How to Analyse Paid Media Performance Beyond In-Platform Data in B2B & B2C Marketing

Harry Hughes: Hi and welcome to today's episode of Unqualified Leads. And I think we're just going to go straight into this one. This is more of a live podcast in terms of it's something that we're seeing literally currently in our day jobs as we speak. So. It felt like it was the right time to go through and discuss this as it is so fresh. And it's something that we're continuing to see out and to measure over the coming weeks and months. So hopefully off the back of this, we'll be able to revert back this podcast in a couple of months time. But essentially, just to kind of lay it out, is that one of the clients that we've been working with, we wanted to, and we do this with most of our clients, but this was just a more interesting case than the other ones. We wanted to test spend. So we run our paid media budget through a number of channels with this particular client. And we wanted to test whether increasing the amount of spend in one of our channels, and this channel was Meta, whether spending more there would linearly increase revenue. And we this out over sales cycles. So this is just over. essentially a quarter and a half, nearly two quarters. we wanted to see what the impact of increasing this spend was. Now, we didn't increase the spend drastically. We increased it around 25%, But we saw some really interesting results off the back of it. So essentially, we increased the spend. And off the back of it, yes, we increased revenue. So we increased revenue by just under 5%. So there was an increase, but during that period, obviously, you know, we spent 25 % more, but we only had a 5 % increase in revenue. So our media efficiency ratio during that period dropped, it declined. for every dollar that we were spending, the return that we were getting wasn't as efficient as what it was prior to that. So you... We were just above our breakeven return on ad spend, which is fine, but generally, obviously, we want to make sure that we are making our ad spend as efficient as possible but we did see a decline in the media efficiency ratio, which was obviously one of the key points that we wanted to delve into is to ⁓ where was this lack of efficiency coming from. off of the back of that, we knew that we needed to ⁓ dive into essentially as much data as we possibly could. So we spent a couple of weeks just putting together all of the bits of data that we possibly could to then obviously then run through and go through into the analysis of it. We'll kind of come on to what, know, some of those bits that we looked at, but the number one thing that came off the back of it. So this is a booking system type business. So what we have to do as pay media specialists for this particular client is we need to generate bookings. Sometimes we do a bit of lead work, top, really top of funnel, bringing leads before they get to bookings, but 95 % of what we do for this client is to generate bookings. Past bookings, the conversion isn't usually the element of the business that we handle, although sometimes we do help to analyze that, but typically we just focus on the bookings and the rest is handled by client. Now, what was very interesting in this case is that when we actually looked at the trend, over the last two years. to see where the trends were where the seasonality differences were. And what was very, very interesting was that the actual number of orders. So the business is also split into two. I'll just add that caveat into there. So they serve, one half of the business serves a certain type of customer and the other half serves a different type of customer. But the revenue split is around 50-50. And so what we saw in customer section A was that the number of bookings actually over the whole two years, it actually kind of stayed within the same median range. So revenue had increased, which was good because we had increased the AOV for this customer. But with the bookings the same, essentially what it meant, and then once we layered in other elements of data, so we added in our repeat purchase rate, and we added in our lifetime value metrics. then we also added in the percentage of bookings that were generated by paid as a source. And so we layered all of those on top of that initial data that we looked at. So essentially what was happening here was that we were generating customers, we were bringing new customers into the business, but we were losing them out of the backend at a faster rate essentially. So we were seeing a decline in our repeat purchase rate over that period of time. So the number of orders week on week, month on month stayed within our relative range, but that's only because we were generating more new customers coming in. And so then what was very interesting is once you then layer in also the bookings via direct attribution source and you see a decline in time over direct bookings, again that's another indication that we're having a decline of repeat purchases or repeat customers coming back. So essentially all of those data points led to us making an assumption of okay, the acquisition model isn't broken here, actually the retention part which is broken here. And so now what we need to do is we need to go away and we now need to analyze the entire retention part of customer A, which is a whole nother point. won't go into that into this pod, but that led to it being a predominantly a retention problem. So that was kind of part one of, why the media efficiency ratio who dropped during this time. But then interestingly, there was a couple of other elements to this. Another element was that we we had a ⁓ revenue lag. So customer B on the other side, the other 50 % of the business, that had a much longer sales cycle. And so we were still waiting. We needed another two, three, sometimes even four months see out the ⁓ to land for that. So we had signaled that we had grown that part of the business by 34%, but we still hadn't seen the revenue land from that. And so we just needed to wait another two, three, four months to see that revenue to come in. And so again, we've predicted the revenue that we're expected to see there. And obviously based on what lands versus the prediction, obviously that will then tell us more information specifically around the conversion and whether that's something we need to look into on that side. So that was another interesting part, which again, told us why the media efficiency ratio had decreased and that actually if we were to go, ⁓ we were to jump another six months ahead and review the same period again, the media efficiency ratio would actually be more effective ⁓ it's taken into account the delayed revenue. So that was point two. And then point three was to then look at the actual channel itself and whether that was directly generating more bookings and more revenue for the business. Now, interestingly, this is always the case when you look in platform, it always looks great for most cases. It always looks great. It certainly looks better than what it is that you're seeing in your CRM. So we were looking at Meta and we're going, okay, this is brilliant. We're getting more bookings. Our CPA is dropping. Perfect. This is all making sense. But actually, when you then take that back into the CRM, and this is again, we're talking about customer A here. When we take that back into the CRM, we take it back to the revenue data. And once you segment it out and we layered on top our number of bookings, our click attribution CPA, our incremental CPA, which you can also get from Meta now, although that was only released in April 2025. So you can't go back all of the way. And so we basically layered on all of those. And then we also layered on top our normal attribution settings, which typically use seven day click and one day view in our prospecting and seven day click only in our retargeting. And that's really as tight as you can get unless you do move it into incremental attribution settings. But interestingly, once you layer those on top of each other, that again gives you another data set understand and to analyze. And actually what we saw there, was that there's still, when we looked at the incremental number of bookings, there is still a gap of maybe 10%. It's kind of changed over time, but anywhere between five to 10 % of bookings that we were seeing inside Meta that actually weren't incremental. Or essentially that's what Meta is telling us. so then that kind of goes potentially in line with what we were seeing with customer A, where we had kind of capped this number of bookings. And so ⁓ us, the next thing to do, is to go okay. So we know that spending more is increasing what we're doing over in customer B, but we're just waiting for revenue lag. But actually, customer A, we've got this retention problem, which we need to go away and we need to understand and we need to analyze. But also potentially, we don't actually need to spend that much more budget here. And actually, maybe spending more budget isn't actually incrementally driving more bookings for customer A. Or at least that's what potentially the setting inside Meta is telling us. And so the outcome therefore ⁓ is number one is to ⁓ wait for the revenue lag to land in three to months The other point is then delve into retention so then, that's a whole nother analysis that needs to be done. So that's kind of point B. And then the third one. is to then run another incrementality test where we take the budget out specifically within customer A. So we'll drop the spend of customer A inside Meta by 10 to 15 % and we'll run that over two sales cycles and then we'll analyze what the impact of that is. Well, ideally not two sales cycle, we'll probably do it for one and then we'll have a look and we'll see whether the actual volume on the backend does drop by the percentage that we've dropped it by. And so that's going to be the real telltale because we just generally increase budget across all, but now we're going, okay, increasing budget across all great. We grew this side, but actually this side didn't grow so much. So potentially one of those issues is that meta is no longer incremental there. So we're going to drop that by 10 to 15 % and we'll analyze and see what the outcomes are. And I think this just really comes back to, know, paid media certainly can be very, very simple. But actually understanding paid media results isn't always so simple and it takes considerably more data and time to understand and analyze actually what's going on because what you're seeing from meta from Google might look superb, but actually you really need to go into the backend and you really have to dig in and see exactly what's going on. And I think it's so worth spending the time and running this analysis, ⁓ you know, every couple of quarters, every half year, every financial year, certainly, to understand and look back and dig out because there will be holes. And that's why I also love media efficiency ratio as a metric and watching that change over time and saying, okay, this is interesting because we're actually, know, every dollar is less efficient than it was previously. And so it's tying in all of these different points using GA4, using your CRM, the in-platform metrics that you have, you know, don't just... ignore those. They are still important to look at. You still need to look at those, but look at the different attribution settings within them. I know that in can look at data driven in Meta ⁓ and you should be looking at incremental settings and the ones that you're using. Layer on top of each other. Have a look at that over time and see what the differences are. because it will tell you information that is very, valuable. that in on what you're seeing from GA4. Look the number of leads, contacts, whatever it is that your conversion action is. Look at the source of those and look at that over time. Has paid search change, has paid social change, has direct change, is that increasing? And layer all of that in on top of each other. And then ideally, if you're running UTM attribution like we've mentioned many, many times before in this podcast, you can then actually layer that in from your CRM. and have that on top as well. And obviously all your financial data to look at the very, very top level. And so it's just having all of those bits of information, which I do understand for some marketers, you won't always get access to that. You're not always going to have access to the revenue data, but you can ask for it and explain why, you know, sometimes maybe even you just have to produce your data and then hand it over to the revenue team to say, hey, can you take a look at this? can you layer that in on top of your financial data? but you need to try and get as much data as you possibly can. And also over a long period of time, just looking at this over half a year, even just a year really isn't enough. You need to be looking at it year on year with a couple of years worth of data to see where the trends are. But it very interesting exercise that we went through. it back to the executives and again, they were shocked as well. There certain things that came out of the data that we produced that they hadn't quite expected and they certainly had assumptions coming into the meeting which again, ⁓ the data we produced of told a different story. And it's just, it kind of really hammered home the importance of spending the time and putting together ⁓ all of those bits of And I think as a marketer, what's really, really important as well is Yes, media budget is your responsibility. So obviously, naturally, you want to defend yourself and you want to defend what you're spending as much as you possibly can. But often what I find when it comes to this analysis is to just be completely open minded and need to be willing to take the answer ⁓ no what it is, whether it shows that what you're spending is actually bad, in which case then... that should spark something inside of you to go, okay, well, why is we're spending more, but hasn't changed, why is that? And that you naturally should then want to go in and understand. So I think it's just going in with an open mind and going, okay, let me get all of the data, let me understand exactly what's going on. And it becomes actually a very exciting opportunity, I think, because you ⁓ ultimately out with it with several action points where you can go, okay, great. There's all of these different points that is... is potentially changing this and we actually could see a huge uplift in revenue if we just do this or if we test this or if we change that and all of a sudden it becomes very exciting again. it's just been a real great couple of weeks for us going through the exercise again. Like I said, we spent more, revenue grew, but the media efficiency ratio didn't quite align with the amount that we were spending and it should have retained its efficiency. Yes, arguably sometimes when you do spend huge amounts more, you will naturally see your efficiency drop because you need to spend more to reach more people. therefore, as you broaden, you naturally reach more people that aren't typically within your tight ICP. But in this instance, I would have expected the media efficiency ratio to have stayed essentially the same, ⁓ not decline. So very, very interesting and really to see what's going to happen when we drop the spend. back down by 10 % within customer A and also when we jump into the analysis and obviously waiting to see ⁓ revenue lag as well. So it'd be very, very interesting to ⁓ back on this in a few months time, but super, super interesting. And the great thing was that everybody came away from it so positive as well because it showed such great signs for how to grow the business how to be more efficient and more effective. Dan: Yeah, I think I can give them some examples of some similar scenarios or at least things to out for. So, first all, ⁓ a Google or a Facebook rep. you, any business or any agency that are working with them, they're typically going to want you to spend more money, it's in their best interest. They're going to be the last people on earth that are going to say to you, spend less. So the first thing is do not be steered necessarily by the reps that you speak to. You as a media buyer, as a marketer, CMO, whatever, you need to understand that. Harry Hughes: Mm. Dan: you need to scrutinise that data and actually look beyond in platform. Again, it's going to be within their best interest to say them the channel rep to make you look at in platform. the agency ⁓ buyer. You need to be able to say ⁓ need to look at several data points to confirm and get like a confirmation bias on whatever. So the first thing is to make sure we scrutinise that. I can give an example of a business that's series ⁓ D ⁓ several ⁓ budget. And again, ⁓ Google have kind come forward them knowing they've got like big raises, big amounts and ⁓ pushed on the narrative of you need to run full funnel ⁓ across multiple markets. And that involves, you know, fully broken out search campaigns. Google Display Network, PMAX, you name it, it's running across several, several markets. You run all of that. There's no measurement of, it kind of mismanaged, I think, after looking it over, say, the last 12 months. So this is like an audit of an account. ⁓ And the thing look at is what was their kind of... resting traffic level and then you know from say organic and general traction PR whatever they were doing and then what was it looking like from the amount of spend that was going on top of it so let's call it 150,000 a month 200,000 a month going through primarily Google and LinkedIn and first of all you could see when you just cross-reference display data lots of conversions in platform Were those conversions recognised the end? Some of them, yes, but a lot of it was fraudulent conversions. Out of say for example GDN, Google Display Network clicks, you might have 60,000 clicks showed in platform. If you tied that back to paid users that were actually landing on site, it was around say 5 to 7,000, something like that, which obviously doesn't correlate. So there was a lot of fraud clicks, a lot of bots that were going through that. So something that any business should be watching out for. Network as well, running for conversions for a business like this was pretty wild. you wouldn't really expect it to work. It would be more of a display activation if anything. And if you were looking for true anchor mentality, I certainly wouldn't have been pushing lots of budget through that. I would have stuck to high intent search as this is a high value kind of almost B2B sale cycle. So. wouldn't have gone out of probably exact match high ⁓ search terms for a business like this within Google. You would have started there and then scaled up and out if you needed to, but realistically likes of LinkedIn ⁓ have been better for you for top of funnel. First of all, Google are going to tell you to spend a lot. you're an experienced media buyer or marketing professional, you're going to understand that. the first thing would be, no, we're not going to put 150k straight into Google in every market and run it. You're just not going to do that. You're going to test a smaller budget and then scale that budget up accordingly with the results that you see. So say it's enough budget a month and you know what your average cost to acquire a customer is. And you obviously would just do all the unit economics, understand the numbers and work it for ⁓ there and scale accordingly. other thing is obviously coming on to that, coming on and taking ⁓ into this account, you try to marry up the spend to traffic volume, website volume. ⁓ and you couldn't. So when you saw additional spend and then you ran it through total users, none of it correlated. So that shows you how crazy the level of budget, like the inefficiency of budget was. This is before you're not even looking at revenue here. This is just looking at budget to traffic ratios. So each month if you tried to over a 12 month period, if you looked at 150k in this month, 75 in the next, you barely saw an ounce of movement within traffic. So that should have been alarm bells for anybody looking at the data if it was looked at properly with that level of spend if you were looking at j4 and you saw that you should be thinking what's going on here if you're spending 150k on top of a resting user of 50,000 users a month you would expect to see ⁓ the per user to be ⁓ and the volume to increase by certain percentage and then from there on out you know what your average cost per user is to visit your site and then eventually you obviously get down to the numbers of revenue realized off of ⁓ those and CAC and everything. ⁓ But so if you scrutinize the data from that you're not looking at incrementality here for revenue this was just even traffic it was shot to bits. And then when you looked at LinkedIn Obviously the cost of a CPM on LinkedIn is very high, the cost per click is super high and therefore a cost per user to the website is also high. So trying to measure 30,000 a month against 150,000 a month of Google, you couldn't even see the impact of LinkedIn because it was just lost in direct and organic. So you could never really see the increase in traffic off the back of that. To add more complexity to this, there was no UTM attribution through HubSpot and through CRM. when you try and, you can see all this front end data, but then you try and look at it and realize it within the CRM you can't. you just got. tons of sort of sign-ups, registrations, whatever, no real understanding of where they're coming from. So what should have should have been is obviously hidden fields across all forms and ⁓ UTMs being ⁓ into the CRO which I know going to touch on ⁓ before we close the end of this this pod out. ⁓ should have been in place that wasn't in place so Harry Hughes: Mm. Dan: What you have to do with all of this in mind is you've then got, and rightly so, say a CMO asking, there's 100k, 150k of media spend going through all these channels, what's actually working for us because we don't know. How do you know? Well, first of all, you can pull, you can switch display off confidently as a media buyer. You should probably have a relevant... level of gut instinct, intuition, media buying intuition, I would have thought, which is that you know displays probably not really having a great deal of impact. Certainly now with the likes of AI, how people are discovering brands, realistically are you looking at, you know, when you're on a website as a prospective buyer, if you see a little banner on the right hand side, wherever that banner might be on whatever website you're on, are you actually clicking that convert in it? Like I'm... Sorry, but I don't see how it's that efficient. If you were to see something within social and it engaged you in the right way, I buy into that a lot more. That's just my personal preference on it. Anyway, so switch display off. Do you see any impact? No. ⁓ Have you just got rid of how many high level of fraudulent clicks that should have already been excluded? Yes. Okay. That's one problem eradicated. Now what do you do? You whittle it down to high intent keyword campaigns and you basically start from the bottom up as to where you would have started from the beginning. And is that impacting overall registrations and money in overall revenue outcomes or your media efficiency ratios? Is that impacted? No, it's not. So means that there was a lot of work going on, be it PR and other work, which is big. for this particular industry that was having a large amount of impact as a base level. So when these, when the likes of Google was laid on top, all it answered for them was they saw lots of conversions and lots of vanity metrics inside a platform. They didn't see anything of value off the back, but that's because none of it was being measured correctly from the very beginning. So you basically have to strip it right back. Then it comes to LinkedIn, proving out impact of LinkedIn, ⁓ can already see ⁓ right businesses are engaging on LinkedIn. Harry Hughes: Mm. Dan: because there's the ability now within LinkedIn to see what companies are engaging via paid and organic. You can pull that and extract that out in a spreadsheet. so anybody listen to this, that's a great way of being able to show early indications and early signs of value of the work that you're doing. So ⁓ were running a campaign and you know exactly what kind of ⁓ descriptions people you should be targeting, you're running all of that and you know, what business is they're likely going to be at, you can now export that information and show that. So it's early indication you're in front of the right people for executives to buy into what you're doing. The next bit is how can you prove that they're actually going to register, sign up to the thing that you're doing? already running, say, ⁓ leader campaigns, you're already seeing a lot of shares, high engagement, good dwell times, good engagement on page. that they go into all the website. Session duration is decent. Comments are good. The right people are commenting. ⁓ of these like early signals are great, but really people want to know is, you know, are these people then going to come and register? Are they going to sign up? How can you prove that out? So for an example ⁓ would Geo Holdout in a new market. So you wanted to do New York ⁓ an example or state, you could run a specific webinar. in that area and there would be two things you'd be looking at is the front end which would be you're running it in New York, are you getting executives signing up to the webinar to listen to what you want to talk about from this thought leader, the right people turning up, what's the quality of these people turning up to the webinar. That's first part, obviously second part being are they staying on the webinar, are they engaged. Basically you break down all of those webinar metrics and make sure that the function of the webinar is doing what it should be doing. Obviously you're not trying to sell them immediately and say that webinars as a sales pitch, should be an education pitch. So educating them on ⁓ on market trends, what's going on for this specific business. Off the back of that in New York, do you see registrations increase and people going onto the platform and doing the action they should be doing off the back of it? Let's say it was an investment or something like that. Over the six months, yes or no? That's then an incremental measurement of LinkedIn's impact for your business overall as a tool. And you're measuring it by targeting a new state that's not been targeted before and looking at whether then, say, the final goal in which you have is, an outcome for that within the next six months. So that's the plan of action with LinkedIn to prove its impact. Because been running in the same areas over and over and over and over. So ⁓ trying... it out, you can, but a lot of it gets attribution gets lost through direct ⁓ organic. So how you do it? The real there is to use a geo is a great opportunity to do it. So if you are doing it states, could be market. States is probably easier if you're in the US, just trying a new state ⁓ and seeing the uplift looks like over time and being realistic with it as well because these people are new to the business, they're cold. So have to be very realistic with what that looks like. So that's one great way of doing that. And then, you know, with Google, you're going to see a lot cleaner attribution because it's bottom of funnel. So if you're doing high intent keywords, as long as you have those UTM parameters, everything set up correctly, you should be able to see more clean attribution to bottom of funnel, signups, conversions, whatever it is you're looking at within the CRM. So stripping that back and then scaling up accordingly you strip these things back. But like I say, the first thing to do is like look at everything, scrutinize the data, especially when you're at volume ⁓ understand what's actually truthfully working for the business and what isn't. like, yeah, media buyer that sees that kind of level of spend going through Google when it's not really like the lines are blurred as to what's working and what isn't. And basically this is, ⁓ I kind of similar to what B2B would have been like a few years ago when venture capital was getting thrown at it left, right and center. and people were just spending and media buyers were just spending budget because there was marketing budget but it was never ⁓ a bootstrap business would spend which would be testing like one channel one ICP scaling up gradually and then going okay now we'll test another one and like what you're talking about where it's like heavily scrutinized and increasing budgets to see if we can scale okay increase it by 25 percent are we seeing scale in six months or is our TAM you know have we maxed out TAM Harry Hughes: Hmm. Dan: in that geo or something and then you go into another one and you look at whether you can then scale up in that geo. It's like things like that that you would typically do whereas with a lot of these a lot of the VC money that was going into say B2B SaaS there are still industries where VC money is still ⁓ don't think it's as aggressive as it once was but it still gets thrown around and carelessly spent. because they're like, have X amount percent of rays that we're going to throw into marketing. Google rubbing their hands together, Meta rubbing their hands together. From my experience, LinkedIn reps, the ones I've worked with have been... Harry Hughes: Hmm. Dan: being very good. I've not really had them telling me to max out spend or saying, hi Dan, your Google conversions could be x amount percent more this month. Why don't you spend more? I don't get that from them. Google on the other hand, I've seen across a few different businesses where they've got them a series A and beyond, and they know they've got two, three million budget. They're just trying to throw tons and tons and tons of budget into the platform. I get it. They're a business and that's what they're trying to do. However, any... Harry Hughes: Mm. Dan: Good marketer would never just let that happen. You would scale up accordingly with your marketing budget. It's hard for me to even comprehend, but I've seen it happen and it blows my mind that you have like... Yeah, obviously my two to three million pound budgets get solely pushed into Google because Google have pitched a Google accelerator for this business and you come on as the media buyer and you're having to battle the channel that trying to take all of this budget and spend it through when you know you that's not the best thing for them. So you have to be obviously trusted as the agency to be able to put your best foot forward, advise them against that and say, let me... Harry Hughes: Hmm. Dan: You know, if we're going to put budget into these channels, this is how we will do it. We're not just going to go, cool, they think we should start this amount of budget because the cost per conversion is 300 quid. So therefore we have to spend X and we have to decide, no, no, no, no, no. What if you didn't have all of that money? What if you didn't just have a hundred grand a month just to throw a straight and talk? What would you do? Be smart and think about how you would do it. Would you even then start with Google? Maybe not. Maybe you would start with a different platform. because generating demand or you know demand activation within that channel could be more cost beneficial to you as a business than trying to capture existing demand and playing against incumbents within ⁓ ⁓ you can be very very smart even with the budgets being huge or you've carried out big raise just be smarter with ⁓ spend. It doesn't mean just pump it into channels and accelerate it through conversion front end conversions. Like that shouldn't be something that we have to explain, but it's insane. And it's just crazy. I never do that. It does happen. And you'd be shocked at how many businesses you go into and you go, what's okay. What's going on here? And the lot of what we're hitting our conversion volumes. You go, okay, well great. Harry Hughes: Hmm. Mm. Dan: what's that like on the back end then revenue wise? Do you know that all those Google conversions you're seeing is that Google revenue? Is the all the conversions you're seeing on the, you you're seeing 5 % in LinkedIn, but is that actually attributing to 8 % of your revenue? You can't answer that. So that's crazy. That does happen. It continues to happen less within startups because you don't have the luxury of your bootstrap startup. Harry Hughes: Hmm. Dan: You're not going to see it as much because you're going to be a lot more careful with your spending what it looks like that you've run into different problems there. If you're a startup and you've got seed investment or a series A or your scale up maybe series A series B, you then have the pressure of we've got lots of budget, we need to spend it and we need high growth. Again, it makes sense because you're looking for growth, but there's a way to spend that budget. It does not mean, you know, just shall I channel spend into conversions? You could test channels you know your customers exist, test angles, test approaches of paid media and scaling paid media up and finding out which angles within those channels work for your buyer and then scale that and then scale out another angle and then another. You don't just blindly go in and spend money into sort of these rigid cutter structures. doesn't work. I could go on a tangent about this for a long time but there's nuance to all of it. Harry Hughes: Mm. Yeah, I think there's a couple of points that just hearing you talk about that just kind of made me think about them. one is I think what makes a good marketer and certainly a good marketing agency is the ability to think outside of just marketing. So this is why I always believe it's so, so important to ⁓ start. whenever it comes to analysis or understanding the effectiveness of marketing is to always work from top down, always. It's the most logical and it also helps you to understand where you need to go in the next steps of your analysis. But just starting with, okay, last financial year, we increased spend in marketing by X the increase in or the change in revenue is Y. Okay, is that positive, is that negative? And then that gives you an understanding and from that you can then start to build out your media efficiency ratio and you can understand, okay, great, how has that changed over time? Is it more efficient? Is it less efficient? And so on. And in some instances, that's gonna be positive. In most instances anyway, you'd hope that it is. Great, we've spent more, we've got more revenue and we're still just as efficient as what we were, sometimes maybe even more efficient. Well, that's great. In case, we know that our marketing's working, but as a marketer yourself, you're probably still gonna wanna go in and... start to do the analysis on the channels to find out, okay, great, how can we make that even better? What channels are working better than others and so on. But obviously if you were to look at that and go, okay, we've spent more, revenue hasn't quite increased where we need to be, we need to diagnose what is working and what's not. And I think the reason why you need to start from top down is because if you just go straight in at the platform level and go right, I know that we're spending money on LinkedIn. I know we're spending money on Google. I want to know what's working. But if you miss that top layer, then this is where you can potentially analyze wrong. Because if you just go straight into the channel level and you go, okay, great. I'm looking at all the data I've got available to me and we've increased our spend on LinkedIn by 40%. But actually the number of conversions that are attributed to LinkedIn certainly hasn't increased by 40%. And actually our CPA might have increased, but actually your top line revenue has increased by you know, 40 % let's say, but you're just not seeing it in LinkedIn. So, but then you see that there's been an increase in Google, a large amount of your conversions are coming from Google. So go, okay, that's where it's working. It's working over in Google. But actually you've spent 40 % more in LinkedIn, but you spent 40 % more on thought leader campaigns and more top of funnel demand creation type campaigns. Well, you're not going to expect to see the conversions come straight from that. That's a, there's going to be a delayed impact. from what you're doing over in LinkedIn for your demand creation that likely Google is just going to hoover up however long your sales cycle is, but that's just going to hoover up down the line. And also you're probably going to see more that comes from organic search or from direct. So your analysis isn't going to show you that LinkedIn's been more impactful, but actually if you look at the top line, well, something is working and you're just not tying it back to LinkedIn because the data suggests otherwise. And so I think that's why it's so, so important to go from the top and work your way down and segment it then by channel. But always, always start from the top. Don't just go straight into the channel because I really strongly believe that you can create the wrong answers if you just go in, yeah, if you go in at just at the channel level. And that's why when I was talking through the analysis that we were running at the start, there's so many data points that are layered into it. Dan: Yeah, misdiagnosed. Harry Hughes: And I think as a marketer, need to be curious and you should want to layer in as many data points as possible. So start at the revenue level, marketing budget versus what comes out, paid media spend versus the revenue come out. And then start going, okay, now what do the channels tell me? Okay. And then layer in the different attributions set into the channel. Have a look at those. Then have a look at GA4. Then also layer in what you're doing over in your CRM with your UTM. lay in all of those different points and each one will tell you a slightly different story. But it's your job then as a marketer to look at all of that combined and then to understand and go, okay, I've taken into account all of the data that I possibly can here. And this is the story that I'm pretty sure it's giving me. And then off the back of that, you'll be able to test and you'll understand, okay, well, I'm pretty sure that it's coming from here. So I'm just going to test that and change that. And I think that's just the real key is just top. down and just go through it in a logical way, but don't skip the data. You need to gather as much as you possibly can. And I think it's exciting when you do it. Dan: going to jump in and say something there as well. just, my mind went off on a bit of a tangent when you were talking about looking at impacts. Another thing with these sort of bigger businesses is they're going to be running like PR and different things in the background. So they might do some like more. Yeah, exactly. So layer that in. So if you know that you run a press release over a period of time and it might go, if it was business, could like Ted crunch or whatever, and different places talking about you. Harry Hughes: Yeah, layer that in. Dan: Financial Times, whatever. If you run that PR piece over a window, look at the impact on traffic through that window on top of your existing traffic. Again, it sounds so simple. I know, ⁓ we did this with ⁓ startup business. I remember when we did our first PR, we knew what our base level was from organic traffic from the SEO we did. And then what we did was we ran the PR through the magazines to our industry we then look at what the uplift was in traffic over that period of time. you would see the spikes. So we knew every time that we would run articles in those magazines, we would then see a spike over the next two to three weeks of X amount of traffic. And then also off of that, was there a higher quality of dwell time from that traffic? And then did we see more users in off the back? And we did. And could see it clear as day, these peaks that we would run from doing this PR and it worked. And then we laid in paid social, you had like the organic amount, then that jumped up. Direct and organic went up, paid social never went up as much as in GA4, which is just the nature of GA4. Then that sat at a new level. So because we had the luxury of being able to measure layer by layer and never taught to do that. We weren't ever taught, do this, do that, but it made common sense. Like, okay, we have this base layer. This is how many users we have the last six months. Now we're going to introduce this. Let's see what the uplift is. Now we're going to do this. Let's see what the impact is. Measured it as we went, always had the percentage differences. So we always knew roughly if we spent this, this is what Direct and Organic did. This is what cost per user did. Like all of that data was managed. And I could say we were never taught how to do that. And I just think like even that is such a simple thing to look for in data. Harry Hughes: Yeah, agreed. Dan: and see the impact of these things that you do. It's like, well, if you know you're doing it, measure the impact and look at what was data like three months before and then look at what's data like three months after. If you run that PR campaign, make sure it's measured in some sort of way. It's not always going to be very easy, but you should be able to see something from it if it's digital. Harry Hughes: Yeah, ⁓ that can also act as a missing piece of the puzzle. know, if you were looking at, for example, user ⁓ or GA4 data, for example, year on year, and only layering in what you're doing from a paid media perspective, but you're seeing big changes in, you know, either drops or increases in traffic or conversions, and you're trying to kind of figure out what's going on. Well, there's probably a whole ⁓ missing that you haven't added in, which is... actually last year we did 15 trade shows that we haven't done this year. So that could be an impact or, actually we spent $200,000 on PR, as you said. It could be a whole number of things. Oh, actually, funnily enough in Q1 last year, we did a huge out of home campaign. Oh, okay. Well, that explains the huge increase in conversions that we had in quarter two. So you have to layer in that aspect as well to really get the full picture. And then I actually think... coming back to this being naturally curious as a marketer and going in with an open mind is it doesn't always have to relate to your channels. So what I mean by that is let's say your media efficiency ratio has decreased. might not necessarily always be due to you being less efficient with your channels. It might be, but you have to be open to look at those other things. So also look, as we just said, what other what else were you doing in marketing? What else were you doing as a business versus the other period that you're comparing to that's different to now? But also you need to look at other potential possibilities. This is why it's really important to layer in the kind of the bigger metrics from the wider part of the business, which is conversion rates. Has our conversion rate naturally changed over time? Well, if we've seen a five to 10 % decrease in conversion rate, naturally our media efficiency ratio is going to decrease. In which case, then that's a whole nother can of worms, which you then need to go in and analyze, which is okay. Why has our conversion rate changed? Why has it decreased? Is it, and that's going to tell you a whole nother possible, you know, number of reasons as to, as to what you need to do. It might be that the lead quality is now, you know, 10 % worse than what it was. Okay. Why is that? And you're just basically constantly chasing these different angles to try and understand exactly what's going on. But that's why you need to look at those wider things, not just in channel. You have to go, you have to cast your kind of analysis net as wide as you possibly can. You know, it might be something as there's another competitor in your space, which is undercutting you. So that's why your conversion rate is dropping. There's another competitor that's just entered the space. Well, okay, ⁓ people are going over there. Retention's dropped. There's so many other wider business factors that can play into why your media efficiency might also decrease. so just be, I think it's just the importance of being naturally curious, but also not just being so restrained to looking at your in-channel platform and just going, okay, top down and then what cast your net as wide as you need to go to understand exactly what ⁓ other causes might be at play. because I guarantee you it's more than just what you're doing from an in-channel perspective. And you need to have a look at the whole ecosystem to really build up a really good story and a really good understanding of exactly what's going on. And I think that's what makes marketing, certainly from my perspective, really exciting, is to then tie that all together and build out that whole story. Dan: I and I think yeah the problem solving aspect of it is brilliant and it's like an thing to sort of into. The HubSpot ⁓ UTMs I that this actually end of this tease up nicely for the next pod because I think if you use the frame of reference that I started with with this amount of money going through Google no UTMs being stored in HubSpot, no real architecture set up. It's just capture ⁓ and never scrutinize it. Just capture contacts and sales deals with it. I think, again, a lot people will be shocked at how many ⁓ businesses literally their CRM like that, even at... Harry Hughes: Mm. Dan: Series D, they've got that far and none of it's scrutinized. is sales, just gets a CRM. Marketing isn't even really engaging with what it looks like. mean, if you take extremes like what you just said and you go, well, money into money out all looks great, but then you scrutinize it and go, well, what's actually working? And you can't because... There's no CRM data that's not even being stored. So I think that would tee up the next call nicely would be to talk about, could use it as a working example. So this business, they're spending 200K for Google, that this is for LinkedIn, whatever. ⁓ no real ⁓ data, say being stored. So we use that as an example. Harry Hughes: The date is not there. Dan: what you would do set up wise, hidden fields in forms that should be there obviously, ⁓ just using, like you said, the HubSpot. UTMs that are native to it, but like sorting out the hidden fields and then structuring what first touch, last touch, all of that looks like through the journey per contact and how you can measure that journey, that customer journey as they come through, how the are stored and people should be taking advantage of it because they might not be there forever either. So whilst they do exist ⁓ this world of data that's being crushed over time, you be taking full advantage of the fact you can still store them ⁓ and capture in a CRM. So I think, yeah, Harry Hughes: Mm-hmm. Dan: that would tee up nicely for the next pod where we should probably run through that and use that as a working example and showcase. I know you've already set that up for a few different people but you could just talk through. what you've seen before and after. And then also you could layer in self-reported attribution because I think as well, a lot of people will look at that in its individual, as an individual, yeah, in isolation, sorry. And they're not using it with everything else and it can't, you can't just use it on its own. It needs to be used of everything as a data point again. So I think using that as well would be a good thing to touch on. So why we leave this one there and then in the next pod, Harry Hughes: Yeah, in isolation. Dan: we can talk about how you can solve that and showcase impact from channels through the CRM as well. Harry Hughes: Yeah, absolutely. For the majority of clients we work with, obviously we work with clients on different scopes, but for the majority that we do work with, obviously we also have access to the CRM. Some companies don't always ⁓ you to have access to the CRM. So we'll sometimes give them guidance on how to set ⁓ UTM as we see best practice. in some instances of clients, we get to go in and we get to actually build it out ourselves. And in some of the instances where we built it out, ideally, you know, we'll touch on it ⁓ in next week, but we ⁓ ultimately up with the dashboard in HubSpot or within Salesforce or whichever CRM you're using, which breaks down, you know, first, last, conversion based and self-reported attribution. So you have a dashboard which generates, you know, all of that. And that ⁓ adds as a layer of data which you can then use alongside GA4 alongside your in-platform as well and it's ⁓ super powerful. It's something I recommend whenever I start working with the client it's one of the first things that I recommend that we build out you know the quicker that we get it implemented the quicker that we get the data in the longer we leave it the harder it is to go back and reconcile. You can reconcile UTM data to a point but it's never going to be as accurate as what you ideally want it to be so It's one of those things that we try and set up as quickly as we can. But yeah, it's incredibly effective. And again, it's just, if you can get more data and you can have more data that you can trust and rely on, then you absolutely should be doing it. And that's why we find UTM attribution so, so important. yeah, look forward to touching base on that one next week. great to catch up as always Dan and I look forward to seeing you on the next one. Cheers. Dan: Nice one, cheers bro.