Janis Zech: Hello, and welcome to another edition of the RevOps Lab Podcast. My guest today is Daniel Foulkes. Daniel, very warm welcome.

Daniel Foulkes: Thank you. Thank you. Happy to join.

Janis Zech: Could you give our listeners a short introduction about yourself and where you work and what you do?

Daniel Foulkes: So I'm the senior marketing strategy and operations manager at CoachHub. So we are a digital coaching company for enterprise businesses. Basically, we provide one to one coachings and also one to many coaching services, but we do it online. So we do this through our own platform, and we have our own analytics. We also have our own coaching lab, and we also develop a lot of the science around behavioral science and how companies interact each other. But I live behind the scenes in the caves of the operations. I'm now in the revenue team, but I started within the marketing team, and then I switched later to the RevOps team. But me, myself, I come from academia, so I did a couple of master's degrees in South Korea and also in Australia in security science, security studies and political science. And surprisingly, political systems are a little bit less complex than business systems, but then you have to deal with all the other human sides. So it's also very complicated, but it's a nice challenge to have.

Janis Zech: Nice challenge to have. Great. And how did you end up in operations then? So coming from academia, studying security and politics, political science.

Daniel Foulkes: So I also have a background as a journalist, so I was also very interested in media and how it played out. And then I relocated to Europe and then joined the tech scene in a marketing team. So they said once a marketer, always a marketer. But because I come from this trying to analyze political systems, it was always about, can I find the data to back this up? When we talk about changes in politics, well, what are the actual measurable things that I can use? Right? I come from a very social science perspective, so within the marketing side of things, I also helped with some HubSpot integrations and implementing HubSpot, and that sort of led me to become involved with databases, how they work, and also the CRM. And from there, it was a jump into operations and how do those systems work, and what is the difference? Like, how is it that Salesforce can be great for some things, HubSpot can be great for others, and then how teams react on it? It's a great study for how organizations grow, but also I find it fascinating to sort of find elegant solutions to complex problems. And then the marketing teams across all industries have a lot of problems when it comes to systems and tracking, and then attribution has been a very big one that we've tackled at CoachHub, and one where we have developed a system that I'm proud to talk about today.

Janis Zech: Great. Yeah. I think marketing, you know, there's like these two sides to marketing. One is the fleshy shiny, you know, like, where you create all the ad creatives and like, do all the commercial shots for cool ad campaigns and so on, and then there's the very technical side of things. It's sort of like under the hood, and there's a lot of work. But you also need to find creative solutions sometimes, but it really has very little to do with that sort of other, like, very shiny part of marketing.

Daniel Foulkes: You're right. There's a side of it. Like, we have some marketers, they do great online and offline events. They go to conventions to have stands that look amazing, designers that do great things. Our digital team does ads, is super targeted, and I come and say, well, guys, the CSV here looks wrong. Like, look at that, you misplaced the semicolon. It's gonna overwrite the data. So I look at the other end. Right? It's about how do we track these things? How do we get those engagements? And, well, this is how we came across it. When I joined the marketing team at CoachHub, it was very clear to me from the start that these are high performer individuals. They do a lot of activities. Right? This is where we sell an online service. Right? So it has to sort of get that gap across in showing something tangible or something that they can touch at an event. And then online, there has to be a follow-up. How to track this is something that it was very important for us to say, well, it needs to have at some point digital. So we need to track when they go to an event, but do they come to the website? Are they in our database? Do we follow-up on them? And then that's sort of where it actually started coming from. So marketing, are you bringing new business? Are you bringing leads? Are you influencing them? And the never ending question of what's the ROI of what you're doing because marketing isn't cheap. And to do very good marketing, you also need to put a lot of money behind it. Right? So at every point, every stakeholder is looking at those expenses and saying, well, are we getting, and what's the ROI?

Janis Zech: Yeah. You're already raising, I think, a lot of very important questions that hopefully we can address in the next thirty minutes or so. But, yeah, like you said, I think in this episode, we're trying to fully focus on marketing attribution, and this is the second episode that we're doing on the topic. Charlie Saunders, we did a show with him a few months back on his custom object that he's using in Salesforce to optimize marketing attribution, and today we want to take a similar look at how a marketing attribution can be done, also based on the example of Salesforce in combination with HubSpot. The last one with Charlie was very focused on the specifics of this custom object, and here we go a bit broader, so on the traits of a marketing model, why it makes sense, how complicated it needs to be, but then also, definitely make sure to dive into the technical setup of it. So before we talk about the model itself, maybe you could briefly explain how sales works at CoachHub and what the role of marketing is in the sales process, like the importance of MQLs versus SQLs, and I think it'd be important to sort of frame it.

Daniel Foulkes: Those are the never ending questions. What is an MQL than an SQL? So within our revenue, on the sales driven organization CoachHub, we have the marketing team, which I'll go into a bit more detail soon. We have a sales development team that sort of does outbound, cold calling, sequences new prospects. We have the account executives who also do their own prospecting as well, but these are the ones closing deals, right? They get the warm leads. They do the discovery calls that the sales development team bring them. And then post sale, we have a customer success team that sort of looks at making sure that the tool usage is correct, right, that they're getting the most out of it, and they also assist on the renewal side. And on top of that, then we have the revenue operations team that sort of helps them both with the tools, targets, tracking, all that jazz. We are at their service, and we wanna make sure that they have the best tool stack. Yeah. The best processes possible. But the marketing team that we have, we do a lot of digital activities, and we do a lot of on-site activities. So we've sort of been dividing them between regional marketers that own their own regions, and then the digital team that supports them individually, but they also grab a global. We also have a brand and design team that also looks at our public relations and also the design. So the marketing team is big, so it's always been a mix of how much can I track you as much as possible? And then how do we attribute it, or how do we weave a story that makes sense about the expenses that you've been doing and the returns that we get? So because we do all these combinations of activities, we look a lot into now for the marketing team, we're really focused on, are you bringing in new revenue? And by new revenue, we mean both new leads, but also new logos that we haven't sold to before. Or have you expanded that logo in a different book of business region, right, in that different zone? So now it's really focused towards, hey. You guys need to drive new revenue, but we also wanna track, can you influence deals? Can you accelerate? Does the deal cycle change when there's a marketing touchpoint? So for us, that's also very important. And we've managed to come to a point where we actually say, well, the marketing team has targets, and we tie them not on MQLs, SQLs anymore. We actually say how much influence revenue do you have, how much source revenue do you have, and also how's your pipeline when it's sourced. Right? So this, we've actually managed to push those marketing results that we track them by to the end of the funnel. We still care about all those other metrics, but it also changes the conversation a lot more. Because then the marketing team is very driven towards the end of the sale, not just the acquisition of the lead.

Janis Zech: And so I understand, like, the source revenue. The other thing you mentioned was impact. Influence. Yeah. Influence. Sorry. Could you explain that?

Daniel Foulkes: So we do some differences. We look at all the prospects, our leads, so to say, and we basically look, do you come from a marketing activity or not? That's what sourced you. But even if you are not sourced by the marketing team, have you been influenced? Have you actually gone to our website, clicked our ads, downloaded a collateral, attended an event, and can we track those to sort of say, hey, even though the sales team sourced you, have you been touched by marketing in other points? And we're also trying to do that not only at the contact level, but, for example, for client marketing to have client marketing campaigns where we say, hey, actually, we are doing this for an account, and every deal that comes in after this particular period should be influenced. Right? Because we've done a specific client.

Janis Zech: My understanding is basically the marketing attribution model that you developed at CoachHub, it serves different goals. So there are the influence goals. There are sort of like the revenue source goals. There is like in the revenue source, there's also like differentiation between selling and new logos. That seems like quite a few goals that you have there. Is it just one model that solves for it, or is it split into, you know, different components? How does that work?

Daniel Foulkes: You asked a very good question. I think at a very high level, it's always about the fact that targets drive behavior. Right? At the end of the day, when you say to someone, hey, these are your team targets. These are your individual targets. They're gonna be really focused. And it's okay. We want people to do those. But it's also about how do we frame it into the sense that, yes, you wanna drive sourced revenue, but well, what does that mean? It means I have to source new leads or new logos and bring them over, and my lead has to be the one that the SD talks to, and then they pass it on to their AE, and if the meeting is good, you know, we say, hey, this also works for the SD targets, but for the marketing team, this lead that they brought gets put into the opportunity. That's you have sourced this opportunity. Right? And there's different things that you can consider regarding, like, your sales cycle. Maybe it's super long or maybe it's very short, and you wanna have, like, a time decay variable where you say, hey, if marketing brings it and it's in this short amount of time, then they source a hundred percent of the op. And you can build in sort of cutoffs that say, hey. As time goes by, you stop sourcing that much because there's more sales activity. So when we think about marketing attribution models, it's not just top of funnel that we need to talk about. You also need to understand, are you trying to deliver what are you trying to deliver for the sales team? Are you trying to get them just leads that they just cold call? Are you trying to just create opportunities? Are you trying to get stakeholders in existing opportunities? And from there is where you have an open conversation with all the stakeholders and say, I'm gonna make this marketing app. Because it's very different when everybody's in the room, when the sales leaders and the sales development leaders understand the model. They say, oh, it's actually not they're not gonna split the revenue by what marketing sourced or what I did. They just look at, this is what marketing sourced, help us track it so we can do more of the campaigns that brought the revenue. Right? Because then we have a different conversation where we say, hey, marketing brought you these leads, and I can see that their win rate is different. It's higher. Their time to close is shorter, and their average deal size is actually larger because marketing brings you these leads so much faster. Right? So we looked at doing sort of this collaborative model where a lead can be sourced by marketing and still be worked by an SD, so at the end of the day, it is gonna track the influence at both teams.

Janis Zech: Got it. And I guess so what I would assume, but please correct me, is then you don't think so much in MQLs anymore. You really think more about, like, the overall influence in the sales process that marketing efforts had. So, I mean, like, maybe, yeah, maybe stop here. Like, am I wrong here, or so MQLs still play a big role or less so?

Daniel Foulkes: This is the tricky — you've hit the nail on the head because when we talk about marketing attributions where you track it so much to the end of the funnel, even if you are in the digital team and you're looking at how much you're bidding on keywords, you don't — especially for us, we're B2B. We're business to business customers for enterprise companies. Someone that clicks on an ad is going to go through a lot of sales process, talk to a human that drives that sales process for the opportunity to close. So, yes, you care about the end result, but you have to tell people, like, look. These are the levers that you can actually — it changes the mindset to say, I can't change the sales cycle that much. What can I do for it? I can accelerate it, or I can bring you warmer leads, and we say, hey, you actually care a lot about MQLs. You care a lot about those SQLs that you get right out. Have they been qualified? But then you really sort of, this is where your responsibility stops. And for that, then it forces you to start having service level agreements with the other keys. Right? Then we say, well, we're gonna have a marketing attribution model because we have an MQL scoring system. So we say, we're only gonna give you leads that meet this criteria, our ideal customer persona, or ideal customer profile. We're gonna give them to you. Can we have an SLA with the sales team of, can you act on these leads before it's seven days or twenty four hours? Or, hey, here's a demo request. It needs to be actioned in less than six hours. Right? When you know these things, you say, ah, a marketer will not join the conversations about pricing and how many seats they're going to buy. They're just going to make sure I'm going to get you the right stakeholders to have this conversation as quickly as possible. Did it work at the end? Yes. No. Let's talk about the campaign that brought it, and how can we do more of that? So in reality, we have a marketing attribution model, so they really care about the MQLs.

Janis Zech: Got it. Maybe I think what could be worth moving to a more, you know, hands on explanation of how the model works, just to move away from that abstract plane that we're on right now. So how did the model start in its most simple form and sort of how did it evolve to what it currently is?

Daniel Foulkes: At its most basic level, it was always about tracking. Can we track as much as possible, and are we tracking it the right way? If your marketing team has a project manager, they are secretly your best ally in a marketing operations environment. I remember I sat down on, like, my first week and looked at our project manager. I'm like, you need to help me. We need a naming convention that tracks what they're doing. It also tracks these other elements that I know are important in the system. And then at the time, we were all a HubSpot shop. How is it that it connects to the next steps in the sales side? So we decided it was clear to us that we need to track as much as possible and set several dropdown field properties in HubSpot because it was one way of bringing some order to it. Single text fields are a curse that every person in operations loves to hate. They're very useful when they're done right, but then they're so prone to errors. But it was about are we leveraging the original source data from HubSpot? Are we logging in things that come from imports? At our basic level, it was trying to find a way of setting up the sourcing to identify if it was a sales development rep through some sales tools or it was an import from the marketing team after an event. How is it that we started setting those up? It started with a — we called it our CoachHub internal value property. So it had to be one of several options. And that was at its earliest level. And then it also became a point of, okay. So you're sourcing this many leads. How is it looking in the deal? And I think when — I know we wanna go down into the specifics, but it's always important to sort of say, you know, businesses, even if you're a startup or an enterprise business, you have to be comfortable with making decisions with maybe ninety or eighty percent of the data being accurate or where there's a fog of war where you don't know that much about it. So you need to sort of at some point say, when is good? Don't let the perfect be the enemy of the good. Right? It's like, you know, when you write a research paper, at the end, you put sort of these are the limitations of that research, and that's sort of the context that you have to read the conclusion in, because everything has limitations. And I think with marketing attribution, certainly, there are always trade offs that you accept willingly or not. Yeah. So we went through these conversations. We had a lot of back and forth with the different sales leaders to sort of understand, well, when you drive a conversation, when you attend a sales event, what is the follow-up that you expect? And it would lead us to complicated things, sort of we'd have marketers racing to be the first ones that created the contact in the database, or else the source would be set incorrectly, and then we'd have to change it, and it would be cumbersome. Then you need to create a process to accommodate for that. But one thing that became very clear for us at the beginning is that we wanna have a model that needed to be rules based, and there needs to be a path to handle exceptions, but it needs to be written down somewhere where everybody can read it, ideally understand it. But it also needs to be as scalable as possible. Right? You need to, at some point, have it run on automations only. It needs to be able to handle a lot of data, a lot of action points happening across different channels and sales activities and teams that do different things. So it needs to be robust, but most of all it needs to be transparent. You need to be able to — if they ask for it, you need to go in, roll up your sleeves and get down into the data and say, actually, it was this engagement of these things that brought these activities, and this is why — and that changes the conversation with the sales team and the marketing teams because we say, look. I can track everything that happens online. We have tracking cookies. We can see form submissions, but offline things, you need to give me a list. Right? And then we say, oh, but so many people registered. Well, is registration the same as an attendance, for example? Right? Is that the same? Does it count for attribution? These need to be conversations that we say, hey. But I actually brought you interested people. You followed up because they didn't attend, but I brought it in. These are not easy conversations to have with the sales team. And you will be challenged on those assumptions, and that's good. But you wanna have those happening in the open between all stakeholders, because at the end of the day, you're all working for the same company. Right? You want the same objectives. So that's something that we really cared about, to sort of bring every stakeholder as much as possible into this.

Janis Zech: I mean, I would assume those are quite a few rules. So I guess, like, a big part of that is just to have the stakeholder alignment to just make sure, okay. Now we align on these rules, so we don't need to have alignment all the time. It's more like a one-time effort that every now and then is tweaked based on the latest, you know, insights from campaigns.

Daniel Foulkes: You always need to iterate on these models. Right? Like, it's never perfect. There's always trade offs, like we said before, but those rules help you to actually make it scalable and in order and also let you automate it at some point. And it also has to do with agreements that you have within the company. Right? You can also say, hey. You know, we're gonna start a new year. We wanted to have certain revisions to the attribution model, and you bring all stakeholders. And sometimes there's disagreement in companies. And that's why we have a hierarchy that some people are in charge of making those decisions. It's on them. But then we make it very clear. Here's option A. Here's option B, and then someone's gonna make that choice. And then once that choice is made, then we all follow. Right? But it also has to do with, like, getting down into the specifics. For example, you might have a company that you can only do purchasing at a certain level with the stakeholders that you get. Right? If you're selling to enterprise, there is a moment that you might get a prospect that goes to events and goes to conferences, trade shows, etcetera, but there's going to be a point that they say, okay, so I'm going to purchase in bulk. Let me pass you to my procurement team, or someone in accounting, or finance, or my legal team for the RFP. They don't go to trade shows. They don't click on your ebook or your infographics. How do you track those? Well, you can also sort of have a transparent method where you say, the marketing team sources the first deal or the first lead that led to an opportunity. Then everything else after would be sourced. And then you have a conversation for how much sourcing do you track it for the team, or do you track it to the specific campaign? Those are good conversations to have, but they also come at a sense of maturity and also budget that your marketing team has. If you are just starting, and you are in a marketing team that is all about growing and getting new leads, etcetera, do you need an attribution model that goes all the way to the end? No. You have conversion. So you have top of funnel information that you can do day to day decisions on. Brand awareness will give you similar things. If you are growing and you have a team across many regions, many languages, and then you start saying, well, where do I put my money? You have trade shows, or you have ads, or you have client marketing events like an after hour. That's where an attribution model also helps you make these decisions.

Janis Zech: So I think this is sort of like a why a model and the different rules are so important, because, like, marketing — like you said it earlier, marketing is expensive, so you have to prove ROI constantly. And I think, you know, this is a typical story. A company is not doing financially well — like, I'm not saying CoachHub is not doing well, but just as an example — a company is not doing well, and then very typically, companies look at marketing first to really try to, like, okay. You know, I never believed in it, and then they cut budget. By doing that, maybe they just look at the directly sourced MQLs, but then they don't understand, hey, there were all these other touch points. The marketing team actually created the brand awareness that helped the sales rep or the SDR to even book that first meeting because the person on the other side of the phone actually knew the name of the company through different awareness campaigns or differentiation campaigns that helped explain the product in comparison to all the other products on the market that do something similar or the same thing. So there's this, I would say, tremendous long tail that marketing efforts can have that is really, really hard to track. I think it's not only about this directly sourced MQL and whether that then turns into booked meetings, but there are all these other touch points, particularly if you do account based selling, that play a huge role. If you're an enterprise, this is very common. You target the entire company, the entire account, and try to get your foot in the door, try to really make sure that all the more important stakeholders have heard the name and have seen the product, have a clear understanding of the value proposition as much as possible, of course, because that will ultimately help you when the economic buyer and the decision maker come to the point where they, you know, are faced with the question, do I sign a contract with this company or that company?

Daniel Foulkes: Absolutely. Because when you have a marketing attribution model, it's not going to be the end solution to every one of your marketing problems, and you can't put everything there. Right? You might wanna start saying, well, can we start tracking the engagement of our design team that participates on the internal sales and making slide decks, etcetera? And yes, but is that the best use of your time? Like, you also come with a marketing attribution to understand there are things that are gonna be outside. There is a cost for other things. Right? If you really, really wanna care about your brand awareness, you can do polling, and there are companies that will sell you that as a service or will track your press mentions. You can choose to have press hits or mentions of a specific category as a target for that team, but, you know, not every metric is a target, and not every target needs to be a metric. There's also the other side where the marketing attribution model will help you get some transparency around certain things, certain channels. Other things, they're gonna be outside of it. If you wanna look at brand awareness, you can try to ask your customers where did they hear about you, but anyone that has ever made a form that says that knows that the kind of answers you get — when you're lucky, they aren't very good. Right? Like, when you get asked at a supermarket, how did you hear about us? Like, I was hungry. You're in the neighborhood. Don't ask me. Right? It's always tricky, but it's a point where the conversation needs to happen with people that understand the limitations of each channel and can also talk to the marketers. Right? They will know their markets best and their channels best, but then also have a conversation with the sales team to sort of say, well, actually, did you see a change in this? I'm like, oh, yes. They mentioned it on that event that we saw each other, that they had seen an ad somewhere or that someone had mentioned it, but it's not just the marketing attribution model. It's the tracking that comes with it that gets the marketing team focused to say, hey. What can I track, and how can I track it as accurately as possible? It changes the conversation to the point that we've had sales VPs that come in and say, hey, this deal, I want to follow-up on it. Show me how the attribution is, because I think this should be marketing sourced. I had a conversation and they mentioned this and these campaigns. Are we tracking those? Right? This is where you switch the conversation, and the sales rep realizes marketing is their own channel for inbound leads. And the more information they get, the more information we feed onto the machine of the attribution model, the more complete — or the more they will focus those efforts later. It's reached a point that we care, for example, about certain campaigns, and we say, hey. Please add the contacts to the opportunity, because if you forget to add them, I lose the tracking of this person and this specific — and then they're like, yeah. Let me add as many as possible because I want the tracking that happens behind the scenes to say, yeah. This person that went to this webinar, etcetera, this came in. Right? You need a model, and this is the part that we've sort of evolved with to sort of say, we want a model that can improve over time, but it's also not bound by time. Right? It needs to sort of — if a rep goes to an event and sees a contact and they give them their business card, yes, talk to them. Don't wait for the marketing team to import the leads. You want a marketing attribution model that can go beyond that, that doesn't care about who created it at ten in the morning instead of the marketers that imported the leads at two PM. You need to have a way of sort of tracking this. We partnered with a company called Magic Robot to help us sort of pull this HubSpot information and sort of deconstruct it and push it out into Salesforce as specific campaigns with engagements at specific times, and also a set of rules that says, if a contact is created on day x, but you have an engagement from a campaign that we have decoded, and it is from a prior time, even before the contact was created in HubSpot or Salesforce, the contact is marketing sourced from this. Right? This kind of logic takes away the pressure for the marketing team to say, no. No. Wait until I do the import. Right? They go do their import, the tracking happens, and then they say, yeah. These leads will get some immediate information when they're created. And once our attribution machine picks it up, couple of hours later, they're gonna show up on the campaigns, and attribution doesn't need to run on demand like that. We don't need to see a marketing attribution model that is up to date the whole time. Up to date the whole time, I wanna see the lead status, for example. It's a different thing. I wanna see if a lead is being worked on or not in real time. It's different than is it from a marketing campaign that happened four months ago. Right? The urgency — you start having a conversation about which data points do you need now and which data points are important to have at a specific point.

Janis Zech: Yeah. Got it. Got it. You mentioned HubSpot, but I know that you also work on Salesforce. So I'm assuming, basically, HubSpot Marketing Hub is for all the marketing efforts, but then Salesforce is the actual CRM. Why did you make that switch from HubSpot CRM to Salesforce CRM? Did that play a role in terms of marketing attribution or was that completely outside of that?

Daniel Foulkes: So we weren't a HubSpot shop for a long time, just HubSpot, and we were running into some issues around permissions, around scalability of that tool. Salesforce, you can do anything you want in Salesforce, and that's a good thing. The bad thing is you can do anything you want. You have to also build it or understand it or maintain it. There is no one true solution. So we were scaling up and we're having operations that we needed things around higher hierarchies and different regions, permissions, reporting that Salesforce had the most robust set of answers for us. But we also recognize that when it comes to the usability for landing pages, forms, email automations, etcetera, HubSpot is the best in the market. So how do we bring those together where they don't really wanna play nice with each other? As I said, we have a very distinct marketing tracking of activities, and we were logging each day. So this is where we partnered with Magic Robot, where we said, guys, we have all this data. Can you please go in, read it, deconstruct it, auto populate campaigns for us, right, read a value, identify from our naming convention that it has a region, that it has a type, etcetera, and create a campaign and an engagement for this contact with this date. These are the moments that a strict naming convention really helped, because we said, if there are eight numbers, it's gonna be four for the year, two for the month, two for the day, always. Year, month, date. So if you see those numbers, that's gonna give you the start date of the campaign, put the engagements in at this point. So we partner with Magic Robot who goes in and reads information in Salesforce, reads the contacts in HubSpot, converts them, creates campaigns, populates them with engagements, and then runs their attribution where we say, hey. We wanna give this much for sourced campaigns and a little bit for influence, etcetera. Because our attribution model, we care about opportunities being able to be sourced and influenced as well. We chose that, but there are some interesting things that it's important to note about. For example, Salesforce, you have a limit of tracked fields, right, of how many fields have been changed by who and what was the previous value. HubSpot doesn't work like that. Every property has a history that you can go back, and you can also query with an API. So at some point we say, well, there are some properties that we have that are static, but some we can also update regularly from the marketing team, and they will update this value. So read always the latest one or the history of it. So now we have a marketing team that lives in HubSpot. They do their day to day activities, landing pages, emails, etcetera, but everything that they report on is on Salesforce campaigns that they create. So it's been an interesting mix. There's always things that we can improve, data changes, etcetera, but it's an integration that we maintain, but it also has solved us a lot of problems because we've been able to do this automatically with a system. For example, when we reach some problems where we say, hey. Why is this opportunity not marketing sourced? We don't go in and change a toggle to say, hey, from not sourced to sourced. We go in, we find a contact, and we say, where is your engagement to this campaign? And if we say, hey, it wasn't created for this or these reasons, we create that engagement. The attribution model picks it up and says, hey. There's an engagement. This should be part of a deal. So this is what I mean that you also have to have a transparency and robustness in your attribution, because you don't wanna arbitrate things. If you track marketing engagements, then there needs to be a marketing engagement that the attribution model picks up, because that's how you trace