Attribution: What Makes Revenue Happen?

Hello everyone. Hi. Thank you so much for joining our webinar today. Today we're going to talk about marketing attribution. My name is Chris. I'm VP of Marketing at TwentyThree. And with me today, I also have Stine. Yes, I'm Stine. I'm the head of global video marketing meetups. That's a whole other story. I can tell you about that another day. But today we're here to talk about marketing attribution. Yes. And today we also have an amazing panel with us. We have Steffen Hedebrandt, CRO. And co-founder of Dream Data. We have Ari Plaut from HubSpot, Product Marketing Manager, and also Halfton Thiem, partner at Obsidian Digital. But let's get to it, shall we? Let's go and meet the panel. First and foremost, let's hear from Halfton from Obsidian. Definitely. So hi to everyone. So my name is Halfton. I'm working as a partner in Obsidian Digital, which is a Danish performance marketing agency. We actually do have an office, but in order to make sure that it's quiet and that you can hear me, I've set myself into our recording room. So that's why it looks a bit messy in the background. But I look forward to share a lot of insights about attribution, which we work with for a lot of our clients in Denmark. Let's also say hello to Ari, Product Marketing Manager, as mentioned at HubSpot. So let's just go over to him. Hi, Ari. Hey, guys. Thanks, Fina and Chris, for having me. My name is Ari. I work at HubSpot as a Product Marketing Manager, which means that I bring our products to market. I figure out how to position and launch our products from the marketing side. And so as you can imagine, both myself as a marketer and also the tools that HubSpot builds focus a lot on attribution. So excited to be here. And finally, let's switch over to Stefan, CRO and co-founder at DreamData and have him also introduce himself. Yes. I'm pleased to be here. Yeah, today I'm a co-founder of DreamData, which is a B2B attribution product. So I know a lot about the pains that especially B2B people feel. My background comes as a head of marketing and B2B hardware company as well. So I really, really felt all these pains firsthand myself before and looking forward to sharing all these stories today. Thank you so much, Stefan and to the rest of the panel. As mentioned, super excited for this webinar. I think we can just basically jump into it. So today we are talking about attribution and we're going to talk about why is attribution important? How does the different models work? How can you start to understand attribution and what actually makes revenue happen? And finally, also touch base upon whether or not there's a difference between B2C and B2B. So let's just touch base upon the first question, which is how would you actually define it or how would you describe attribution to your family, to a tech savvy person? And first up, I think we're going to go with Halvdan on this one. Halvdan, could you give us your opinion on this? Definitely. So marketing attribution for me is a matter of properly attributing value to all touch points in the given customer journey. And this is something that has become ever more important in the last few years because when I started working with digital marketing 10 years ago, I think it's much more simple today. Like you had one device, you didn't, you maybe had a phone, but you didn't use it that much to browse with. So the device that you did your research on was the same one that you bought the product on in the end. You didn't have all of the social media that we have today and all of the different channels. So the customer journey had less devices, it had less platforms, it was probably quicker in some cases, and it was just easier to attribute where did the sale come from. That has become a lot more difficult. So in a customer journey where you have both Google and Facebook and email marketing, display advertising, programmatic, everything involved, which of these actually attributed to creating a conversion and which were just there by accident, that's marketing attribution to me figuring out what actually brought value to the table. Let's also just jump over to Ari and have him answer the question, what is marketing attribution and his view on this? Yeah, I think that was a perfect definition for the more technically savvy for the marketing experienced. I was thinking about this in the context of my family Thanksgiving week in America. I was thinking how I would actually explain attribution to my mom. I would ask her what she bought last. She'd probably say she went on Amazon, she bought an umbrella, and I would say what are the things that you did before that? She said I go to a page on someone's website, I went to their Instagram, I clicked an ad on their Instagram, I opened up their email newsletter. And so then all attribution is that the company that you bought this from is trying to figure out what helped people like you to buy over time. So that they can make your experience even better the next time. So that's how that attribution reporting is from the I think the less technically savvy explain it to your mom angle. I think that's a really good definition. But maybe Stefan you want to give it a final flair and give you a shot? Yeah, so I think the two guys did well. So I'm going to take the extra technical level for me because today now, I'm going to be talking about the marketing experience. Attribution is very much a question of getting a technical infrastructure right to measure every single digital touch point leading up to a purchase, which that's what we're trying to do at Dream Data. So the thing that Ari said about his mom, we're trying to put put an infrastructure in place that can actually measure every time a little piece of the puzzle is added. We will try to understand what was it and what happened. That's also a good definition, I think Stefan and somewhat of a technical one. But it's good to have the whole spectrum from the maybe a little bit more broad definition to the more technical one to get around this subject. Now, in terms of attribution, there are quite a different number of models. If we just take a look at it quickly on attribution, the question is like, what are the most commonly used models and what should they be used for? Because we have so many different models and all have maybe their justification, but in its essence, they are very different. So let's talk a little bit about the different models and what the most common use cases are. Let's turn over to help them first and foremost, maybe to give a bit of a flair on this one. Definitely. So one thing that we've noticed for our clients, and I think this is the case with a lot of Danish companies at least, is that the most commonly used is probably also the one that's most off, which is the last click attribution. So last click attribution is the standard model in Google Analytics, whereas you attribute 100% of the value to the last touch point in the customer journey. The reason why we view this as particularly wrong is that we're, as an agency, very focused on growth. And last click as an attribution model is not very good at showing where the growth comes from, because this shows you what was the last channel they used. And we really want to know what's the first channel that we use, because we want to add on growth and we need to know where the customers came from to begin with us so that we can end up spending more here. So we see last click being used a lot. We would much rather use first click or some other models at least. Stefan, do you want to comment on this one? Yeah, so I very much agree with health and Google really introduced this nice way of showing that you actually got some money back from the money you gave to Google with their last click model. But the truth of the matter is that the demand do not start when the credit card transaction is completed. It starts somewhere else. And that's what we as marketers are looking to understand. My own experience as like a head of marketing was that in the true black and white scenarios between last click and first click, I was always more interested in the first click attribution because I wanted to know where did our customer journeys start? Because like where did they close? It's not really for a person who's living off delivering leads to the salespeople. Then you want to know where the journey starts, not where the journey ends. It makes perfect sense because basically with attribution, it's all about figuring out, okay, what is driving revenue? And there are multiple touches on this journey, most likely at least, both for B2B and B2C. But Arne, maybe you want to touch base upon the other models than the first and the last click and give it a little bit of flair maybe on which other models are out there and why we should be looking maybe also on other models than first and last click. Yeah, for sure. For sure. So the way that I'll go back to the happy to get into more specific technical models in a second, but going back to explaining it to my mom because that's what I have in mind. The way that I think about the different models is that like let's say that you're the coach of a Olympic track and field team running team and you have a relay race. You have four people running a race first leg, second leg, third leg, fourth leg. As a coach of that team, you want to be able to optimize every person at every position. So if you just gave all the credit to the person who ran the anchor leg, the closing leg, you'd miss out on the first three. Or if you just gave all the credit to the first person, you'd miss out on the last three. You can give credit to different people in that relay race depending on what you want to optimize for. So I think that's first click last click. Neither of them tell the full story. I think it's even short-sighted to say that any kind of like a full path or multi-touch model tells the full story. You kind of need all of them and it depends what question that you're trying to answer. No, no, I do agree with you, Ari. But obviously, we have described a thing on the previous night, a few different models, I think seven or eight just on that slide. Right. So should we as MaruTES be using all of these? Are there some that are less useful than others? What should we rule out? Maybe do you want to jump in on this one as well? Yeah, definitely. So I think this is really a question of what situation your company is in. So we're working with clients, both who are very aggressive, like young companies and startups who really want to add growth sort of like at all costs. We're also working with companies that are a bit older and who maybe have a very high marketing spend already and who are not interested in adding on marketing spend or adding on growth, but maybe it's like making their business more lean and getting even more out of their ad spend. So the proper attribution model to be used really depends on the situation of the company. Because if I had a company where I wanted to diminish the ad spend by 20 percent, but keep the remaining conversions, I would look at the very conservative attribution models like possibly position-based and definitely last click like what brought the user in to begin with, even though that meant up exhausting my upper funnel. But if I was a young company like a startup or a SaaS company really looking forward to add growth, I would go with a very aggressive growth oriented growth models attribution models like first click would definitely be one of those. So the proper attribution model really depends on the situation of the company as well. Like are you there to add growth or are you there to make a more profitable business right now? And the answer probably depends on that question. So it depends is the answer and to a certain extent it obviously always does. But maybe we could take that and play a little bit with that scenario because it obviously depends on how your user journey typically looks like. But I was thinking, Steph, maybe you could elaborate a little bit on the user journeys that you're seeing in your daily work both now and in the past and give it a little bit of an aspect in terms of how you've leveraged attribution models in the past in this regard. Yeah. So nowadays we work with a B2B focus and their time is a major player. And like it gets me more weary towards the first touch what actually made revenue happen. But I think it's very important that you actually do look at all the attribution models because you're looking at a life cycle and you want to understand what pieces of content is good at beginning journeys, what pieces of content or like websites is closing your journeys and like each part of your like you can say your online ecosystem plays different roles. And if you don't understand the roles of these pieces, then you might risk shutting down stuff that actually works well. I can jump to a quick anecdote from me and my time at Airtame where me and Halfdan wanted to be super data driven. And we started to look at, okay, not closing any deals on mobile phones. All the purchases happens on the desktop. And then we shut off all the mobile phone ads on Facebook. And there was no problem for the first week or two. But then suddenly you would see like the demand just going really in the wrong direction. And that was because we didn't understand the correlation between like the mobile phone was probably there where people were browsing, searching a little bit information, just browsing Facebook, then getting the seed from the website. And then at a later point when they're ready and understand the product and they open the desktop and purchase. Really good point. I just want to turn a little bit back to the original question on. Sorry. And that's for the whole panel on what model do you actually prefer? We're just going to go quick round across the panel here. Single or multi-touch attribution model? What do you prefer? 30 seconds. Which model and why? Let's start with you Halfdan. All right. So if I had to choose one, it definitely be if possible, definitely multi-touch or data driven. Because I think data driven is sort of like the ultimate one. And I'm sure we're going to talk more about that later. If I had to choose one from the static models in Google Analytics, I would probably go with first click. Because in our company, we work with a lot of companies focused on growth. And that's where we see where people come from. And maybe over to you Ari on the same question, which model do you prefer? I think just that was a good point Halfdan made about if it's possible. Multi-touch requires deal data and marketing data to be connected really well and process to be seamless. But in an ideal world, multi-touch and something that encompasses a lot of the buying journey. So either like a full path or a kind of a W model that takes the biggest touch points into account. And maybe over to you, very short Stefan. Okay, I'll lean towards something that attributes money to like where journey starts. That's how I would say it. Our W shaped model in Dream Data is where did we see you the first time? When did we get your mail the first time? When did you buy? Those three steps get set at 25% each. So we try to do like an even split of value. But it depends on the role you have and what the goal is that you want to achieve. All right. I have a question coming up guys. It's from Anders. He's asking is the real challenge really the different models? My concern is more about acquiring the right data and even trusting that data. Maybe Halfdan you'd be up for answering that question. I think that's a great question because the main issue I see in terms of marketing attribution is actually not the attribution part as Anders is talking about. The main issue we see is the marketing measurement part actually getting the data to begin with. And there are several issues to talk about here. One would be the integration between online and offline data. One would be making sure that data is flowing from all channels and systems. Another one would be new proposals like ICP from Apple incorporated in the Safari browser that makes it more difficult for us to track. I mean, there are so many once we have the data in it's relatively easy to sit and adjust the models and find out what what had been working for us. But getting the data in to begin with now, that's that's kind of a hassle. Stefan, you also raised your hand. Yeah, like all attributions only as good as the data you do it upon. So if it's the wrong data, the calculations that you end up with still going to take you to a wrong place. But I think the biggest shift that we will see in marketing the coming years is that people will start to want to own their businesses own data. Meaning that right now a lot of companies rely on Facebook storing the data or Google Analytics storing the data. But after nothing, you know, is it like after 180 days Google Analytics just start sampling data rather Facebook start sampling Facebook deletion data after 180 days and Google Analytics start sampling it. And I'm not sure if it's on volume of 500,000 businesses. I definitely agree on this. And I also think that we're going to shift towards using systems that integrate a lot of features. And I think HubSpot would actually be an example of this because if I have all of the data from my email campaigns and my CRM, I bought in one system, it'd be easier to integrate that and see what's actually working. It's a really good point Hansen. If you just quickly turning back to the different models here, we basically see an outline of what models performance marketeers are currently using where we see the majority of people actually using multi-touch models out there. Secondly, it's first click, thirdly position based and all the way down to fourth place is last click models and beyond that. Ari, does these numbers surprise you or do they go in line with what you're thinking? My first response is it's a little bit of a misleading set of data. I think because multi-touch all it means is you're giving credit to a few different touch points along the way. Position based, there's a couple of different position based models, which are multi-touch. Time decay is a multi-touch model. Linear is a multi-touch model. So multi-touch is kind of this superset that includes a bunch of these. But yeah, I think like the general trend that I think is a good thing and like to think that HubSpot is part of that trend is that attribution is getting easier over time. That like Hefton was saying, there are systems that you see a kind of consolidation, whether it's HubSpot, whether it's Salesforce 360 or the different like kind of big players out there pulling data together in a nice way that makes it easy to access it to make a multi-touch model. Whereas before when it all lived in completely different places, all you could really do was was a single touch point. So yeah. Yeah, and I think one thing that is also going along with this trend is just the rise of customer data platforms as well. Where like if you use HubSpot and use all their services, then HubSpot is really good. But a bunch of companies like have a like a tool stack of several different tools where you can use these customer data platforms to kind of glue the information from all of the tools together on like one user. And I think that will grow extremely the next couple of years. Before we move on to the next question, also just want to highlight that to the audience that if you feel like it, please do ask question on the right hand side. We'll tackle it either along the way or in the end for a Q&A and fireside kind of session to make sure that every question goes unanswered as much as possible. But just quickly over to the next question first, Stine. So a question or a commentary almost you could say. So we're really curious on your story in regard to attribution. When did attribution make a big difference for you? When did you have a hard moment or even better a solid revenue impact? You know, suddenly starting to leverage attribution better. Maybe you want to go ahead first on this one. Definitely. So we've seen the same story play out 100 times by now. And the case is that almost half of what we do at Obsidian Digital is advertising on social media, Facebook and Instagram in particular. And we've seen the same problem play out all these times. And the problem is that Facebook have a different attribution model than Google Analytics have. So Facebook's attribution model is that all conversions that happen within a 28-day click window and 24-hour view window are being attributed 100 percent to Facebook, which is a very aggressive model and not one that I will consider the truth. This certainly gives Facebook more conversions than they actually helped create, especially if you look at incremental conversions. But the issue is that if you look at the data in Google Analytics, which a lot of our clients do because Google Analytics in Denmark and I think in most countries, have this really dominant position because it was the first proper free measurement tool online. So this really dominant position. So people are used to working in Google Analytics and Google Analytics is a cookie based platform and you cannot change that. So when we generate a lot of top foreign traffic with Facebook ads, especially on mobile, as Stefan mentioned before, and our clients were to measure the impact of this in a last click cookie based system like Google Analytics, they did not see the value that was actually being created. So what they saw was that in the period that our Facebook advertising ran, the conversions from direct traffic were rising a lot and the conversions from brand Google search ads were rising a lot. But in reality, if they had measured, they had seen that the Facebook ads led to people Googling something or led them to go directly to a page and then convert. So we have shut down projects when we did not know attribution was an option years back, which we thought were profitable looking at Facebook, but we could not see the profitability looking in Google Analytics. And that was because Google Analytics as a cookie based tool just cannot see this. So we actually did the mistake of shutting down profitable campaigns when we did not know this. Once we started actually understanding this and using people based measurement or using a better cookie based measurement, not only relying on Google Analytics, we were able to properly adjust spend to the campaigns that actually brought conversions. And this has made a significant difference for so many of our clients. So not using Google Analytics in the standard version with last click attribution, only cookie based can make a huge difference. Interesting story. I'm also curious, Ari, on your side, you know, Hubsan, big company, you obviously both do a lot in terms of attribution yourself, but also consult with a ton of customers and prospects on the subject. And do you have a short story that you want to share in regard to attribution and where it actually made a huge difference? Yeah, for sure. I think just like you're saying, we use it. Our customers also use it. The one that comes to mind is I just chatted with a HubSpot customer yesterday. It's a three or four person marketing team based in the Northeast in the US. And maybe it's a 60 person company, something like that. And they I feel like we've all seen that slide. That's like, you know, what's the what's the quote? You spend you waste 50 percent of your ad money, but you don't know which 50 percent. It's like the most popular slide in marketing decks ever. But John one of me got quote half my advertising is spent, but I don't know which half. That's right. Yeah, I feel like it's that one. And then the seven thousand marketing app slide, those are like the two most popular slides in marketing decks. But anyway, so this was coming true for this company. They didn't they were they knew marketing brought value, but they didn't know how to prove that. And as a result, they hadn't grown in a couple of years. They hadn't added any headcount. They hadn't added any budget. And they what they did, which was interesting, is that they they ended up. So sales sales get celebrated everywhere because sales has a revenue number. And when they hit the revenue number, everybody knows what this company did was actually assigned each of their four marketers a revenue number in the same way. And so at the end of a given month, they could see that they were using HubSpot, but you could use any tool. They could see how much revenue that marketers campaigns had influenced. And so at a high level, it was on the marketer level. But then zooming in, they could do it on an asset type level, like on landing pages versus blog posts. And they could do it on a channel level as well. Email versus social media, et cetera. What that ended up doing, they realized that they they doubled the marketing influence revenue year over year. It got them a seat at the table with with sales and with leadership. And they were able to think they've added at least two headcount for next year to grow their team by over 50 percent and a whole bunch more budget. So that in terms of specifics, they used a few different dimensions and they use mostly the W model on the full path model because they wanted to show the influence on revenue all the way from the first click all the way to. You know, the things that happened in between the deal create and the deal close. But yeah, that's the most recent one that I think stuck out to me. All right. Thanks for that. Ari, I actually have a question coming in here. Should add impressions be attribute value or only click? Is there any of you who would like to answer that one? Hes and I see you are nodding right now. Definitely. So I think this is a great question. I was I was having a hosting a workshop with a large Danish company recently and I was looking through their programmatic advertising and the programmatic advertising platform had a standard attribution window of 90 days click and 90 days view. So they were attributing attributing a lot of value to the tool. But when we looked at a data driven model and Facebook attribution, we saw that fish were attribution estimated that it only brought 0.2% of the value that the programmatic tool estimated. So there was a 99.8% difference in the trivial attributable value. And when you I think that views have a value, but I think that the value is diminishing quickly. So I would attribute views maybe within a 24 hour window or maybe within a 12 hour window. But some tools have a five day, 28 day or even 90 days standard attribution window for views, which I think is crazy with the lack of a better word. Actually, now that we've been talking about this for a bit, I think it's time to put up a poll. Are you ready to try it out? Yes. Yes. So what attribution model do you prefer today? Let's go. You should see the poll on your right hand side and then please do feel free to answer it. So we also kind of get an idea of where you stand out there. It's not just our view on the world, so to speak. But while we wait for the answers to come in on this one, just want to move a little bit along here and talk a little bit about another topic that that's close to heart in our on our side. At least it is basically whether or not you should mix and match when it comes to your website and the actions that that happens, happens there because one thing is campaigns, etc. But you also have a ton of assets and touch points on your website. It's not just happening outside of your own domain. It's actually also happening on your domain. You know, there's social, yes, but you also have your blogs, your landing page, your emails, your videos and your pricing pages, etc. And the question is, should we add to the value to what happens on your own website along the user journey in these kind of models or should we keep the two things very separate? So let's hear it from the audience. Sorry, from the panel on this one, maybe Stefan, you want to go first on this one. Should we mix and match? Yes, so this is something where you might want to be really specific about what you say because what because like are we discussing like channels now? Are we discussing specific pieces of content, etc.? One thing I would say before you at all able to do any of the things we're talking about here is that you need kind of a customer data platform that's able to identify you and keep storing every time you take an action. Meaning you view a page, download an e-book. I know 23 as well can give data points on how much of a video did you watch, but without like a customer data infrastructure to capture this on each user, you will not be able to do any of it. Ari, maybe you want to give it a little bit of flair as well. Should we include one page activities? Yes or no and why? Yes, the answer is yes, you should include them. I think the answer of mix and match is a little bit tougher like kind of like Stefan was saying in HubSpot, at least the way that we think about it. There's you can break down your attribution based on certain dimensions and that dimension could be channel. So you could give the social media channel X amount of revenue credit versus the email channel. It wouldn't make sense to include on page attributes when you're breaking down by channel. So in that sense, I would say mix and match. Maybe not, but we also have in HubSpot, we have a dimension for like an interaction type, which is that interaction is what was happening on the website that's tracked through the HubSpot tracking code in our case. And that could be a form submission. That could be a video view. That could be a meeting booking. That could be a live chat interaction. So people spend a lot of time working building websites. It makes a lot of impact on the way to close. So I think it's it would be foolish to ignore that element, but maybe not mix and match depending on what model you're looking at. I don't know if you all can see it here, but and thanks for voting, everyone. We can see that first click and position based is equally popular basically. And after that comes linear and last click. Is there do you have any comments on that health then? Does that surprise you or is it does it make sense for you? Well, I'm happy to see that the most growth orientated models are first click and possibly position based have a high position as well. The only thing is that actually I've done a test on a few accounts where I have compared the numbers from Facebook attribution when you use a data driven model to all of the static models. So I've compared data driven on accounts with a lot of data where I would assume that the data driven is sort of like the closest to the truth that we have. And I've compared the two static models and equal and linear as attribution models are the ones that are performing the worst compared relatively to the data driven model. Some also have to see them in the bottom for that reason. Yeah, I have to maybe you can say a little bit about how much data you actually need to do a data driven model because unfortunately the inconvenient truth about it is that it's not that many companies who have enough data to use this model. Oh, yeah. And I think this is a whole question because that is a lot of companies does not have enough data to use a data driven model. In Facebook attribution, at least there is no minimum requirement. So you can do a data driven model with 50 conversions per month, but you shouldn't do it because the numbers will obviously be wrong. If there is sort of like a rule of thumb, I do not know what it is. We tend to not look at it unless there is a few thousand conversions per month. And even then, I'm not sure if that is too low. For those who are newer to attribution on the line, can you explain what a data driven model is? Definitely. And that's a good idea. So a data driven model would be allowing a machine learning algorithm to create a model based upon what it thinks actually broad value. So the idea is if we try and simplify it, that if a lot of people saw like a Facebook ad as an example, and the people who saw that Facebook ad did not convert more than people who did not see that Facebook ad, then a data driven attribution model would not attribute any value to it. So if it had no real impact on conversions, if it did not bring incremental conversions, it will not be attributed value. So instead of looking at did you see the ad and then buy, you're trying to answer the question, did the ad make you buy? Because there are a lot of advertising platforms where you get a lot of views. And I'm thinking of programmatic in particular here, where those views may not actually have been proper views. You didn't really notice it on the page. And if you attribute values to those, you might end up getting too high a number. I know that HubSpot actually have a predictive lead scoring that I think works a bit in the same way. Is that true, Ari? Yeah, we do have predictive lead scoring. We haven't yet worked out. One thing we're still thinking about is how to bring that predictive algorithm into attribution reporting. So that's why I think it's important that people understand it. We actually don't have a great solution to it right now. But I think it's like you're saying, that's a problem for larger businesses with a ton of data to solve. Most customers, most companies, I think, don't have clean enough data or don't have enough data to get there. But no, good question. And I think this is especially the case in a country like Denmark, where a large company in Denmark have fewer conversions than a large company in the UK or the US would. Like relatively speaking, a large company in Denmark is not very large globally. So having enough data, especially in Denmark, is kind of difficult. So if we look at our client base, I would say that 95 percent of the companies that we work with do not have enough data to create a data driven model. Only five percent do actually. And that's a problem because we end up using static models, which may be true and may not be true. So I actually put out another vote. Is attribution a priority for you today? Please, everyone, give us your votes. But I can already see that yes is very popular. It is. Which is good. I just want to go quickly back on the subject that we were on before. So we were talking a little bit about on page activities before. If you want to talk a little bit more on this subject, maybe over to you, Ariana, a second. How do we best include on page activities in one of our models, if not the model? Because we were talking about before splitting it up with models, and we were talking about how to do that. And I think that's a very good question. One of our models, if not the model, because we were talking about before splitting it up, which kind of makes sense in some cases. Right. But how do we go about this? I'm happy to answer, Stefan. Did I see you also raise raising hand? Wanted to get on that. You can jump. Okay, start. But I think what you want to do with this one page analysis or what you want to call it is, did you want to connect whatever takes place with revenue? Meaning that you don't want to be counting amount of leads for the sake of that. But you want to be able to see if people watch a video for five minutes, are we making more money? If they download this white paper or they sign up for a demo call, is there a difference in quality of the conversion compared to revenue? So you want to get the system in place and know Hubspot does this really well, where you do a certain activity on a web page and you see that in the journey towards or if you made money on this, made revenue on this person, plus this part of the equation. Ari, do you want to comment quickly on this before we change the topic? No, I think that's a good point. I think you're at the nail on the head. I think it's about the key there, being able to connect things in the right way. So like, are the assets on your website actually talking to an attribution system? Whether that's stuff in software, whether that's HubSpot, whether that's 23 for video attribution. The more of those things on the website actually tie into whatever you're using for attribution, the easier it'll be to connect them to revenue, which is the most important thing. So I want to change the topic a little bit and talk and basically look ahead towards the future and ask the question, is it going to be more difficult in the future? Is it going to be easier? What are the trends out there as you see it right now? I'll be short because I know that both Ari and Stefan have a lot of input to this as well. Let me start with the challenge. I think that attribution is part of becoming more challenging due to more people using ad bloggers, which is blocking the scripts that are used to attribution and the implementation of ITP, Insilient Tracking Prevention in the Safari browser is definitely making it less accessible. Data is less accessible and it's more difficult to create a proper attribution model because you simply lack data. I think a few things are helping us. The introduction of more tools like Facebook attribution and DreamData is helping. Also the consolidation of tools. Instead of using five tools, you may use HubSpot and have all the data and that system helps. But certainly I see clouds in the horizon and they are a bit of dark, I think. Stefan, you also had a comment right there. Yeah, but I very much agree with Halsten. But my opinion would be that there's two major trends. There's Apple pushing towards more privacy, which is good for us as consumers, but bad for us as business people. The good thing is that almost every system today, like online system generates data. So I think we'll see a rise now of platforms that's able to make sense of all these data where you can aggregate the data and start to make sense of it. Back in the day, you had newspapers, you had TV and then you had radio ads. That was pretty tough. Nowadays, there's a lot more digital touch points to everything that goes on. And to me, that does speak to that. It will become easier to make sense of. Let me just say attribution is not about getting to 100 percent of the truth, but it's maybe about going from knowing 30 percent to knowing 60 percent. Because that will give you a data set that when you do act upon this data set, you make more revenue. Just as the example as Ari told about this company that started to understand attribution and how the whole journey connected, they were able to grow their business by going from knowing 20, 30 percent to knowing 60, 70 percent instead. So, Ari, the same question basically goes to you, but also maybe you can touch base upon what are the trends that we also should be looking out for? We are approaching 2020 really fast here. So maybe also touch base upon what should we look out for in terms of 2020? Yeah, I think the two competing strands like these guys are saying, it's getting easier to do technology wise. It's getting more difficult to collect data because of Apple, because of ad blockers, and also I would add because of regulation. The GDPRs of the world in 2020 specifically in the States, California is enacting a similar lot of GDPR that makes it harder to keep and store and collect data. It makes it more restrictive in terms of dropping cookies onto websites and onto browsers. So I think those we still haven't quite figured out what what that future is going to look like. Like, what is what is a cookie list world look like? I think that's like a question that we in the tech industry and we in the marketing industry need to grapple with. And I don't think we've solved it. That said, as we still have access to cookies, tools like Dream Data, like HubSpot, like, you know, CDPs like Xsegment make it easier to collect and aggregate data and report on it. But that's the trend I'm looking at is how much data will we actually be able to get at? So another question from here is basically around as we touched based upon in the beginning of the webinar, B2B and B2C, what are the differences? Obviously, you know, a number of conversions is key. And I'm sure Halftone is going to talk about that as well. But we're going to go a little bit fast both for Stefan and Halftone on this one. Yeah, so there are two major differences as I see it. The first one being that B2B in general, as you mentioned, Chris, have fewer conversion and that makes attribution more difficult to work with. The second one being that Facebook is really good at B2C because the people the person considering buying a running shoe is probably the same one that's going to end up buying it in the end. But that's not the case with B2B attribution. So you have all of these different stakeholders on a lot of different channels and a very long path to conversion, which makes people based tracking not usable. So we have a lot of advertising on Facebook for B2B companies where we can see that we're creating a lot of value, but Facebook realizes none of it because the person buying the product was not the same as the one who reacted to our ad. So in terms of attribution, it's certainly much easier for us to work with B2C company compared to B2B company. Super cool. And what about you, Stefan, about this question B2B and B2C, the differences? Yeah, it's I would say B2B first of all is much more complex. There's easily like 6-12 months. So when the ad platform set their cookie, it will get deleted somewhere down the line of these 6-12 months and then tracking is lost. That means if you're not able to connect where you paid for the demand and the guy who come or that guy girl who comes with the credit cards and buy, you would stop a lot of growth activities because you can connect that journey. So which means like a lot of B2B companies are growing much slower than they could if they're able to connect like understand that this is one company who's starting a journey here and then they're like finishing a journey there. So I think like if you're a little bit, you can a little bit broadly say everything we've said so far only applies to the B2C scenario because like it gets so much more complex when you add multiple stakeholders on the customer account. Now, I think we're getting ready for a Q&A actually to the experts now that we have them here. Just one thing more thing to the B2B scenario. I think it's very important also to think as a holistic business when you're doing a B2B attribution, meaning marketing can start the demand. But like in the air time, it was 95% of the big deals came through sales. So sales play a huge role and you're not able to like measure that in any way or like support fixing stuff Q&A is coming in. Then you like with attribution, you want to say I have a thousand dollars. I want to spend a thousand dollars for the biggest impact that I can do. I do that by giving it to the salespeople, to the marketing people, the customer product people, customer success people. And if you're not able to see full funnel holistically, you end up making silo decisions that are good in a silo and not like across the board. So just jumping quickly to as mentioned before, the Q&A session. So we have a few questions. I'm not sure we have the time unfortunately to touch all of them. But Stine, maybe you want to pick one of the top questions here. So we have here, what is your top three tips to start V2B attribution? And maybe we can just because we are unfortunately running out of time, we can maybe take one. That's a good one. Halton, you go. Yeah, so making sure that you measure as much as the customer journey as possible would be my best tip. And you can't do that with Google Analytics and you can't do that with Facebook Analytics and the Facebook Pixel. You need other tools to do that. So Ari mentioned the CDP like segment before. I would definitely look into a solution like that in order to make sure that you both own the data and that you measure as much as the customer journey as possible. Because as we mentioned earlier, you can create the perfect attribution model. And if you don't have any data to apply it on, it's basically worthless to you. So start by defining all of the possible touch points and making sure that you track those and you have a system that can see when a conversion was made so that you can attribute the channels that brought that person in. Ari, do you want to go with that? Yeah, sure. I was talking to a customer in the UK this morning and their words were attribution is more about the process than about the tech. And I think that reflects what Halton had said that with HubSpot, we have a CRM and a sales tool and also a marketing tool. And the big alignment we try to get is like in order to get any kind of revenue attribution for a B2B long sales cycle inside sales process, you need sales team to be consistently following up with leads to be associating them with accounts to be associating them with opportunities or deals to be closing them to be putting them through stages in the right way. And if that's even if there's two out of the six sales reps aren't doing that, like it ruins the whole thing. So that's not to put it on sales or on marketing, but making sure that both those teams are aligned and that the process and the data is consistent. It's 100% the first step. Thanks for that. And Stefan, do you have last thing to add? Yeah, so just to recap, like start storing and owning all your data and train everybody to do do stuff that gives a digital reflection. Like if you call on your own phone, switch to like AirCall or I don't know, can you call through HubSpot as well? Like do generates digital touch points. You could through HubSpot, but you could through AirCall, you could through Trilio and integrate it back into HubSpot. But either way. Then moving a little bit to the tactical level, I would think we'll see a big shift towards a microconversion just like picking up an email instead of trying to push the sale. Because like when you get the email, you can start associating a lot of stuff. So a lot of businesses will start to try not to sell, but just get the email because then they can get back to the same people over and over again. Amazing. So we also have another question. We have actually a lot of questions. We're not going to touch base on all of them, unfortunately. But one of them is which tool do you prefer to measure attribution? And just to give it a little bit of a twist here, maybe a tool that is not your own. So we go through all of the talent here. And obviously you are more than welcome to also throw in a 23 or Dream Data HubSpot mention as well. Who would like to go? Sure. So I'll start. If you're a B2C company, I would use Facebook Attribution because it's a people-based measuring tool and it's free to use and it actually works pretty well. If you're a B2C company, I would try and following up on the last question, I would set up a measurement that allows you to recap as much as the customer journey as possible. And there are several attribution apps that are able to integrate into segment Dream Data included, but not the only one that I would try and mess around with in that case. Yeah, so I had to struggle to find an answer, but I think actually the easiest thing you can do is to ask people, where did you hear about our product? So like every time you sell something, send them a mail, just ask where did you hear about this product or this service? Or like, can you remember where you heard about this thing the first time? That would actually give you a lot of insights to what's working, what's not working. Yeah, or do you have a comment on that one? Yeah, I mean, it's a good question. I think it depends what you're trying to do. The one I think that's most commonly used, that's not HubSpot in our space, is Visible. And I think Visible is great at tying in data points. I think it tends to be a little bit on the kind of complex side for most of our users, but I think it's operated independently for a long time. They're under Marketo and Adobe now. And you need to use Salesforce to make it work. That's the other thing. Yeah, okay, cool. At high level, it's a good tool, but it's limited in a lot of ways. All right, I think that's actually it for now. Thank you to everyone who participated. But before you go, we've just put up another poll asking you whether or not you enjoyed the webinar. Hopefully you did. Hopefully. But also do feel free to send us feedback, questions or even good ideas for new topics in the chat or in the questions area. But that being said, thank you so much for attending and thank you so much to the panel. Yes, thanks to the panel for attending as well. Hope to see you at our next webinar. Yes. Bye bye.