Alex Birkett: Content Growth Modeling

In this INBOUND interview, we chat with Alex Birkett about his INBOUND 2020 session My Cheat Sheet: Content Growth Modeling!

He shares a tip, hurdle, and myth when it comes to growth modeling your content strategy and efforts. As well as we talk about who should be in the room during his INBOUND2020 session and why.

My Cheat Sheet: Content Growth Modeling

All models are wrong, but some are useful," the famous statistician, George E.P. Box, would frequently say. Nowhere is this truer than in content marketing, where SEO tools offer suspect data, Google's algorithm is still never fully predictable, and sometimes results come much faster (or usually, much slower) than expected. Still, modeling your content program and projecting ROI are important activities. This session will cover the nuts and bolts of a simple Google Sheets content growth model that anyone can use to get a better understanding of their content program's expected value.


About the Expert

Alex Birkett

Alex Birkett works on user acquisition growth for HubSpot's freemium products. Sitting at the intersection of SEO, conversion rate optimization, content & partnerships, Alex leverages data to prioritize the highest impact user acquisition projects to bring thousands of net new users to HubSpot's growing suite of products.


Full Transcript

George Thomas 0:00
Hey, inbound attendees, it's your boy George V. Thomas. We're back with another episode of inbound speaks where we talk to inbound speakers about what they're speaking it inbound. Anyway, it's almost over. We've only got a couple more that you've got to listen to that intro to get to the actual good stuff. And speaking of the good stuff, let's roll that bumper and get into the good stuff. All right, I'm here with Alex and Alex. Your session at inbound is a cheat sheet we'll use air quotes there but content growth modeling before we dive into the deep end of the pool and start talking about tips and hurdles and Miss and all those good things. Why don't you take a couple minutes to wax poetic about all that is Alex and let the inbound attendees the viewers listeners know a little bit about you and what makes you tick.

Alex Birkett 0:47
Okay, cool. Well, first off, happy to be here talking about this. It's one of my favorite topics content mixed with the quantitative side. So to that end, I I'm a startup guy at heart I work at HubSpot now, which is the largest company by far that I've worked at Before that I was at a pre seed stage startup Oman starter it since raised a bunch of money, and then C XL, where I learned all about the data driven arts, AB testing, statistics, all of the quantitative stuff that I've sort of brought into my work now at HubSpot, and with the kind of quantitative content growth modeling. Outside of that, I look at myself as kind of a reluctant dilettante. Like I'm interested in way too many things that divert my attention from focusing on one thing. So I'll flit from playing guitar to kayaking to reading books about statistics to just all over the place. But I've got this kind of like generalist mindset which I don't know it's it's, it's, it's cool for connecting dots. And it's cool for just maintaining interest and like all the various things and I'm working on.

George Thomas 1:44
Yeah, I feel like maybe we're brothers from another mother because focus is difficult people when you're trying to like look at all the things that are squirrel, squirrel, but hey, you got it. You got to do it sometimes. So Alex, I want to dive into kind of a general What is the content growth modeling that you're going to be talking about? But But before that, and maybe it will be even be in this section? We'd love to start with, like, who shouldn't be in the room? And of course, we mean, virtually this year. But who should attend your session? And why when it comes to inbound 20 in content growth modeling?

Alex Birkett 2:20
It's a good question for sure anybody who's working on content marketing, that could be a content writer, that could be a Content Marketing Manager, and it could be a director of content. And I think anybody who is managing those people as well, so if you're doing if you're a VP of marketing or VP of growth, if you're doing because a lot of those, those roles are gonna be doing forecasting anyway. But they may not have thought about it on the level of like content specifically, and may have been like a pro program level, kind of like where do we expect in terms of like our revenue growth or our Mr growth or our like, conversion growth, whatever that is, but anybody who's essentially looking at content marketing and working on it, and would like a little bit more kind of like illumination in terms of like what results you can expect from your program and, and in individual level in terms of like, what can I expect from a given piece of content. So content writers are typically not involved in this modeling. But I think you can benefit a lot. If you're a content writer or a content marketing manager from this more. It's almost like a financial kind of a modeling side of things like putting that economics light on it. Love it. So let's dive into this. I'm going

George Thomas 3:19
to give you the first floor, the 12th floor elevator pitch and just ask you the question, What the heck is content growth modeling? Like how would you explain that to the inbound attendees?

Alex Birkett 3:31
So there's on a qualitative level, a contact growth model is something that shows you how your company grows. So there's going to be inputs and outputs now put is going to be sort of the business metric that you care about. And that could be top line growth, it could be something like revenue, and it could be bottom line, it could be something like profit, and then the inputs are going to be things like Alright, what are our acquisition channels? How many people are coming through via those and like, what's our conversion rate? What's our average sales price or like our average order value, depending on the type of business you're in, but essentially the model It shows you like, Alright, how, what levers Do we have the pole, and like how how are those levers contributing to like the metric that we care about. So you can actually like pull that down to the content level and do those same things. And in content, you're just going to have different values for the inputs and outputs, the output is still probably going to be a business metric. Preferably, it'll be something like revenue and conversions, for leads, whatever your business kind of like caters towards, it could be pageviews to that's a little bit upstream from the business value, but then you've got all these inputs on and if you're doing keyword research or something like that, it could be something like the search volume and the keyword that you're going to write the average conversion rate of an article that you put on your blog. It could be the average, you know, traffic if you factor in like social shares or direct traffic, all of these things sort of like act as inputs and can help predict the output, which would be the business value that you're sort of gold on by your company. Love

George Thomas 4:54
that. So and you can take this to the content side, you can take this to the you know, modeling side And of course, we don't want to give away the whole forum. But what's one tip that the inbound attendees could start to leverage or use or build in between now and your actual session?

Alex Birkett 5:11
One tip, I would say, just running a model in the first place is going to give you an interesting thought exercise. So there's this great quote by George Up Box, it's a, all models are wrong, some are accurate, or I mean, all models are wrong, some are useful, right? So none of them are going to be precisely accurate in terms of like what reality actually says they're all based on, like our perceptions and our assumptions, whatnot. But the simple act of doing a model will sort of show you what you should expect in terms of like, Alright, if I write 10 articles, what should I expect for output? Like how much traffic is that going to be? How many conversions is that going to be? And then if you have company goals, like is that going to contribute to the company goals that you're expected to hit? So instead of like realizing like you're disappointed after the end of the quarter, you didn't hit your goals for some reason. This gives you an upfront expectation to say like, well Even if we hit everything that we were supposed to, even if we knocked it out of the park park with every article that we wrote, would this even achieve what we wanted to achieve? And like that assumption that you're gonna knock every article out of the park is also probably an unrealistic assumption. So you want to bake in a little bit of uncertainty. So just doing the model itself, I think it's, it's less about like forecasting specific values that you can tell your CFO or your co CEO, it's more about, like giving you that thought exercise up front to say like, is this actually going to contribute to what we want it to? To do?

George Thomas 6:28
So Alex, before people step into the virtual room for your inbound 2020 session, are there any myths around content and measuring content and success that you're like, Oh, my gosh, if I hear that one more time like that we can kind of hit a reset or reboot before they actually pay attention to the content that you're gonna deliver it inbound 2020

Alex Birkett 6:52
Yeah. All right. So what I'm going to say might actually negate everything I said before, but it's really hard to build a content model because of the data So like all models are based on assumptions, and the accuracy of your assumptions sort of like translates to the accuracy of your output. And with content, what we're dealing with is a lot of like pretty untrustworthy data in terms of our inputs. So like keyword research tools, there, I love them, like href is one of my favorite tools. I use it every day. But to the level that you can trust, like the specific keyword volume of a given given keyword. It's like, I don't know, if it's like, Best Live Chat software, does that have 500 searches a month? Or does a 5000? Does it have 200? It's not really known, like what the level of inaccuracy is on a given keyword, which, of course is going to taint your assumptions, if you're like, building those into the model. So one thing is just like understanding that it's clearly like all of these values are not gonna actually be like 100% accurate, and not letting that stop you from building the model in the first place. Because like I said, it's more of a thought exercise. But I think when people when people dive in, like they'll either do one of two things, it's the one extreme would be to not build a model to display it by ear to the Write content and kind of like, spray and pray like hope that you get a bunch of visitors hope that people come in via like, whatever channels and that's clearly not like the greatest approach. But then on the other side, you have like, more like analytical like, kind of more my side of like the equation where you building these very robust, very sophisticated models, like, I don't know, doing like time series analysis and like CS, parsing out seasonality and like, just getting way too in the weeds. But if you just take a step back, you realize, well, these values aren't actually like, incredibly accurate. So let's not spend a ton of time just like making all these assuming that all of this is gonna be like 100% like scientific. So you want to take a middle ground approach like I wouldn't like treat it like this very, very certain science, but I wouldn't throw it out like you know, just spray and pray and guess you know,

George Thomas 8:45
yeah, I totally agree with that. And I might even say in life, we don't really know where we want to get to until we've gone almost just a little too far. But yeah, let's back up that right there. That right there was the sweet spot. So I love the fact that you painted kind of both sides of This. I'm super curious though other than laziness, which to me is the immediate like response of like, I'm just going to spray and pray because I don't want to take the time to do this. Is there another major hurdle or hurdles that you've seen that have stopped a lot of companies from actually maybe pushing the edge pushing the envelope when it comes to this modeling for their content?

Alex Birkett 9:21
There may be an assumption that like by doing all this modeling and planning things out up front that you sort of prohibited creativity. I don't think that's true. I think you're really just trying to get like a scoping on like the results that you're expecting, you could still be super creative with the concept you actually do produce. I think there's a valid objection that says, alright, like this model, like given the uncertainties with regards to the inputs, like is it actually worth doing? And I think that's actually a valid thing to ask because I'll be honest, like there's been content that I've written that it's expected to get, you know, x pageviews and x leads but then like, it doesn't do anything, even though it ranks perfectly well, then I've had other ones where the keyword search volume and like ah, Stretch showed, like nothing, and we barely put it on the roadmap and the content ended up outperforming all others and ended up being like one of the top performing blog posts in terms of like actual business value. So there's there's room to like, leave yourself open to surprises. But I don't think building a content model. exclude that from being part of your approach. I think it just gives you sort of, you can do sensitivity analysis and say like, well, to hit our goals, do we need 10 more articles? Do we need to reduce the cost of articles by $100 per article, like what do we need to do in order to hit like kind of like the boss's expectations, but I think you can still leave yourself open to serendipity, surprise and creativity within those constraints.

George Thomas 10:40
Yeah, so so much good there. I think, viewers listeners, you should hit the rewind button and one thing I'll say that is coming to my mind if you want to get super nerdy with your content, and like understand, like, there's a lot of terms that Alex is just like flinging out thrown around like a ninja that. I know most of you are probably like What the What like This really is going to be a good session for you and your content. And Alex, I'll tell you, I, too have had that moment where I create a piece of content. And I'm like, this thing can do nothing. And all sudden, it's like, boom, boom. And you're like, wait, what, what just happened? I don't even understand this. So always make the content here, but also use this to make the content. So that's one thing. And by the way, if you're listening to this interview, I literally touched my heart and then touched my brain, but you should be watching interview and you'd be able to see that, Alex, I am super curious. Like, I love this last year when I was doing these, I'd be like, okay, the audience stands up, you're done talking. They give you a raving round of applause. Obviously, that's not happening this year. I speakers, we're all sitting on our couch with the applause meter on our phone. Like, yay, they loved it. But when the sessions over like, what's the one thing that you hope like you're kind of big aha moment you hope they grab hold of and run with as they move forward.

Alex Birkett 11:59
Think about all that. incremental pieces that make the composite of what you want done with a piece of content. So do you expect that it's going to get organic traffic? Do you expect that it's going to be social shares? How do you expect to distribute this piece of content, I would just say, put a little more thought into like the purpose of a given piece of content, the, the, what I want to say the individual component, each individual article should have dedicated purpose and like without a purpose, a model is going to be useless because you're not going to be able to predict what output you expected. So think from a level you could say like, Alright, we're gonna read this article on x topic, but like, what do you expect from that? Is this is this to get backlinks is to like help your sales team? What's the piece for and then you can sort of structure a programmatic analysis of like, Alright, what's the ROI of this program? What do we hope it to be, but at a very small level, like what what's each individual piece of content supposed to do? In our ideal world when we're planning

George Thomas 12:53
inbound attendees, did you hear that its inputs its outputs, its purpose, meaning What's the goal? Why are you creating this content and why you're trying to figure all of this out and signing up for Alex's session inbound 2020 will be here waiting for you in the next episode.