Maximizing B2B Saas Paid Ad Campaigns with Eran Friendinger, Co-Founder & CTO at Voyantis.ai
In this episode of Exploring Growth, host Lee Murray talks with Eran Friendinger, co-founder and CTO of Voyantis, and an expert guest about modern B2B SaaS marketing strategies. They discuss the changing landscape of paid ads and the importance of testing and feedback loops. The conversation offers practical insights on optimizing campaigns and leveraging AI to boost customer value and marketing results.
Have a guest recommendation, question, or just want to connect?
Go here: https://www.harvardmurray.com/exploring-growth-podcast
Connect on LinkedIn:
Lee – https://www.linkedin.com/in/leehmurray
Eran – https://www.linkedin.com/in/eran-friendinger-5b38506/
Eran Friendinger
00:00:00
With the advent of AI and AI, the cost of production has gone so much down. If you look a few years back, the cost of production for creatives was the prohibitive factor. Now it doesn't cost that much. And the prohibitive factor is not creating that. It's testing each and every ad that you create your little money to see if it works, and if it doesn't, you kill it. If it does, great, you double down triple.
Lee Murray
00:00:26
All right. This episode is for B2B SaaS. If you're running paid ads and frustrated with the quality, we're going to cover some interesting ideas here that I think can help you tune your campaigns for optimal results. Today I'm chatting with Iran co-founder and CEO of Volantis. Welcome to the show, Aaron.
Eran Friendinger
00:00:49
Thank you. Lee. Happy to be here.
Lee Murray
00:00:51
Yeah. So I was asked my guest give a little background of kind of how did you end up here? to give some context for your ideas.
Eran Friendinger
00:00:59
Okay.
Eran Friendinger
00:01:00
so. I'm a software engineer, by training.
Eran Friendinger
00:01:05
and my first job was in the industry with a company called techie, which was at the company, back in 2011 dealing with, brand, brand advertising and advertising according to, trends in social media. So what people were talking about and kind of, tailoring creatives and, targeting to, to to the conversations that were, almost real time, and then and then after being there about for two years, I found that, that I'm very attracted to performance marketing rather than, brand marketing because I'm a data scientist by training as well. And, and, I like when I've got feedback data to improve whatever I'm trying to the type of feedback I'm trying to create. So when I left the turkey, I started a company called Alliance, which is a play on the audience in science, where its audience and science and audience were started. I started with two friends, so three of us set out to build an predictive analytics platform for mobile advertising back in at the end of 2012, and and to help them understand the users and basically harness their first party data.
Eran Friendinger
00:02:34
that company was acquired after 18 months, by Market Dotcom in the UK, an e-commerce platform, and I stayed there for a few more years. so that's kind of what the trajectory that took me into the marketing space to begin with. and I was already coming into that with a passion for analytics and big data. My main takeaway from audience was that the predictive analytics platform is step one, but for Marketers to get analytics and insights is not enough to actually translate this into actual business impact. And they often find themselves at a loss where they actually need to translate those insights into actions and actions with automations, right? If you know that a specific audience would respond well to a specific type of messaging because of what happened in their data and in the in their interaction with your product, you know, into some kind somehow automatic into creating relevant creatives, targeting those relative interests and then, actually serving ads to them and then repeat the process again and again. So building this kind of automation loop requires engineers, data people, operations and so on, and which usually are not at the disposal of the go to market things.
Eran Friendinger
00:03:57
and so I At this point, I've kind of, my takeaway from this was whatever I'm going to to build next, I need to take it all the way through, all the way through to automating the loops that would create business value. Then when I met my co-founder and our CEO, Ido we talked about that a lot, and we thought about the fact that generally go to market teams, whether the sales teams or marketing teams, or product marketing teams, they to, be able to close that loop. But it's also very important to find the kind of use cases for automation, because there is a lot that you can just go like, just running in a, in a spare wheel and not really actually creating value. So we, we spent a year and a half looking for, those use cases. And that's what all of this does today. We can expand on that. But this is how we got to that. And today where this does exactly that, that we help companies activate their data, turn their data into actual ROI rocks.
Eran Friendinger
00:05:02
We help you send the ads to the right people, bringing your high value customers. We help you activate them once they're in and get them to actually become a payer after they start testing your product. Kind of from 0 to 1. And we also help incentivize the right upsell at the right time with the right message. So throughout the journey.
Lee Murray
00:05:26
All right. Well, I mean that that's a pretty extensive background. And it's and it's cool to see the progression of how you went from the technical to really what I would define as like more psychological, where you're getting into the interest of the buyers, which I think is where the where it's one. I mean, that's where you're going to win people is, oh, wow, they know me, but how do they how do they really know what I'm thinking? Right? so getting several layers deep, let's let's jump into this because I think you have a lot of value to give marketers. how do most B2B SaaS marketers approach paid ads.
Eran Friendinger
00:06:02
Yeah.
Eran Friendinger
00:06:03
So, it really depends on whether you're talking about search or social ads. But the common thing is that you, when you set it up to begin with, you'll probably be just trying to you're always in a need to teach the network something so it can then harvest, turn starting to work for you. So there's always this to show you that you have to pay, right? Then you pay it with marketing. Spend both for the network to learn and for the marketer for the network to learn. Just because the first time you ask it to bring you a paying customer network is not going to do a very good job because it will okay your business. I'll try to get you to potential customers that would bias US product, but what will they buy your product? And me as an ad network first advertising. I don't know anything about your business yet. I'm going to get much better at it once you start providing me with feedback, you'll spend a few dollars on my online platform. I send some traffic here, your direction and your direction.
Eran Friendinger
00:07:06
Then you say, okay, this batch of of clicks is already nothing. These clicks actually do have some good traffic and then the network keeps improving. So that's the tuition you pay at the network level. And whenever you store it within your network, you pay to get the tick tock. You think every time and every marketer knows that, right? There's also the tuition that you pay on yourself. So how do I what's the best setup for the campaigns? What am I targeting? Landing pages and so on. Once you get all of that working, you start leveling up, right? Start increasing spending and start improving your performance. the main contract or the main API or protocol between marketing and network is that feedback. And that feedback needs to happen one. Quickly. Quickly. claiming the first seven days after it is shown. That is correct. two. It needs to be one thing that you said you should. You want to optimize for sign up. Okay. One optimize for an optimized vote purchase.
Eran Friendinger
00:08:11
Also choose one for each campaign. And also it has to happen frequently enough, right? because there's a minimum number of of feedbacks of persons that you need to provide to the network for the network to be able to. So when you begin, you'll probably be optimizing for visits. So website visits or registration or sign ups just because you have not yet spent not reached a level of spend that will drive anything else that you look at, like drive started or got 14 purchases is just going to be so few and far between that one is not going to learn anything. And then it just that starts working. It starts spending, increasing your spend and you'll get no value on that. You can say, okay, I can go down the funnel I no longer need to look at visits. I can look at registrations. I no longer need to look at the station. I can look at whatever. And so you keep leveling up those events, those good presents, and you need to pay that tuition fee for the network every time you do that.
Eran Friendinger
00:09:12
But it is worth it if you're going to increase the span at some point, you reach a point where even if you want to go deeper down the funnel, you can't even if you increase them because they happen too late. So if you think about most such businesses, they either have a freemium model or a free trial model, right? to have a B2B, product that actually makes money off the bat and that's it's more transactional or volume based, consumption based, in which case, still most of the volume and consumption will probably not happen in the first few days. And then you when you reach that point, that's it. You're kind of stuck. And this is where you're, you're you're going to have this split brain between what your team is being measured on and what you are measuring the network on. It's going to be different. You're measuring the network on until I started. But you're being measured on paying customers right at the quarterly level. That's that's that's your goal. You have like spend goals, volume goals and cost goals, cost per customer, or maybe even worse goals like actual internet.
Lee Murray
00:10:19
That's a good point. And it sounds like you're just saying setting expectations on both sides of expectations of the platform, expectations of what you're trying to do internally. So, you know, automatically you're you're you're painting a roadmap or you're putting together a roadmap of a much longer cycle than just jumping on and getting results so that that's, you know, become real, real obvious. And it sounds to me like from a content marketer standpoint, it sounds like you're almost like working through top middle, bottom funnel. Yeah. Data. Like you're trying to get a, you know, a wide tam and lots of data that you can then refine and refine and optimize through your call to actions through the funnel.
Eran Friendinger
00:11:05
That's exactly it. And at one point, you can no longer go deeper down the funnel, although you want to. Right. That's that's the big pain there. And then you have to go with analytics, right. It could be AI. It could be just regular predictive analytics. It could be simplistic.
Eran Friendinger
00:11:21
But you have to start playing with your data to actually get a quick enough feedback to provide the network and your home team. Forget about the networks like you've run a campaign. You spend, I don't know, $1,000 a day. It's been three days. You've spent $3,000. Is it working? Should you kill this campaign? Should you actually increase the budget? No. No paying customers yet. Only sign ups. What do you do? Yeah. Right. So probably what you do is you say, okay, there's no paid customers yet, but those users that came in three days ago, how many of them actually Have been coming back day after day. How many of them have invited other teammates? How many of them used a specific feature that is very indicative in the platform of reaching a paywall and maybe actually becoming a paying customer, and so on. So we're going to start doing all of these like, slicing and dicing in your analytics. Assuming you have the resources, the analytical resources to do it, which is not trivial.
Eran Friendinger
00:12:20
And once you do that, you actually want to kind of like make a decision, you know, this is how I'm going to measure. I'm going to measure, not like signups, but my sign ups for corporate email sign ups that had at least three team members, right. And so on. And AI is just the natural way of extending this Excel like, analysis into micro segments. So look at all of the data, all of the behavior of your of your cohort of users within your product. And basically they're telling you what the customer product fit is, right? Every single individual customer. And then for each. This is basically the lead scoring right. It's the equivalent lead scoring to sales. When you look at all of the data that you have to say how good are they? How good is my product as a fit for this customer. And you're going to use it. Every evidence that you can gather, whether it's behavioral evidence or user attributes or device attributes or location.
Eran Friendinger
00:13:21
Right.
Eran Friendinger
00:13:24
So that's that's actually the way to end there.
Lee Murray
00:13:28
Okay, I like that. Well, we'll come back to the lead scoring here in a minute. what's so you're talking a lot about like, honestly, like, kind of just start up like go to market and and trying to really get go to, let's say, Google paid ads and you're going to, you're going to really get the platform to learn what you're doing. You're learning at the same time. And as you're as you're building, you're getting you're getting more, you know, educated. you're going to start to sell people. And of course, it depends on your model and how you're, you know, selling your, your product, and what stages they have to go through. But at some point you're going to, you're going to probably get pretty close to the ideal market and you're going to dial that in. what's at that point then? A better like, what's a good way to target more valuable customers? More so than just new customers?
Eran Friendinger
00:14:24
so in order to target more valuable customers, I would say there are two main levers or maybe three levers, for a marketer.
Eran Friendinger
00:14:36
one is the actual targeting. Like demographics, locations, interest, keywords. On the second one, would basically basically be your creators, like, right. Whether it's the text that you put in the results of the Google search or the video ads that you show and so on. The third one is the feedback that you're providing back to the networks. So let's start with the creatives themselves. I think it's it's somewhat counterintuitive. And until you think about it for a second, but an ad is an audience, right, in the sense that for each specific head, there's an audience that will join us on Twitter. So we need to find that audience. But somewhere out there, there's the audience, loud or small, that would fit this up. This does not guarantee that that audience is a good fit for your product.
Eran Friendinger
00:15:29
Yeah.
Eran Friendinger
00:15:30
As is often seen with gaming ads where the ad has nothing to do with the game, right? So it's not enough to find the right audience for this, which is what the network does very well.
Eran Friendinger
00:15:42
Right?
Eran Friendinger
00:15:43
You need to make sure that the audience, that the target is also good for your product. So different ads means different targeting. And if you want to look for the high value The customers could try different ads and see which one of them drives high value customers, right? so that's, that's that's a huge part. And I think with the advent of AI, the cost of production has gone so, so, so much. Now, if you look a few years back, the cost of production for creatives was the was the prohibitive factor in why not just why not just done lots of great things because it costs money to, to build to create them. Now it doesn't cost that much. And the prohibitive factor is not creating the that. It's testing them each and every ad that creative that you create, it will you know they doing this money to see if it works and if it doesn't you kill it. If it does, great. You double down, you triple now.
Eran Friendinger
00:16:41
Okay.
Eran Friendinger
00:16:41
So that's one aspect that created the other aspect is the targeting. And that aspect is slowly going away because the ad networks. So you could you could track like you think a few years back you could be target get interest groups in an inverter and so on. Right. This is great the way they have. Every network is geared towards broad targeting. Even in search where you have keywords, they have target, they're pushing you towards broader targeting. And the new advertising products like Max and universal are campaigns and play and Tactics Plus and so on. All of those are new products are basically telling you to stop telling us how to target. We know better than you. You just give us the creative and the and the signal and the feedback, and we'll find the best customers for you. Well, there's always going to be some level of targeting. You're going to be around like geo and demographics because you just there are some models that you don't want your ad spend to go at all. But the networks are getting more and more opinionated about this.
Eran Friendinger
00:17:50
so I would say that's an interesting part, and especially when you think about conversational AI and how they should monetize it to ads. This, almost by definition, even keywords are no longer relevant in the world of conversational search. But what keyword does the whole conversation going out there? How can you, like say what you want to target? So the AI networks like this is an evolving thing right now, but I'm guessing it will be like, just give us the the video, the the image, the landing page and the and the feedback signal. Just let us do that. Just all very, very complex AI machine learning.
Lee Murray
00:18:29
Yeah, I would agree with that. I see that more and more as a user to just use different tools and platforms and apps and stuff, and how they're targeting me as a consumer. you know, they've got it pretty dialed in.
Eran Friendinger
00:18:43
They do. so out of those three, we we covered three areas. We covered everything. Then you've got the signal itself, the feedback.
Eran Friendinger
00:18:50
And I think that is the most under under, under looked, so it almost overlooked, a lever that you have as a marketer because it is super strong. So you always like, you use it and you improve it at the beginning it's going to scale. Then you get stuck. As I said, however, there's this whole concept of value based bidding or value optimization or target towards the the audition, names for smart bidding. But the idea is that the user is not either good or bad for your business, and your product is not either good or bad for this user. Different users are going to get different value product, so you should price them different in the sense of what bid would you put on a possible impression for that? And the best way to do that is use max conversion value, target value optimization, all of those marketing strategies in the ad networks, that, however, requires that your feedback is not binary. It's not a yes or no. It's not a good or bad.
Eran Friendinger
00:19:53
It's on. How much? How much dollars is this specific user going to be worth to your business? If you do that, if you can provide that kind of feedback to the network efficiency experts, that it's like you can get gains of tens of percentages 20, 30, 40%, even even see 100% improvement in loss, which will then let you scale significantly. and some businesses have it easier doing that than somebody who's a bit hard. Yeah. So the ones that have it easier is anybody that is like trying to make a excuse my French, but just trying to make make a quick buck. Not in the sense that they're not providing value, but instead that they try to make back their money very quickly. And that would usually be, industries where it's like a bootstrap business, often e-commerce is that when you're running on very low margin and whatever you're showing you, you want to make it back on the first order, but you don't have time to wait like 90 days or so with one year to make them make money, unlike SAS companies.
Eran Friendinger
00:21:02
But it's flipped. So if you're an e-commerce company or or, I don't know, a hyper casual gaming company where you make money very quickly, users would not stay long in your app. then you can use value optimization in the network and actually provide the dollar value that you're making. That would work amazingly well. Okay, so it's the e-commerce market. If you're not using value optimization, you should just do it. You don't need a GUI for it. You don't need anything to prove your performance. but other businesses, especially those that cultivate long lasting relationships and loyalty with their customers and land on expansions of revenue, especially in circumstances where it's not only that people will take time to start buying is even when they start paying you, they are more likely to pay you more if you wait more. Exponentially. Or as their team grows and they upgrade to higher tiers. In that case, you cannot send the value within the first seven days. There's just no actual revenue to send. Then what do you have to do? You have to model the value and then you die.
Eran Friendinger
00:22:09
And if you can use AI to capture all of the behavior of a single individual customer within your product in the first seven days, preferably in the first 24 hours, if you can, by the way, and you can model this into kind of a predicted lifetime value, which is what we do at work as well. And if you have the right methodology of sending it back to the network and signaling it, what a good user looks like and how they work, then you get those guys I'm talking about. So we work with customers like Neil and Hannibal, NerdWallet and so on and so on. We've seen a lot of of of improvement, as I said, north of 20%, often north of the 4,050%.
Lee Murray
00:22:51
Yeah. That's amazing. And it sounds like the thread through this whole conversation is feedback. It's like looking for those signals from the very, very first ad run all the way to customer retention and long term lifetime value. Yeah.
Eran Friendinger
00:23:05
Yeah. And it's as applicable not just to the case of paid marketing.
Eran Friendinger
00:23:10
It is the acquisition. If you think about CRM of users that have signed up for a platform but have not yet become payers. Then there's a bunch of ways you could, think about how to incentivize them to move further down the line, but it's only going to work if it's relevant for some of your free users are just not relevant. if you're if you're a tool like Miro and let's say, a teacher signs up for a service to demonstrate something for his class of students, the chances of that that account becoming a paying customer, it doesn't matter what you've done. Like they don't need to pay. They won't pay. They don't know how to pay. so sometimes it's not relevant, but if it is relevant, you want to know how and when to act. Yeah. and again, if you apply AI and machine learning to to crystallize all of the behavior of the user within your product, then you can prioritize what you, who to talk to, you, who to approach, when to approach, and what the messaging should be.
Eran Friendinger
00:24:16
And should it include some kind of of, financial incentives such as a discount or free trial or whatever? That's also very valuable. not to mention the fact that when you're talking about actual salespeople or talking to SAS customers, that's a very, very, pricey resource, right? Yeah. So you really want to to borrow that?
Lee Murray
00:24:43
Let's get back to that, because we were you mentioned it earlier about lead scoring. So, you know, talk more about how, you know, lead scoring to prioritize.
Eran Friendinger
00:24:51
Yes. So with lead scoring, you need to, basically, score your customers either by their potential to, to become a better customer or even better just by the potential to become a paying customer. Times how much will they actually pay? So they they expect to see the the expected revenue from that customer. and you really want to do that to be able to prioritize a very, expensive resources such as stars and salespeople. and you can do it on, I would say either a CRM or leads or the your user base of free users, where you have more data about their behavior because these are just leads, but your CRM of users is significant.
Eran Friendinger
00:25:43
All the parameters between what we call hindrances, right? People that are actually using their products and then they while using the product, they I click the top two sales button, right. So these are the three streams that you could look at all of them together. Because usually you look at them separately like this. You've got people that are taking care of animals. You've got people that are taking care of, like just the leaves just chilled out some for the other, people that are trying to reach out, like as account managers to the free users. And you should actually look at it together. It's a, it's it's a joint pool of customers. And you should always prioritize the cost of the results that you're going to basically assign to them with the possible outcome and the probability of that becoming a customer. So that's that's how I love that.
Lee Murray
00:26:39
This is a very eye opening and, and, and affirming for some of the a lot of the ideas I've currently had, you know, I live in the content marketing world, more so than the, performance marketing or product marketing.
Lee Murray
00:26:52
And, so it's good to see, you know, a lot of the same ideas are being translated there. and, you know, to the nature of the we, I name my agency Signal Media because that's what that's the when it comes down to it, what you're really trying to get at is the signal, like, what are all the signals exactly? You know, and almost like as we're talking about feedback, it resonates a lot with me. I mean, this is how I think, but, it's almost like, you know, they have, sales enablement, positions or, you know, they'll have like, kind of like the role behind the roles. I almost think that there should be some evolution in marketing where especially B2B SaaS, where there should be a feedback driven role, where it's almost like they're like a CRO type of role, but they're more focused just on gaining feedback at all the touch points. So they're the signal, you know, Communicator. Not necessarily that they would push out messaging, but that they would be taking in all the data.
Lee Murray
00:27:56
Making sure it all comes in at every, every point. And I think like talking to the AI piece of it. I think then if you can bring all that data internally as much as possible, you can apply machine learning to that, that new layer, and it tells you so much more. It'll tell you that the it'll probably affirm a lot of the hunches that you've had. you know, like you're mentioning kind of hunches.
Eran Friendinger
00:28:22
Yeah. Yeah, exactly. Always like when you look at the predictive model, a lot of the signals there are very intuitive. Oh, yeah, that makes sense. I knew that. What it does is it formalizes the structure of the relationship between those hunches.
Lee Murray
00:28:37
And it will give you new, innovative insights into not only the product, but your audience of how you connect both of those together. I think having that new layer that's been, sort of told to you, and confirmed. I think that'd be super valuable.
Eran Friendinger
00:28:57
I mean, I see it all the time.
Eran Friendinger
00:29:00
It's definitely not.
Lee Murray
00:29:01
Yeah. It's amazing. Well, thanks. Thanks for coming on the show and bringing all your practical wisdom. I mean, you have a lot of experience through the years of, going from technical to psychological and now applying all of that thinking, technically with AI, I think is just a huge, a huge thing. So thanks for for being on the show. If we want to send people your way, where do we send them?
Lee Murray
00:29:25
well.
Eran Friendinger
00:29:26
A they can email me at Iran. Iran at points I it's the and I or you can just reach out on LinkedIn and that would be great.
Lee Murray
00:29:41
It's awesome. All right. Well thanks again for coming on.
Eran Friendinger
00:29:44
Thank you for having me. It was a pleasure.