was also doing an interesting experiment, on LinkedIn because I’m a very voracious poster on LinkedIn, people who follow the show, they probably know. And so I did a series of experiments where I had ChatGPT, right? LinkedIn posts. And the big thing I noticed is that one, that the posts are much more generic, but I can tell a little bit more about my strategy and I’ll make it less generic. But secondly, a Grammarly still finds grammatical suggestions in what ChatGPT says.

AB: Really? Yeah. And that’s why do you think?

AL: Because I don’t think GPT is optimized for having beautiful grammar, it’s optimized for giving synthesis, like for summarizing large quantities of natural language information and then providing suggestions while also tapping into the whole internet as a database essentially 

AB: Yeah. That’s true. Yeah. And, and I think Grammarly will always provide different options maybe as like there is not one way of writing a sentence correctly, sure. So I don’t know. But yeah, I think it’s a great to like ChatGPT is nice. I think I’m quite excited about like I seen they started to develop some apps like you can integrate with external system. I think this would be like, I haven’t tried. I don’t know if it’s available on the market, but it’s basically, you can pull the data from other system. It’s it’s a bit like see it. Okay.

AL: I mean, we are doing this for our product… actually just this week. So today, I mean for those who are listening, it’s August 18, 2023. Just five days ago, we went live with DeeDeeAI. 

And so, we called our generative AI capabilities DeeDeeAI a little bit similar to Einstein AI, I suppose. And what it does is that this is for our Segment product. So, for those who don’t know Segment is essentially a drag-and-drop solution to great SQL in Marketing Cloud to do very refined Segmentation duplication whatever. And so now people can just say, hey, did I want this segment from, you know, my last orders in Australia and France with volume of this and a value of this. And, and based on the natural language, we can translate that into a configured segment in DESelect which is not just SQL but actually for people who don’t know SQL, it’s actually a visual representation in our UC. And, and then people can still iterate and that’s just step one. We have a lot more ideas there to go further.

AB: You are using a ChatGPT API for that?

AL: Yeah, that’s right. So this is, we’re integrating with OpenAI for that. So, which is it’s a paid service, right? Because I have a solution for companies and we’re using some, we have integrate with some other APIs essentially. But then on our site, we need to add a layer to, you know, train the model, and help them understand, okay, this is how we, you know, this is what our data looks like and how we visualize it. And I mean, I need to add some guard rails, and stuff like that. So yeah, we do that.

AB: Is there like any plan to… get rid of OpenAI in the future to build your own fully full system?

AL: No, and yes. So, no, we don’t plan to get rid of OpenAI. And, and here’s a thing like, so as a side note, but an important one, I think… integrating generative AI capabilities into your product will typically rely on one of the big providers which for the time being will mainly be OpenAI, but we’re gonna see other providers as well. Obviously, the other tech companies are trying to catch up real fast and they will probably, I mean not catch up but they will deliver good solutions too. But what I’m trying to say is that this will become ubiquitous. And what I mean with that is like today, if you would start assessed company, gonna do it in a Cloud, right? Unless you maybe some very fringe niche financial on-premise product for security, which also happens but, you know, except for that, like very rare situation you would by default do in the Cloud. So, my prediction is that maybe a year from now every new SaaS company by default will have generative AI capabilities integrated into its solution because customers will expect it because they will expect that level of performance.

AB: Yeah. Maybe I saw a lot of… a lot of… yeah, most of the us now are using already like… some artificial intelligence integration I think is not all of them are really well integrated. I feel like sometimes to me, it doesn’t make sense that like it’s more here to, for the sake of being here. I think it’s really interesting to find what’s the usage that really provide value to the customer. And that’s a challenge because in some cases, in some cases like I think it’s still early to find like the real value of it. But, but yeah, like that would be interesting. Like, I think like people will enjoy like a lot of… people that are not technical or even people that are technical could enjoy using like a DeeDeeAI, I think.

AL: Well, I mean, if I can lift a little bit more of the… (spoiler alert), but so I mean we, we’re getting really positive feedback, on the initial feature. So the initial feature I described is essentially what we call copilot, right? You tell DeeDee like, hey, this is what I want to do and it’s segments for you, right? So that’s copilot. So that’s already live with a few customers, but we have already have a big waiting list and if people who are interested go over to DeeDeeAI and you can sign up for it. But, but we’re gonna go further and the next main capability that we’re already looking into. And then by the way, we also need to talk about just because we sidetracked on the ChatGPT. I just remember, we do talk about native ML as well. But, you know, just to stick with you, the other thing I wanna do is… what we call tribal knowledge. So, so Arthur, you know, you’ve been, you know, working with customers, you know, how in every organization, there’s just one guy or girl who knows the data, right? And whenever someone needs to create a segment or isn’t sure about a filter, they go to Frank or Suzie’s desk and they go, hey, Frank or Suzie, what filter should I use? What, what criteria should I use? And so, typically, this person, this kind of person is very proud of their work but it’s also not ideal for them that people come and pull their sleeve every five minutes, right? So, so there’s a bit of annoyance there from the point of view from the organization. However there’s like a massive risk if Frank or Suzie gets hit by a bus tomorrow, like who will know the data? So here’s, the idea for the tribal knowledge. The idea is that as people are segmenting, in DESelect with did or not like even manually, we quietly listen, and try to figure out or rather quietly tries to figure out how are people segmenting and why? So that when a new person joins that, they can just say, hey, did you know, I’m new in this company, tell me how people segment and it will just spit out all the typical segments. And then you can go build me one of these segments. Yeah. And I put, it will do that. But the, and then I think that here is a really important ones in a really interesting use case. You will be able to ask, can you tell me something about that filter? Why did you use this value in not the order? And that’s where it gets really interesting because now Frank and I, are certainly scaled, right?

AB: Yeah, I think that’s a good idea. I think we need that we also need like things to write documentation easily. Have you, have you heard about Scribe? So it’s…

AL: I think I heard of, I haven’t used to it.

AB: I haven’t used it neither, but it’s like something that basically as long as you create the process like let’s say you set up your segment in Marketing Cloud, Scribe will take some screenshots and they will build the documentation for you automatically, which is like very impressive. I think this is like the type of artificial intelligence usage that are very interesting compared to generating a small text on like another platform like for example, I saw Elementor they propose like something, do you know, element say propose of generative? Yeah, I mean you an, of, this platform… but like the way they integrated artificial intelligence is not that great like things. They just propose you to generate, the text, on the text widget. So, you know, like it depends, of the company. But yeah, it’s interesting.

AL: Yeah. In fact, I think we still use Elementor for our own website but, I might be wrong. It’s been a while since I had to go into the Admin Panel myself, but about Scribe, I mean, I just have a quick look at the website. Looks super interesting. I mean, look, I don’t know the product. I’m not endorsing this, is on affiliate marketing thing. I will say, I really believe in the value of good documentation. I think it’s one of those things. It’s interesting. Like one of the last… guests on the show was, Genna Matson from HowtoSFMC. And she also like underlying the value of good documentation. Now, obviously, we were talking specifically about Marketing Cloud. I think this is important for everything. In fact, just before I swear, just before this meeting, I spent the last two hours updating some of our own internal documentation which is like very specific business knowledge. I will make sure it’s like downloaded from my brain and available for the whole company. So, I mean, yeah, seeing something like this, is pretty cool. So thank you for the recommendation.

AB: And, there is two things. There is the documentation. And there, I don’t know how you call that in English but like in French, we call that, the “norme”. So it’s like structure of, yeah, it’s like the norm of your code. The way you write code is also very important, I think because…

AL: Or like a convention, maybe “coding convention”?

AB: Yeah, exactly. Code convention. I think this is so important because… like a code that is well written in Marketing Cloud, it’s so much easier to understand. So, I like this is something that we are working a lot in our team, to make sure that people write, the correct way. And also like there is some choice to be made to be made because for example, you can choose between AMPScript and SSGS… and some people will write like a full page in AMPScript when sometimes you should using be using more like SSGS for like a cloud page and AMPScript for small script only. So, so this is like an open discussion. What’s your view on that? Like what do you think? Do you have like a coding convention, for example, a DESelect and how do people make choice when they choose, the language to write?

AL: Cool, man. Let me interrupt that one. So, yes, the short answer is yes, of course, we have code conventions… before I jump in there. One thing that came to mind as you were as we were talking is, I know some people have used ChatGPT for reformatting of code. So reformatting it very quickly for those who don’t know like if you write code in a certain way in an overtime, keep adding and adding and adding. At some point, there’s sometimes ways to do it more efficiently if you can start from scratch, but that’s like typically the thing no developer wants to do because it’s more time consuming and you’re not really adding new functionality, just making code more efficient. But it’s super important because if you don’t you’re creating technical depth and sooner or later, it’s gonna hurt in a bad way. So I actually suggest a engineering done to try that. Now we tried it with BT to let files be re formatted or re factored and it didn’t quite work it as we want. And I think it’s partially because… but you’ll be able to explain better to me but I’m guessing it’s partially because it cannot adhere to our specific coding conventions, right? So, if there was a way to, and I’m sure we can give some instructions to ChatGPT, it’s just that I know from my own practice that you can provide instructions to ChatGPT and it listens to it kind of for a while and then it starts doing its own thing. Again. So if you could just say, “hey, ChatGPT, these are the rules” and then reformat that will be super powerful. So that’s probably a company in itself.

AB: Yeah.

AL: You could start for that use case… to your question of… just a little bit how we think about our code. Now, our code by the way, I mean our own solution is in a JavaScript. It’s for those who know front end is react back in those notes. So I honestly don’t know at these days what we’re using. I used a quote four years ago when we started this thing that it’s been a four, five years ago, time flies. So I’m past that point, but… what we do of course generate SQL and all the SQL, we create is already optimize as much as we can because we want to make sure there’s no, there’s none of those infamous time outs in Marketing Cloud. Yeah. But, we have other ways around that too. I’m sorry, I feel like making a say on a side here. But aside from just having as optimal optimized code as we can to make sure that we don’t have time out, we also have other ways to work around that. And I don’t think I have to go into too much detail right now. But so far as to say one of our customers, has a huge database and when is a huge database? I mean, they have 170,000,000 contacts and they have one data extension that has 1.2 billion records.

It’s a huge customer not just from ours but, from Marketing Cloud. Pretty cool. Unfortunately, I can’t mention their name just yet, but yeah. So we can handle that stuff. So, and then maybe my last thought on what you said is I guess the question is also like when do you use AMPScript versus SSDS?

 AB: Yeah.

AL: I’m gonna be really honest. I’m not a super expert in either, like I know how to write in JavaScript  and AMPScript. I think what you said makes sense. I think the reason why, so what you said was AMPScript for small stuff, SSGS for more complex stuff. And that intuitively makes sense to me because as us is JavaScript which is way more powerful than AMPScript, which is very specific language to Marketing Cloud. I think why they drop into AMPScript is that because most people start using Marketing Cloud even if they want to learn the technical stuff, JavaScript is its own language, right? So you really need to be a developer to understand what’s going on.

AB: Yeah, but I think also like it’s like in most of the case like as soon as you build like complex solution, you have to use both and you merge like both language into the same… piece of code. And because there is things that are easier to do with AMPScript because I think it’s like, a new version like it’s more recent than SSGS. I think SSGS is an old version of JavaScript basically. But yeah, yeah. After like it depends but I think it’s really interesting is if as an organization you define when to use what and also like you try to standardize and a naming convention for pretty much like everything. Because like if you do this effort as the beginning, then you will have like a clean Marketing Cloud instance and everyone will benefit from its event.

AL: Clean Marketing Cloud instance. Now, you’re just talking nonsense, let’s get real. But, but I mean, there’s also a few use cases. I remember that SSGS can solve specifically, that AMPScript just can’t and I think one of the use cases is if, when you wanna deal with rows in an e-mail, so a concrete example, you’re a retailer, someone makes a purchase with three items being able to show those three items in the e-mail I believe can only be done with SSGS. Not really with AMPScript in a good scalable way.

AB: Yeah. I’m not sure about this one.

AL: Well, I mean someone can maybe check fact checkme. You like hopefully maybe one of the NPS is listening to comment, in the comments on the video that’ll be it’ll be pretty well.

AB: And, and just yes, like DESelect is… I was wondering like DESelect is the organisation on the AppExchange for Marketing Cloud. That, that is one of the, you guys are one of the most popular apps basically on the AppExchange for Marketing Cloud, no?

AL: I believe so. I believe we have more reviews than any other provider.

AB: Okay. Yeah, that’s a huge… what you created over the time.

AL: Thanks you’re very kind. Well, I feel we’re still only getting started there’s so so much more we can do, and so many more customers to serve and so much more stuff to build. But I’m pretty happy where we got so far. I think it’s something to be proud of you.

AL: Do you think like from your point of view, there is a stage where you would like to integrate DESelect with the other Marketing Cloud marketing platform, for example, like Adobe or I don’t know Marketo or like, you know, like or even DESelect becoming like your own platform. That could, what’s the future plan for DESelect from your point of view?

AL: It’s a great question. I think about every investor asks me about it and every single one who works for us asks me to. So first of, I will say we are really already a platform in our own right. What I can in with that is we have multiple capabilities and we actually run our own infrastructure, right? So, our solution like Engage is built on Google Cloud platform actually, which is massively scalable. It can filter… 1.2M messages per hour for a given customer, which is huge… and I can probably scale it up infinitely by the way. But so it’s its own platform. But to answer your question, when, and if we support our platforms, I will just say stay tuned to.

AB: All ready… that we at like we wanted to know, but yeah, that would be interesting. Like I think it would be exciting like in the future like, but yeah, but if you say stay tuned, I will not push.

AL: I’ll leave the audience with a cliffhanger there. Yeah. Hey, coming back to if we can geek out a little bit longer about code principles. There’s actually another sort of design principle customers can make not just between SSGS or AMPScript, but also flow between AMPScript and SQL. Now, let me interrupt that. So AMPScript is more of a front end language. Typically you use it to personalize your emails. SQL is a backend language using query your database. So what’s the link? Well, we see. And this is something that we didn’t really think about when we start out to be honest, but we saw a very smart customers use DESelect to create queries, where essentially they did all a personalization already in the data. And so when they write the e-mail and afterwards, they can use drag and drop personalization fields and they don’t need developers time anymore to create those front end e-mail templates or at least much less. So, I think that’s another interesting design choice you can make as a company.

AB: So if I understand correctly, it’s like, you link AMPscript with your e-mail template. No, with like the SQL basic?

AL: You don’t even need AMPscript. I mean, there are personalization fields. I know there’s script personalization fields too, but I think there are some drag-and-drop functions too to just get field names. So say, your first name, right? Like a very simple one. So what people can do? I, so I’ll give one example say, you have a use case where you want to send an e-mail to a person and you want to… refer to their last purchase like, you know, you got a pair of sneakers, you wanna say, hey, Arthur about your sneakers, whatever. So there’s two ways of doing that. Essentially, you could do something like in AMPScript, say, I’m gonna create a lookup to the table or purchases and then the last item and then whatever, or you can say, well, I’m gonna write a SQL query that merges contacts with their last purchase. And so I, the data I have it. Now, if you don’t have any other options, these options are kind of equal, right? It just really depends where do you want to manage your code? But if you want nontechnical marketers to use the e-mail itself without relying on developers, it’s probably better that you do the coding in the back end in the, excuse me, in the a, in the data in the data. So our customers who have these like segments, they figure this out, and they do all the personalization as much as they can at the level of the data which they can do with non technical users too because it’s an intuitive drag -and-drop in. Yeah. And then in the e-mail, they hardly have to write every any AMPScript anymore.

AB: Yeah, yeah. Yeah. But I think from what I see, we have a lot of customers who that would benefit from DESelect, I think that’s impressive like the number of Marketing Cloud users who don’t have like technical knowledge, and I think, and I think Salesforce is not pushing like Marketing Cloud to be a product that is accessible for everyone. Because, I think that they see themselves as an industry more than a company, and this industry also rely on partners. So, so, so I think like the fact that there is complexity is not… is that 100 percent an issue, you know, like they don’t need to make the product as easy as possible for everyone. And I think where DESelect bring value is like, in this challenge that people without technical knowledge can start using the platform. Is that what you see as well?

AL: I mean, yeah, first off, thank you for the kind of words. Look here’s. The thing, I think there’s I think Marketing Cloud is a super performance platform and it is very flexible. Now, the downside of I of that is it’s technical, but the people who originally started ExactTarget, I think they knew what they were doing. I think they were specifically catering to a very technical audience to a very technical user base. Just like Pardot. There’s a very interesting interview with Adam Blitzer, you can find on YouTube where he talks about his journey with part and he says they made the design shows to do the opposite thing, right? They, they went for super easy but less. And I think there’s a customer base for all now. So you need to really understand it because that’s really the history and the legacy, of Marketing Cloud. And I think it’s hard to get rid of without losing some of the value of the flexibility. So it does unfortunately mean that… some customers they may be buy Marketing Cloud thinking it might be more, you know, end-user friendly if I can call it that way. But yeah, that’s often where we come in, that’s often where we come.

AB: Yeah. I don’t know. But I think it’s a great platform in or the case like, yeah, like I remember like the first time I opened in Marketing Cloud was a bit shocked by the design of it, like the UX design of it because it was…

AL: What are you talking about? I have no idea what you’re saying.

AB: I used to work with like started products. So more like may seem, you know, like this kind of small… polished products. And then when I opened Marketing Cloud, I was really, is that the platform that I will work every day with. And at the end of the day, once O, I think you need maybe one, one or two weeks of practice, you know, like to get used, to the interface. And once you start using it, you understand that there is no limit to what you want to achieve with the platform like more or less like everything you want to achieve is doable because you can cut in the platform. Yeah, I think it’s yeah, that you think, is that as well?

AL: Yeah. I mean that’s why it’s so powerful. But Arthur, let me because I really appreciate your candor. Let me ask you this question. What do you see as the biggest challenge in the Marketing Cloud community today? That nobody is talking about?

AB: There is, there is a challenge with the sandbox. I think like… there is a challenge where people cannot learn Marketing Cloud without purchasing the product. So it’s really difficult if you are like, a Trailblazer to learn about Marketing Cloud because you don’t have access to the platform. So all you learn is like theorical knowledge is like learning how to drive on GTA for example, like.

AL: I’m a gamer, I know what you’re talking about.

AB: This is a challenge and also, yeah, like the customer cannot. Yeah, there is friction at this stage, to upskill like, the new wave of like of Marketing Cloud like specialist or consultants. I think this would be interesting to have like a proper sandbox where people can sign up, and use, and also like customer… like the entry ticket for Marketing Cloud is quite expensive. So, so, so, so like… so customer cannot really try it without, you know, like it would be nice if customer could try the platform before signing it.

AL: Yeah. And that’s what we are doing now too. We, we also start with a free version of  DESelect Segment so people can try it out and they can install it themselves. But, I see we’re kind of coming up at time soon and I feel we can probably go keep going on for a few hours, but still let me maybe end with this question. Then if I may, because you start talking about, the challenges in learning it. So I’d like to end these interviews typically with the question. What advice would you give to Marketing Cloud, newbies or marketers? Maybe coming from marketing agencies like yourself. We’re trying to transition to Marketing Cloud. What would be your advice to them?

AB: Google is your friend. I will say that because a lot of the answer are already on the internet, like on the Stack Exchange, on YouTube, on the article. So, so, so the more you read every time you have a challenge, like say should read and find the answer on the internet because it’s available. And they are not the first one who face this challenge like me personally. Like there is for most of the things that I build, I have to search on internet. Like, I don’t know, from the top of my directly. So, so it’s like I don’t have like, yeah, like storage, you know, like I a, I just search the information, and I think by doing that, it’s you will work and it’s not about knowing already. It’s about being able to search the information. I think this is a real skill of that. It’s like Marketing Cloud. But I think it’s everything else that is like in our industry, like how do you search the information? How fast, you find the right answer? And, and, yeah, that’s my advice.

AL: It’s funny that you say that I used to say that consulting is just being able to use Google, very well. Oversimplifying, but there’s a lot of truth to that, so.

Thank you for, again for your flexibility for making time. You’ve been a very engaging guest here. So, yeah, I hope to see more of you in the future.

AB: Thank you! Goodbye.

Discover how our platform instantly optimizes your Marketing Cloud

Discover how our platform instantly optimizes your Marketing Cloud