#233: Attribution in B2B - Why It's Broken (And How To Fix It)

Speaker A: G'.

Speaker B: Day. It's George from the B2B playbook. If you're a B2B marketer and you're stuck spinning up, um, some random campaigns based on whatever the business asked you to do that week, or you're chasing low quality leads and you're watching sales take credit for business that you help create, but instead you're stuck justifying your activities to leadership. Well, we made something for you. The B2B incubator is our, ah, demand generation program for small in house B2B marketing teams. It's built around our 5B's framework and it gives you the strategy, the templates and the system to finally take back control of your marketing and actually align with sales. And we show you how to use AI the right way, not to do more of everything worse, but to know exactly what to focus on and to do that better. It's a top rated demand generation course. We've had over 315 graduates, including marketers like Allison from Rivet who went through it and with their team are now driving 70% more pipeline for the business. So if you've got some training budget sitting there that's unused, this is exactly what it's for. Applications for the next cohort close on 26 June, but of course we have our self service program available if that timing doesn't work for you. So next time you take that first sip of coffee in the morning, head to the B2B incubator.com join us and take back control of your B2B marketing. All right, let's get on with the show. M. Welcome to the B2B playbook where we help B2B teams punch above their weight.

Speaker A: We're your hosts, Kevin and George. Each week we break down how modern B2B companies build demand, win deals and grow revenue using our 5B's framework.

Speaker B: So if you work in marketing, sales or customer success and you want to make a bigger impact in your business and hit subscribe so you don't miss an episode.

Speaker A: And remember, revenue growth starts with people, not platforms.

Speaker B: A lot of these big platforms kind of is in their best interest too. For the metas and the Googles, the less granular detail they give us as advertisers, exactly what search terms are driving, what kind of traffic and exactly what actions are being taken on meta, uh, converting that to, you know, uh, exactly the bottom line that you're driving. The more murky that is, the more they can say, hey, give that to us. Because we can see that information with our algorithm.

Speaker A: Just, you can't, we can't be waiting, you know, two years for the data to tell us what channels are working, what channels aren't working, how much they influence each other. Even if that data was right, we really need that directional data from day one and right away to figure out are we going in the right direction in month three, in month six, rather than waiting two years before we change direction. Attribution at the end of the day is an estimate. The word is pretty definitive. Show things like a range or confidence levels rather than single numbers, um, that would really help communicate just how much variability there could be within the numbers that come from attribution.

Speaker B: Welcome back to the B2B playbook. Today we are putting attribution under the microscope and seeing what lies is it telling you, what truth does it reveal and what can you actually build from your blended data. We are of course into season 8 of the B2B playbook. Seasons 1 to 5, we gave away our 5B's framework for revenue growth. We walked you through exactly what that 5B's framework is step by step. Over seasons one to five, season six and seven, we answered a lot of FAQs on what building a demand generation engine really looks like. And now in season eight we are going into how demand gen engines fit into wider go to market strategies and architecture of course. Kevin, for those who are just joining us, they have no idea what a five piece framework is all about. Can you please tell them what it's all about?

Speaker A: Yeah, of course George. So the five piece framework starts with Be ready which is all about deeply understanding your dream customers and making sure you know every single detail you need to about that audience then and be helpful which is a second B, we talk about how to build helpful content, uh, online in particular and that really is the basis for developing relationships and trust with that audience online. And then in B.C. the third B, we talk about how to amplify and focus in that relationship building effort, whether that's through amplification via uh, paid channels or doing the groundwork to work your way into some of these audiences. And then in Be Better, we talked about how to optimize the whole B2B marketing process that you've built up until that point. And finally in Be the Best, we talk about what the big guys do at the beginning of town with advanced strategies, tactics, how you can bring that into your own practices and also how to use the whole framework as a flywheel so you can continue to get value from it and continue to improve.

Speaker B: Very good, thank You, Kevin. And again, listeners and viewers today we are talking about the myths and the truths of marketing attribution. And Kev, it's a very topical one or I guess it's always front of mind for a lot of marketers, isn't it? Especially in B2B where attribution can be a tricky thing. You know, B2C, we have much higher volume, a much shorter window, often between the activity that's happening digitally and then the action that's taken by the consumer to buy the pair of shoes, the handbag, whatever it is. We're normally dealing with just one person. Often there isn't really a buying committee going behind the scenes and saying whether or not, Kevin, you should be purchasing those new Nikes. Do you have new Nikes on?

Speaker A: No, not today.

Speaker B: No, not today. Look, maybe in some families, I don't know if your purchases in your family, if they run by your wife Christine at all, um, if there's a deliberation stage, if there's a budget, I'm not sure.

Speaker A: Of course, of course. Okay, I've got a look in my

Speaker B: house, in my house. Whatever Jordy says, guys. And so there's no buying group at all. So attribution is probably a little bit easier in my household than yours. Um, but the point is in B2C it's much easier. And Kevin, B2B isn't that simple a story, is it?

Speaker A: No, it never has been. And I think a, uh, lot of the things that kind of lead us astray in the B2C world, um, attribution in that context often promises complete clarity, um, that you know, with things like multi touch attribution you can really get to the bottom of exactly what touches uh, really helped uh, get a purchase across the line. Now in B2C it might be a little easier. There might only be a handful of touches before that happens. Maybe uh, influential ones are uh, particularly clustered towards last click. But even in that scenario, but even in that scenario, when you're looking at multi touch attribution, it still often overweights clicks and those interactions that can be easily tracked. It overweights things like, you know, last click as well. Uh, and it really relies on uh, the data that you can collect uh, online, which as we've talked about in the past and is obviously evident to most people in the industry, is quickly diminishing and disappearing, uh, as privacy things come into play. Um, so rarely does um, the modern day multi attribution really answer or answers those promises of clarity that it purported to answer. Um, and Actually, um, and actually, a lot of the activity that influences a purchase, Even in that B2C context, isn't really captured. You know, you've said that Jordi makes the final decision. Maybe you're the one that does the buying, but the decision making could have happened completely at the other end with whatever Jordi was browsing or being influenced by, um, in terms of content that she consumes. So even, um, in that B2C context, uh, it's not perfect. And then when you bring it to the B2B space, it's even less clear. You've mentioned there there's a buying committee invol multiple people. Now you've got multiples of that same process happening across different people, different messaging, different, uh, priorities into making one collective decision for a business. Um, it just makes it impossible to glean any real value from that sort of attribution activity, um, we talked about before. You know, you can get sort of a directional sense from attribution from the data of where, uh, you're putting your efforts and where that's working. And you should still make that effort, um, because it does give you that directional sense of where things are improving. But, uh, as we've said before, you really shouldn't be data driven. You should be data influenced, uh, and, um, data led, uh, rather than 100% data driven.

Speaker B: Yeah, yeah. Data informed versus data driven. And to your point, Kevin, what doesn't get tracked in my B2C purchases is anything golf related. Because for that, I go in store and I have cash. Jordi can't see that there's no record in our transactions. She doesn't know where all the cash is.

Speaker A: Yeah, there might be an audit in the George Kid and Us household after this episode.

Speaker B: No one from the ATO is listening, are they?

Speaker A: No, no, hopefully not. Um, but yeah, I mean, you know, in this episode, we'll discuss, you know, how do you do that process of blending things? Um, why would you blend measurement, um, and how to combine those different things? Uh, maybe things like incrementality, test media, mix modeling, and how as small teams in B2B marketing, we should really start to treat our data. What you can build to build trust in your data without drowning in complexity and just ambiguity in it all.

Speaker B: And, Kevin, this topic's becoming even more important, isn't it? Because of the privacy changes that just keep coming in place. And we definitely see it just heading that direction. Right. I mean, we're all going to be getting less and less visibility over time. I think that's what people Generally want. That's the way the policy is continuing to shift. And also Kevin, a lot of these big platforms kind of is in their best interest too. I mean when you think about it for the metas and the Googles, the less granular detail they give us as advertisers or you know, our clients in terms of exactly what search terms are driving what kind of traffic and exactly what actions are being taken on meta converting uh, that to you know, uh, exactly the bottom line that you're driving. The more murky that is, the more control they can say hey, give that to us because we can see that information with our algorithm. Just you can't.

Speaker A: Yeah, yeah, 100%. I think it definitely will get as advertisers and marketers less and less data over time from these ad platforms. And then when you look at than the other group of big companies, uh, you know, the providers of attribution software, the providers of the analysis piece, most of them are ah, promising multi touch attribution, some sort of attribution that's really just focused on um, you know they promise a lot of sophistication and actionable insights and all that kind of stuff but often when you get down to it it's really just repackaged last click analysis. And because of that uh, it's really just more cost and complexity for not much extra value. Um, and particularly in B2B small teams, you know, we need actionable insights, we need stuff that we can actually rely on, that we can get around um, without the vanity metrics that a lot of these platforms focus on. And uh, we need it sort of before it's too late. In a B2B sales cycle, you know, we, we have long sales cycles, 18 months. We can't be waiting, you know, two years for the data to tell us what channels are working, what channels aren't working, how much they influence each other. Even if that data was right, we really need that directional data from day one, um, and right away to figure out are we going in the right direction in month three, in month six rather than waiting two years before we change direction.

Speaker B: And that's the nature of the B2B cycle, isn't it? I mean they're much longer cycles. It obviously depends on the market. But if people are signing with vendors for two to three years that means it's going to be two to three years before someone is in market. And if your market is an enormous, then there's going to be less and less people in market. So people always ask us Kev how long is it going to take for us to see results? And it always comes back to, well, how much of your market is actually in market ready to buy? And for those that aren't, how often do they come into market? Because that's a real determining factor of the marketing work that we're doing, actually having an impact. Because if you're just doing it in bursts and you haven't been present long enough to make an impact on someone's decision making, well then you know, obviously there's not gonna be a good chance of that marketing activity having any kind of success because you haven't done it for long enough. You haven't stayed in front of people for long enough until they enough have come into market ready to buy and had their mind changed by hopefully marketing. On this show we talk a lot about being remembered in the right moment when your buyer actually has the need. Well, here is our moment. If you're thinking about switching agencies and you're finally ready to bring one on for the first time. We're actually a B2B demand generation agency. We run LinkedIn ads, Google Ads, meta, uh, all the channels that your B2B buyers hang out in. And of course we do the strategic component that is completely aligned with our 5B's framework for demand generation. This is all we do. So when that moment comes, we want you to think of ourselves. And if you're still just figuring out your next move as a marketer, next time you've got a two minute coffee break, grab our free upskilling report from the b2b playbook.com you'll get instant personalized results showing you exactly where you sit compared to other B2B marketers and what other ones driving real pipeline, uh, are doing. We'll also send our best free resources over to you based on your answers. Check it out@the b2b playbook.com okay, back to the show.

Speaker A: Yeah, 100%. I mean that's just a handful of the sort of front end issues you have with, um, where the current state of attribution is and why it's so broken. Such an issue, particularly for B2B teams, uh, when you then think about, you know, the net effect of all that work. Even if you did all that work to do the last click attribution, if it's blended attribution, data driven attribution, what have you, even if you set that cost aside, you still sort of run into the issue of well, will those people have bought anyways? There's no way to tell from the attribution, whether they would have bought anyways, in which case maybe you didn't need to run the activity, they would have converted anyways. Maybe we're uh, confusing correlation with causation and a lot of the attribution tools available just don't answer that fundamental question, uh, which is really kind of important when you're looking at investing so much into attribution, into this activity, even if you're not spending a lot on the platforms. Uh, you know we have AI tools these days to run a lot of analysis. You can pull the data yourself, what's available and try and run it yourself and do the modeling yourself. Um, that concept of multi touch attribution at the moment that most companies focus on just isn't really going to answer those kind of questions around is it actually having the impact, um, or is it just correlation? Um, so in terms of building then trust with um, with maybe blended measurement, which is what we kind of like to approach. Um, the question for attribution within the B2B context, um, we'll give you our framework um, for doing that and really it starts with defining the role of attribution and the way we like to align within an organization. To talk about attribution is to say that attribution is a useful short term optimization play. It's another data source, but it should not dictate strategic budget allocation completely. Um, recognize that as one input amongst many, as we said earlier in the piece, be data influenced, not data driven. This is just one of the things that can influence that decision, can influence the data that you look at and ultimately the decisions that you make when it comes to optimization and budget allocation. Second, we would say you should still run uh, multi attribution, multi touch attribution. If it's available to you, you should do it. You should try and keep it um, you know, as lean as possible. Like we, we tend to, you know, if you have the resources to test different tools for doing that, sure. But make sure to compare with you know, that sort of lean, um, you know, hacked together way of doing it. Maybe it's in a spreadsheet somewhere with a different, few different data sources pulling together. Maybe it's with an AI tool to help you automate that process each week it pulls automatically pulls the data from different sources and does a bit of analysis on the back end for you just to do that. Multi touch attribution and use it as uh, another source of data but make sure to also combine it with things like marketing, mix modeling and incremental test, incrementality testing. So marketing mix modeling, uh, that really is looking at, you know, channels beyond just those that you readily have data for. Maybe you have some data around, uh, your branding activity that you can throw in there, maybe it's email activity that you can throw in there and try and do some regression analysis, uh, with again with AI potentially or something like that, to give you another sense of how the overall contributions of all channels is playing in your marketing mix and how the channels that you're specifically looking at is playing within that mix. Because what it might reveal is that multi touch attribution might say, you know, 90% or 80% of the leads are uh, coming from Google Ads, so you should obviously invest more into that. But if you look at it through the lens of marketing mix modeling, maybe it suggests that, okay, the periods where you're running brand activity, that's maybe a little harder to measure traditionally. Maybe it's your email blasts, uh, maybe it's uh, even a TV campaign or a sponsorship that's not easily trackable. Maybe it's a piece of content that went viral organically. Those kind of things might be timed with the leads coming in. So actually you haven't changed your investment into Google's ads at all, or it just happens to correlate with that pickup because you happen to launch at the same time as this campaign. With all these other channels going live, um, does that warrant 80% of the budget going to Google Ads going forward? Probably not. You probably still want to cover those other things, um, to help make that decision. Once you get those sort of indications of where the direction is, then you can test things like incrementality, um, then you can use things like incrementality testing to help you figure out exactly which ones you should focus in on in terms of tactics. Because you can start uh, another campaign and see if adding or removing a particular channel or tactic actually helps with the conversion data, uh, with the actual leads that you're getting through the door. All this is pretty late on, isn't it George? Um, when we talk about measuring success, most of these things that we've talked about so far is those lagging indicators. So you can only sort of do these things, um, really well and with sufficient data. If it's a shorter sales cycle, if it's things in the lower funnel, in most cases, if you're doing a bigger piece analysis, you probably need to do it at a year, two year sort of time frame. If you're starting a new campaign, perhaps a lot of this isn't as relevant. So you need to also mix in things like qualitative signals. So early on, those leading indicators might be, what do people respond to when they filled in lead forms around how did you hear about us? Maybe you collect sales anecdotes, um, at the beginning of, uh, the sales cycles of what do people ask or talk about when they first come in as a contact. And also make sure to monitor things like DMs and comments, um, that might be a little harder, maybe a little bit more manual, uh, particularly when you're talking about DMs into sales teams, into the CEOs, uh, inbox and things like that. But those are also important because those are the signals that reveal which channels are, uh, working that the attribution process are probably missing.

Speaker B: Yeah, absolutely. Combining it with those qualitative signals is so important. And again, those qualitative signals, literally meaning asking, how'd you hear about us? On all your intake forms, as you've already suggested, sales talking to prospects, asking, how did you first hear about us? They're those leading indicators and you want to keep a record of those. You're not going to capture that information about every single person, maybe 30% or so if you're lucky. But you always want to measure what's being said there versus, as you said, whatever your attribution is telling you. And the truth is always somewhere in the middle. That's the reality. And to your point, Kevin, around correlation versus causation, the other thing to take into account is in B2B, we have, as we already touched on, there's much longer sales cycle. So there's a very real time lag between the marketing that we do now and when we see that impact. You wouldn't expect to see, uh, a lot of impact from marketing activity that you've launched now for, you know, three, six months. That's very generalized. Of course, it could be much longer in some industries, but that's something that you need to keep in mind as well whenever you're doing that attribution. And then in terms of incrementality testing, that's something you and I used to do a lot of when we worked with clients that had much higher volume, especially in B2C. But the reality is probably for a lot of people listening, and we work with some pretty big B2B companies, even the big B2B ones, a lot of them don't have enough data to draw really definitive conclusions, um, based on that incrementality testing. And so a lot comes back to the Qualitative signals, looking at that of course your attribution and then just the answer somewhere in the middle.

Speaker A: So far listeners that process looks like you know really setting um the expectations when it comes to attribution internally and then getting what you can available on the quantitative side um through you know, multi touch attribution, marketing, mix modeling, incrementality testing, all these different things. But as George mentioned there that's probably going to be lacking uh leave you wanting and you really need to mix in sort of the qualitative signals, um, wherever you can capture that. And you know the mix between taking importance of those things is probably closer like 50, 50 or even 60, 40, 74, 70, 30 towards the qualitative side rather than the general expectation that you're going to have really good quantitative data. It's like 80, 20, um towards quantitative that's just not realistic. And then from there you really need to focus in on communicating that uncertainty. So when sharing reports that cover both qualitative and quantitative, uh, with leadership with um, the rest of the business it's really important to highlight that attribution at the end of the day is an estimate. The word is pretty definitive um but really in the marketing attribution world it's really just an estimate. It show, show things like a range or confidence levels rather than single numbers um, that would really help communicate just how much variability there could be within the numbers that come from attribution. And then as a last piece there in the process really try and align with cross functional teams, align between your sales team, CS team, finance on the definitions of success and frame. So all those things around leading lagging indicators, what they actually mean in terms of metrics that are being reported and also uh, having a clear understanding of um, what attribution brings to the table that should create some guidance directional um help rather than pure complete clarity or 100% clarity.

Speaker B: Yeah yeah. Those leading and lagging indicators and separating what tells us that we're on the right track versus lagging which is looking back how do we know that there was some success in what we did? I think they're important to distinguish to whoever that you're reporting upwards towards. And probably the only other thing I would say is don't you don't set yourself up for failure by over promising. I feel like Kevin, as marketers we face so much pressure because everyone thinks marketing is the on switch for the business. We know that's not the case especially in B2B that it's a coordinated effort between marketing and sales and You've got to really work with at the business if you want to drive a solid outcome and don't promise the world and promise perfect attribution to whoever it is that you're reporting to in order to justify that marketing is working. Um, that's just something that I think is important for marketers.

Speaker A: Yeah, very, very important. I think, um, you know, when, when we go into those conversations about attribution internally, it's often, well, we just want to see what's influencing things. Here's like, here's, you know, the list of requirements. Give us the numbers, and it's sort of like, okay, but we need to have a conversation first about setting those expectations around how reliable are those numbers actually? And what is possible? Ah, you know, can we get to 80% accuracy? Can we even get to that, let alone 100%, you know, it's not as simple and click as maybe what finance imagine it would be that all leadership would think it is. It's not a perfect world. Uh, and as we've discussed a lot on the show, most of these touch points are not going to happen, particularly the early ones are not going to happen in a space that's trackable. That is just the reality of it. Uh, if you think about your own buying processes, you can explain in that way that you can't really actually track a conversation or recommendation from a colleague. Um, so realistically, you need to be tracking that, uh, in different ways, trying to hit those leading indicators, somehow taking into account that there's going to be a chunk that you just can't measure. Yeah.

Speaker B: As we always say, Kevin, if attribution or perfect attribution were everything, we would have stopped this podcast a long time ago. Maybe we'd still be doing it for fun. I still find it very fun.

Speaker A: Well, it might have been about 10 episodes about how to do attribution, and then we'd, uh, just run through tactics about how to refine it. But, yeah, unfortunately that's not the case.

Speaker B: But even Kev, when we speak to a new client, I mean, the number of new clients who say, okay, what. What do you think we can get our cost per lead to? And there's no historical data. They've never run, you know, activity in that particular channel. And they say, oh, are there any industry benchmarks you can go off and you go, well, unless we've already got a client who's like, for like, or you can get a benchmark that's like for like. It's a very big. It depends because, yeah, sure, say we find a client from a similar industry and you find out what the average cost per lead is. What the hell is a lead? How are you defining a lead? You know, like that makes a massive difference. Are they selling exactly what you're selling? Do you make the same amount of money, like off that lead? Is your sales team set up the same way? How many of those turns opportunity with those close, like, we've got to compare apples with apples.

Speaker A: And uh, I mean, even in that case where we are comparing apples with apples, just say hypothetically, you know, how many times do we come across an account where the tracking is reliable, where the conversion data is reliable, where the attribution data then is reliable? We've almost always. I don't think, in fact, I don't think I've had an account where we've gone in and we haven't had to make any changes on the tracking. It's just a degree of difference between, you know, have they just got the basics down? Have they not got the basics down at all? Or, um, have they got any of the sort of signals that you potentially would be interested in down as well? And you know, along that scale, I would say most are not really in the latter two stages, um, at all. So, you know, even if you're comparing life for like, I think it'd be very hard to say, yeah, based on my experience there, that this would work here. Um, because we just don't know how long was that, uh, reliable for?

Speaker B: And Kevin, that ranges from what seemingly are, ah, very large sophisticated organizations to the much smaller ones, which, yeah, you'd expect that not to be perfect, but listeners, viewers, you'd be very surprised at the state of what we see.

Speaker A: Yeah, that's for sure. All right, key takeaways for this episode. Multi touch attribution is never the silver bullet. Uh, it remains click obsessed. It's built on disappearing data, uh, as we mentioned, and it offers really very little causal insight just by design. Blended measurement is essential. It helps with that. Uh, you know, you can augment attribution with, uh, marketing mix modeling. You can add to that incremental testing to understand what really drives growth and to look at those qualitative data pieces as well, to blend that measurement piece and then finally to really build trust, um, on that data and attribution piece. It's, it really comes down to clarity in communication, you know, aligning teams and expectations throughout definitions, setting expectations around that. It's going to be a range rather than a definitive number. And to use that mix of quantitative and qualitative data to guide decisions, um, to be data influenced rather than data driven.

Speaker B: Kevin I'll just add that so much of this comes back to you can be more confident in the marketing activity you're doing and relying on perfect attribution less. When you execute the first stage of our 5Bs framework, be ready, which is really deeply understanding who your dream customers are, uh, what the buying committee looks like, where they hang out online and offline, and what's important to them.

Speaker A: Because that essentially guides what your leading indicators will be and what you should be looking out for early on in the piece.

Speaker B: If you want to learn more about our 5Bs framework, make sure you head to the B2B playbook.com we have some great overviews of exactly what's in it. If you want to get deeper into it and execute it, we have resources there too. We of course have our programs course that guide people on exactly how to implement it with the templates, the tools, the strategy, all that beautiful stuff. As always, you can find links to everything that we discussed in the show notes and Kevin and I are just so grateful that each week more and more are tuning into the B2B playbook. Um, if you know someone who would get value from the show, please pass it on. Um, it'd be a huge help to, uh, us. We'd really, really appreciate it. Hopefully it would be a great help to the person that you pass it on to as well. Um, thanks as always for your time, listeners and viewers. Thank you Kevin. Take care and we'll catch you next week.

Speaker A: Thanks George. Thanks listeners, reviewers. Take care and catch you all next week.

Speaker B: A quick note before you go, listeners. You can find more great content and get in touch with us at Ah, theb2b playbook.com Be sure to subscribe to

Speaker A: this podcast and our newsletter while you're, uh, there to get the latest news, tips and resources from our Playbook.

Speaker B: We'll be back the same day and same time with another episode next week.

Speaker A: Thanks for tuning in to the B2B playbook. Remember, success will be B2B. Marketing starts with the buyer.