The ‘AMP Up Your Digital Marketing’ Podcast Presents:

Dr. Ian Howells on AI in Marketing at Scale

One of the greatest challenges marketers face is creating campaigns that can both scale and are unique to the person viewing the content. Its where the premise of ABM originated; it's how marketers get the most bang-for-their-buck. It is undoubtedly accurate that a well-tailored campaign that speaks directly to the reader at the exact time they need to hear the message will convert better than a generalized message using the spray and pray method. 

One of the greatest challenges marketers face is creating campaigns that can both scale and are unique to the person viewing the content. Its where the premise of ABM originated; it's how marketers get the most bang-for-their-buck. It is undoubtedly accurate that a well-tailored campaign that speaks directly to the reader at the exact time they need to hear the message will convert better than a generalized message using the spray and pray method. 

Enter AI into this mix and it combines the best of both worlds - AI in marketing at scale while allowing for micro-targeting of verticals, groups of people, or those with very specific attributes. It also allows marketers to ask better questions, and questions that haven’t been considered yet, to discover new attributes that could be relational, which helps to scale marketing campaigns.

In this episode of AMP UP Your Digital Marketing we meet Dr. Ian Howells, Vice President and Head of Marketing at Sage Intacct, an accounting software. He speaks to tailoring their marketing approach to allow for scale while creating an unparalleled level of micro-targeting to their target markets. In this discussion, he challenges the way marketers think about verticalization and grouping together industries just because they share a common thread, and how to leverage AI to influence them better. 


Glenn: Welcome back to the show. Today we're speaking with Ian Howells. Ian, welcome to the show. 

Ian: It's great to be here.

Glenn: Ian, can you tell us a little bit about who you are and what you do?

Ian: Yeah, I'm the VP head of marketing at Sage Intacct. We're a cloud financials company that was acquired by Sage for just under a billion dollars two-and-a-half years ago.

Glenn: Oh, very nice. And that must have been interesting, just going through the acquisition. In terms of some of the areas that you've been focusing on lately, what's the area that really pops up for you, what's got you most focused or passionate about right now?

Ian: Just it's interesting really the space that we're in has driven the approach that we've taken. So, we sell cloud financials, we have cloud budgeting and planning, cloud HR and people and analytics. And Gartner kind of characterized the market really well in 2018, it called it the largest enterprise software category at about $37 billion going through a generational shift to the cloud. 

So, what you have is this ginormous cloud market really shifting from on-premise to cloud. And each vertical is shifting at a different pace in each geography. So, kind of the key thing that drives what we do is normally you're looking for opportunity, but here the opportunity is so large, you're looking at where do you start? And there are three core things that we do. We kind of break the world into what we call the micro-verticals, and we use AI at scales to kind of personalize the scale.

Glenn: So, okay, so AI means different things to different people. And so, just give us some context here. How would you define AI in this? And then, I mean, is it just machine learning, is it – how far do you go down that path and how do you do it?

Ian: Yeah, absolutely. So, if you look at the key problem we're trying to solve, to give you the context to the AI approach, is what we see is this, this potential to kind of hyper-personalize the experience at scale to give a really great a level of engagement. We think that ABM, if you're over-ABM-ized you kind of go from an opportunity to thousand prospects, fine. 

So, you got to – how do you deliver really relevant content using AI at scales. So, for us it's really how do you get in the mind of the buyer rather than the seller? And how do you understand the intentional behavior of the buyer? So, for us, we're looking for, you have a fit, which is not just similar customers but really, really similar customers driven by the features of AI. Intent, which is, what kind of intent does the buyer have? Recency, which is, have they been looking in the space – two years ago or two days ago? And engagement, how engaged are they with your company? 

Really, AI looks at, you know, thousands and thousands of features to kind of drive that kind of that – what we call the fine model. And then the other kind of key thing is, when you put yourselves in the mind of the buyer, very often people think about their own funnel, you know, are they top of the funnel, are they middle of the funnel, are they bottom of the funnel? Whereas the customer, we think about them, are they a target, are they aware of the space, are they considering a purchase in the space, are they in a decision or a purchase phase – which is very different? So, you put your mind in the funnel of the buyer rather than the funnel of marketing. 

And that's just when we're powerful because, you know, a company may be looking, researching the space for a year and suddenly comes to you, they're bottom of funnel even though they're just entering your funnel. So, you know, it's really putting yourself in the funnel with the buyer rather than the funnel with the seller is what AI allows you to do. And then realize you deliver very specific messages with – very often customer stories, that are not just similar but really, really similar to the buyer.

Glenn: Okay. So, let's use that as an example. I've been searching for a platform for the better part of a year, I come to your site, where and when does AI kick in and how is it going to deliver to me what I should be getting, which is more towards the bottom of the funnel?

Ian: So, there are key things, how do I hyper-personalize the message to you and make it very, very specific? I think that when you talk to most marketing people, they agree that a very specific message to a very specific audience is better than a Superbowl ad, you know, where you're having millions of viewers and there's a tiny percentage of people that's relevant to. You know, people get that but you still see a really bland market. And so, I think the matching the content is really important. I always find it fascinating, I go around a show, I look at the message – which I really often reflect on the website and the stand, you know, I always ask myself, could I apply this to any vendor in the category or any vendor at all? And it's surprising how often you can, because they're just so bland and horizontal.

 So, the key thing here is, how do you – what we call – micro-verticalize the message? So, the key thing is, you know, you don't want to do one-to-one messaging, because it's just you can't scale that way. How do I have a cluster of very similar people with very similar problems that really relate to each other? That's what we do, we use AI to kind of target that kind of space with that micro-vertical approach. So, really the kind of analogy we have is, you know, when you think about focus – if anybody says they're focused on manufacturing, they really don't know what they're talking about. Because you literally take what they just said is, if you're focused on manufacturing, that means you're saying that planes, trains, cars, toys and pharmaceutical companies are all the same. And everybody gets that they're not. Whereas, you know, people kind of intrinsically get that because that's the things we see in daily life, but then you still see people say, I'm focused on financial services, which is equally as crazy because you're really then saying that Lehman Brothers, USAA and Wells Fargo are the same, which they're not. 

So, what you're really looking at to use AI to match the very specific message to very specific people is clusters of really, really similar people. And we have a whole process to go around that to match the content to the AI-based personalization. And really the key thing here is, what doesn't work is looking for zip codes or NAICS codes, what you're really looking at is, you know, synonyms and the really clusters of like-minded people that you find through research rather than zip codes or NAICS codes. 

And then, for example, with us, you think about at not-for-profit, that's very, very broad. And within not for profit, we can personalize the faith-based organizations and even catholic diocese and then what you have is a cluster of very, very like-minded people with very similar situations that self-reference each other. They probably have a very, very similar problem, but just at different stages in their buying cycle. So, we can match the – using AI, the exact kind of message, to the exact kind of people, at the exact stage they're in. And in financial services it's applicable as well. So, financial services is very broad and we can drive that down to asset managers who basically have a broad portfolio that involves real estate to family offices. That's really how we kind of blend the mix of very specific content with very specific messages to very specific audiences using the AI approach. The key thing-

Glenn: Is that making the assumption that they came in from a known source, maybe an ad, so you know a little bit of information about them beforehand or is this based on somebody interacting either with a certain page or maybe they downloaded something that's in a certain segment?

Ian: All of the above and an awful lot more. The key thing is AI gives you many, many things to look at. I think the key thing then, you really made a great example, which is, you know, do they come in from an ad? 

So, the key thing when you have this kind of ABM at scale, AI at scale, is that messaging has to be very relevant, as I've said, but people have to get in seconds because you're doing it at scale. Or we go through a process where you say, you know, think billboard not paragraph, think about the customer. And it's really the start of a conversation. 

So, an ad is a start of a conversation that somebody looks at for maybe five seconds and then maybe come to your website and look at something for 30 seconds, and they're looking at something for minutes and an hour. So, really, you have a consistent kind of content and think for that, the expanded conversation. 

And then really the key thing also is, it shouldn't be an explanation. And inherently it should differentiate you being unlike your competition. Because if you're competitors are bigger, you're saying, hey, we're the same, bit smaller. So, you have to inherently think, is it very understandable and is it inherently differentiated? We always look for, kind of, other kind of sources of content consumption for inspiration because it changes the way people expect to consume content. So, for example, you know, we believe Netflix has a massive impact on websites because you binge on content. So, when you're finishing a Ray Donovan, they'll either recommend the next episode or maybe Game of Thrones. So, it continues to allow you to binge the recommendations. And we need to think like that for our websites and the way we expand the conversation and the recommendations to people within this space as well.

Glenn: Is your content derived internally or do you use external agencies for this? How do you approach that? Because there's a cascading issue, right. Every time you go down a path, there's more and more need for content that's specific to that area. 

Ian: Absolutely. So, I think the key thing here – that's a great question – is, what we say our customers are our source of knowledge. We aim to be best-in-class in micro-vertical marketing by being best-in-class in gaining customers. So, you know, if the knowledge is in the customer's brains, you can really have great conversations with them. And so, that's what we kind of focus on is, that really, you know, having in-depth conversations, research and market by talking to customers.

Glenn: And how do you layer chat into this or do you?

Ian: I mean, on website we do. But I think the key thing before that, which is just kind of expanded from the previous questions is, if you outsource your interviewing of customers, which is probably one of the most important things you do, you lose a lot of knowledge. So, we've been fundamentally critical to that internally because there are certain questions that you ask that you get amazing answers to that you just lose if it was outsourced. So, you know, we believe fundamentally that that kind of process is really understand the customer really, really well, micro-verticalize the interview process and then identify the micro-verticals, drive it into product marketing for then the external content that's going to be generated, drive it into demand generation for the content that's been generated by product marketing. So, you need this whole kind of cascading process. And it has to be done, to be best-in-class, to be done internally.

Glenn: And you do that with the combination of surveys, both, online and maybe phone?

Ian: We do pretty much exclusively phone or actually we do a bit of both. But the key driver for us is really in-depth phone interviews with customers. And we do kind of quarterly and annual surveys. But we see the customer interview process as being fundamental to what we do.

Glenn: And is that typically being done by more in the marketing and communication side or more in the product marketing side in terms of the actual calls to the customers?

Ian: So, I actually went and document them in the book Crossing the Chasm that we work with Geoffrey Moore very closely, about how to go to market. And the key thing we learned there was, you have a group of people dedicated to understanding verticals and micro-verticals and we call that our industry marketing group, which is separate to product marketing, separate to demand gen as well. So, that's just dedicated to discovering and understanding in great depth verticals and micro-verticals.

Glenn: Got it. No, powerful stuff. So, circling back to chat now. Do you use chat on the website and tie that all in? And is that driven by AI as well or is that when a human being can pop in or a combination?

Ian: Yeah, we use web chat on the websites, because one of the key metrics is AI is interesting that it allows you to do analysis, but also it generates a lot of data, which people kind of often forget about. So, it allows you to see how important it is to – when somebody hits your website to get back to them very quickly. So, chat is very – kind of a core way to respond to a customer very quickly. So, we integrate that into our website and have a mix of AI and humans.

Glenn: What's interesting about the data is you can get buried in it. And so, as the senior marketing leader what is your dashboard focused on? What bubbles up to you to tell you where to dig?

Ian: So, I think the key thing is hiring great people that can filter data. But for me it's, you know, within marketing looking at the whole opportunity model is very important and then looking at it from the full funnel from lead, to MQL, to up to QDC, to opportunity, to close then looking at it by micro-vertical. And I think the other critical thing is, making it really, really easy to access the data. So, if you have to download it into Excel and do an analytics, it's so painful. You don't ask the questions, you do it once a month. The key thing is we should be looking at least every single day and when things are going up or going down, really asking the question, why? And having the drill through capability to understand why very quickly because the market changes very quickly. So, we have to access by opportunities, we have to access by the full funnel, we have to access by looking at it through micro-vertical approach and then the kind of the key levers that you use to drive it and then look at it every single day.

Glenn: We're seeing more and more senior marketing leaders really having to have some technical chops and where do you see that? Just to give us context, did you grow up on a particular marketing side, was there an aspect to marketing or have you done it all in marketing?

Ian: So, my background is I've got a PhD in databases and in computing, so I'm kind of a-

Glenn: You've got those chops.

Ian: So, I started life as an engineer. And I think in the database industry, so if you're an engineer you just look at things in a kind of a different way, you're much more kind of analytical process driven, what AI companies does well. So, I think an understanding for and a feel for data is really, really – the feel is very important. You can just see things very, very quickly, getting kind of obsessed by analytics and looking at it from the high-level and then drill, so it's very important. 

So, for me is, the key thing is hiring really smart people who are kind of very data-driven that also have talents across the full range of marketing. So, marketing is interesting, you have a number of skills and acquiring them is very important, because it gives you a – even if you're not doing it currently, it gives you a view across marketing as well. So, I think what we're seeing here is kind of interesting, you think about traditional enterprise marketing is much more focused on, you know, interacting with the customer, then we saw massive high volume by marketing, particularly through the source space where I had some experience. 

And then what we're seeing now is that kind of – and then we saw now is kind of a wave around ABM. So, we think about all those three put together, I think what we're starting to see is we call it ABM or AI at scale how can I get the hyper-personalization of an ABM but do it at the scale of an inbound marketing organization. And I think the, from an experience perspective, it really, really helps to understand ultra-high-volume inbound marketing which is open sourced epitomized all over that. And then SaaS as well, with the messaging and understanding of an enterprise software kind of persons. We get people with mixes of skills from each of those kinds of areas, which are very different. And having the blend of insights from across – one of our great guys worked at Zinger and at CIBIL. So, that's a good example of two totally different companies. We have a different insight on things.

Glenn: Yeah. Ian, where do you see AI going? Over the years technology never tends to be done, right. So, there's always opportunities for growth in any technology. Where do you see that going and/or where do you think it needs to go that kind of get it to the next level?

Ian: That's a great question, I actually thought about this over decades. For me it's you can see a pattern emerging over the last 30 years. And one of the biggest changes we've experienced in our lifetime, it sounds crazy, the relation database was ginormous. You suddenly database and ask questions without going to an IT guy and waiting for a week. And the first impact when that was – running queries, but the next really big impact was, it was really easy to run applications on top of it. So, you saw CIBIL and SAP and PeopleSoft. You saw a whole generation of applications written on relation databases that were fundamentally better than what you had before. But they didn't call themselves relation database applications, they were just better at what they did. So, that was one kind of phase one. 

I think what we then saw was applications written on software, you know, big data infrastructure and they were just fundamentally better at doing big data analysis. And I think what we're going to see here is applications written on top of the machine learning infrastructure or AI infrastructure, which is just fundamentally better at doing what they do. For example, in our space knowledge and techs is very well understood, you're looking for outliers. So, you know, within the general ledger there are – some industries clearly say that, you know, 10-20% of entries in the general ledger may be incorrect in some way. Anomaly detection can discover those very, very effectively, but you won't call it anomaly detection, but you'll just have a simple application to go through your GL and highlight those that are most highly likely to be incorrect. 

So, I think we'll see these kinds of micro applications built upon an ML or AI infrastructure applying to various verticals. And we'll see that in timesheet management, we'll see that in anomaly detection around the general ledger, we'll see that in getting insights that we never even looked for. 

For me, I wrote something quite a few years ago that kind of epitomize a lot of what AI can do for us. It's bizarrely as a phrase by Donald Rumsfeld, everyone knows he's like, you know what you know, you know what you don't know, but you don't know and it's like – so, when you look at that from an AI perspective is you maybe traditionally you'll ask five questions once a month, AI can ask thousands of questions you've never even thought of and find the things you weren't looking for that are fundamental to your business. So, I think that's what the kind of insights that we can get is, you know, the things we don't know that we don't know that can have a ginormous impact on our business.

Glenn: From your perspective, and knowing that AI just, both, creates a lot of data, but frankly needs a lot of data in order to be truly effective. Do you see that data residing within the company or to some third-party AI platform such as an IBM?

Ian: I think it will resolve – you know, we're finding great success by looking at the data within – what we call intelligent general ledger, intelligent GL. So, I think the key thing is looking at A, the data, and we can import external sources of data in as well through some static ads. Then a key thing I've learned through AI over the years is how you structure, what they call features. So, you can take raw data and you can look at – take a look for fraudulent transactions, you can look at it through “day of the week” code, every column can be a day. It's the way you think about structure, the data can be very important to increase the accuracy of your insights in the data.

Glenn: Ian, this has really been insightful. If there's one thing that our audience could put into action today to really have impact with their digital marketing, what would that one thing be?

Ian: So for me, there's one thing that I always believe and I believed for many years is just hire great people that are multi-talented, that have done lots of different things, that are intellectually curious and then give them the common goal and ask, we want the best in software industry micro-vertical marketing and then just align them around that common goal. So, just hire great people.

Glenn: Awesome. Ian, if people want to get in touch with you, what's the best way? 

Ian: So, my Twitter handle is Dr. Ian Howells – @drianhowells.

Glenn: Ian, thanks so much for being on the show. 

Ian: It's been great speaking with you.

Ian’s Bio: Dr. Ian Howells is the Vice President, Head of Marketing at Sage Intacct, Inc., a cloud-based accounting software for small and mid-size companies. To reach Ian, you can find him on Twitter

Want to appear on AMP Up Your Digital Marketing? Contact Us and let us know how our listeners can benefit from hearing your message.

The predictive index employee advocacy case study

Ramin Edmond

Ramin Edmond

Ramin Edmond is the former Content Strategist for GaggleAMP. Outside of work, Ramin likes to run, hike, and take pictures of Boston's best views. You can get in touch with Ramin by connecting with him on LinkedIn.

Thinking About Starting an Employee Advocacy Program?