See all Insights

The Firm of the Future, Part 2: This Machine Will Tell You What To Do Next

What is the future of marketing? Well, there are of course lots of potential futures, but our money is on machine learning being a key theme. We began our pursuit of and investment in machine learning two years ago and we’re excited to share the journey, results, and future predictions with you during this free one-hour webinar.

Join Chris Butler, Dave Mello, and Mark O’Brien as we discuss our thoughts on machine learning’s impact on the discipline of digital marketing, and showcase the tools we’ve created to make the most of this exciting new development.

Webinar

Unlock this Free Webinar

To access this webinar, please complete the short form below.



Transcript

Transcript

Mark O’Brien: Hey everybody. Welcome to the Newfangled Webinar, ‘The Firm of the Future’ Series. This is Part two, titled ‘This Machine Will Tell You What To Do.’ This is Mark O’Brien, and I’m not alone today. Once again, I’m joined by some Newfangled friends, and today we’ve got Dave Mello our Director of Technology. Hey Dave.

Dave Mello: Hey there, how’s it going, guys?

Mark O’Brien: Great. And Chris Butler, our Chief Operating Officer.

Chris Butler: Hey Mark. You are not alone.

Mark O’Brien: I’m not alone. I’m not alone. It’s great. It’s always fun when we get to go group webinars like this, and today we’ve got quite a bit to cover. We’re going to start with Chris’ sabbatical. I guess it was 2016?

Chris Butler: It was 2015.

Mark O’Brien: Oh, 2015.

Chris Butler: Actually about in the spring of 2015, Mark brought up an idea of me taking a sabbatical and frankly it was because I was pretty burnt out at the time. It was a good idea, because I, up until that point, had really not taken a substantial amount of time off and needed to, so the initial intent there was just to recharge, and they say that when you get away from work, especially if you’re burnt out, you need a couple of weeks to really recharge, assuming you’re not thinking about work at all, and I found that to be true.

For me it was about two weeks in, literally it was the Monday of the third week, and I was walking my dog and I had been reading a couple of books about machine learning and listening to a few podcasts that were bringing that up and all of a sudden I just started to think about all kinds of ideas that would be relevant for Newfangled, and it just sort of hit me. It was like a light switch flipped, so I think that two week thing is a meaningful measure.

On that day, I came back to my home office and I ended up filling up a couple of books in my notebook, just with some ideas on how that kind of concept could apply to the kind of work we do and the objectives we have for our clients, and so that was a really interesting explosion of ideas and brainstorm, and as soon as I got back, I put it all on a whiteboard in my office and sat down with Mark and Dave and just talked about these ideas, and thought about how we might apply machine learning to the kinds of things that we’re doing, but also the problems that we were seeing at the time.

We had some issues with trying to get the best intelligence to our clients in a timely manner, so that was a couple of years ago, and we’ve been working diligently since to build a system, that reflects those ideas, and it’s evolved heavily from the initial idea in working with the two of you as well as some other people here at Newfangled, I think we’ve put together something that not only represents the best of those ideas, but is not way better than even those ideas were because of everyone’s contributions here.

Mark O’Brien: Sure, absolutely. First thing, you cannot stop Chris Butler from working, no matter what you do. The less he works, the more he gets done, which is wonderful, but this machine learning thing, this is a phrase that gets thrown around all the time now, and I think the majority of the population doesn’t really know what it means. Dave, can you give us the layman’s version of what machine learning is?

Dave Mello: Sure, I think for what we’re talking about today, it’s really creating a system that will be able to find patterns out of large sets of data for you to be able to find what’s important, and really bring that to the forefront, so when Chris came back, he was talking about being able to spot patterns within sent emails and stuff and basically just being able to pull that from large sets of data became challenging.

Chris Butler: Yeah, the difference between what I think we typically have done, and firms like us have typically done from a programming standpoint and a machine learning approach is that, typically you think of a query you have or something that you want to do with data that exists, and you process that query, and you might automate that process happening routinely, but you don’t write those queries in such a way as to allow the process itself to evolve.

You think about well, I want these three data points. A, B, and C, and you just write out a query that says get me A, B, and C on this whatever. I don’t have to hit go every time. The idea of machine learning says, “Okay, given A, B, and C, and maybe every other letter to Z, what will you discover and what connections are made over time by routinely pulling that data, meshing that data together, and what can the machine tell us base upon having empowered it to do so, so I mean there’s a ton of coding that happens in order just to allow the machine to do those kinds of inquiries, to give it the latitude to do that, because it’s not a person.

Machine learning, I think a lot of people think of that and they think about artificial intelligence in the most proper sense. A machine that things for itself. That’s not really what it is. We’re thinking it, we’re creating the illusion of a thinking machine why giving it a little more querying latitude and processing latitude. Is that a fair way of saying it?

Dave Mello: Mm-hmm (affirmative).Yeah, of course.

Mark O’Brien: I think there’s two branches to the prediction element, the telling you what to do. We’ve got the sales side of it, and this is something that I use all the time now, and this tool is completely changed the way we operate as a biz dev organization, and that’s really exciting, and it’s funny, so many of our clients and we ourselves rely heavily on person-to-person referrals and relationships, that kind of thing, and a typical argument against digital marketing and content marketing and marketing automation and a CRM approach, a typical argument against that whole set of approaches is that oh well, we’re a relationship=based firm, and that’s too sterile or whatever it is, and what I’ve found to be really interesting about this specifically, is that it does tell me what to do, and it tells me when it’s time for me to pick up the phone and get in touch with somebody.

It actually assists the personal relationship, which is really cool and exciting, and now, it’s as common for us to close business that resulted from me getting in touch with somebody, because I saw what they were doing, and I saw the patterns based on the system, as it is for us to close business from people who voluntarily got in touch with us, which is pretty amazing. I would not have guessed that two years ago when we were siting in your office talking about this whole thing. I would have never expected hat would be part of the result, but it is.

Chris Butler: Right, well, one of the biggest contributions, Mark, that I think you brought to how this system has evolved is that when I came back, I really was thinking of this more as a marketing system and tool, and less as a business development tool, and you’re right that since then, in terms of this machine will tell you what to do next, that’s true for a marketing and a business development person, and that’s something that I think is a natural evolution that was brought to the intent of this system, because what we’ve discovered in our time pursuing this kind of work is that the distance between marketing and business development is just narrowing. It’s so tight.

The other thing I was thinking about is things that firms like ours and our clients struggle with is making sense of the data that they have at their disposal, and we were doing about this in a podcast episode a bunch of weeks ago, for those of you who listened to that, that our management team here used to spend a lot of time arguing abr what’s true, and the reason we did that is because netdeep in data, and we didn’t really have any guiding principles sifting through it and figuring out what’s actually true.

The more data you have, the more insight you don’t have, you need some kind of told to help you bring that insight, and so, machine learning, that’s really what it’s for is to help human brains make sense of data that is beyond the human brain’s capacity to sift through.

Mark O’Brien: And in terms of the roles the three of us played in that, just to be super clear about that for the audience, so Chris conceived of this on a sabbatical as he mentioned, and then really you and Dave did a lot of back and forth brainstorming, and Dave’s done all the development on this, and Chris has done all the design on it, and so there’s tons of back and forth, it’s hard to know where one begins and one ends, although, the strategy was really shared. You guys had a lot days going back and forth, but you built it all. You definitely designed everything.

Chris Butler: Correct, and the conversations extended into our leadership team. Mark, you played a role in shaping it in terms of giving us a sense of what would be useful to be seen. Additionally, our former Director of Strategy Chris Creech played a big role in helping us shape some of the algorithmic approaches to evaluating marketing activity that we’ll show later on. Lauren Siler played a role in contributing, Holly Fong has played a role. Dave and I don’t spend a ton of time in the marketing automation space.

Dave Mello: Right, we really have relied on them a lot to really bring their knowledge to the forefront here.

Mark O’Brien: Mm-hmm (affirmative). Yeah, my role in this was ‘Hey this’ll be cool.’

Chris Butler: Everybody needs that.

Mark O’Brien: The way it works at Newfangled is someone says, “Hey this’ll be cool,” and then Dave, before you even get the work cool out, Dave has it built. That’s what happens, and it’s really exciting, and that has bene helpful because I’m responsible, along with Lauren, for business dev here, and so I am always thinking about what tools I would need and I would use in order to make that as efficient as possible, and it seems that my business development needs to do parallel pretty well with the business needs of our prospects, so if I think something’s cool then others will think it’s pretty cool too, so that works out well.Yeah, the whole teams been involved, and it’s been very exciting, so let’s start looking at some of this stuff.

Chris Butler: While you’re pulling that up, just as a reminder to those of you who are listening, the mission of this thing was pretty straightforward. We’re talking about human limitations, and something about Newfangled’s model that is sacrosanct to us is interfacing with our clients and all of our strategists meet with our clients once a month. Some of you may be on the phone right now listening, and there’s a limit to how effective you can be duplicating that experience.

Once a month is a good clip, but you need intel more readily than that, and that’s what the system is built to do. It’s built to give you the answers you need when you need them, which isn’t just once a month. It would be every day. It could be when you take a certain action, and so we wanted to employ machine learning to better analyze your prospect patterns, to give you that intelligence when you need it. That is what this system’s mission is.

Mark O’Brien: Here is how I wake up every morning, is it really is the first thing I look at after I get up, and this is the daily activity report, and so as you can see here, we’re blurring out the names of the innocent except for John, who is definitely still innocent but we got permission from RP, who’s an awesome firm out of the Toledo area. We were just visiting with them, they’re a client of ours, and they’re exemplary in every way. We love them, and John’s been really active in our work together and he’s been very active on the site doing a lot of research, which is great, and we can see that, obviously, but basically what I do each morning is just check this out, and say, okay, who, from yesterday, of note was on the site yesterday. The list is pretty long, and so we’ve got the number of pages that he viewed, that’s the 11. Their name, their title, their firm, and their location.

What happens is, as I get these emails each morning, and as I get the website conversion form email, so anytime someone downloads a white paper, registers for a webinar, signs up for the blog, we obviously get those emails as well, and it just allows me to develop a peripheral sense of what’s going on/ this is the third time someone from this company has been to the site this week, or this one person seems to be doing a lot of things, what’s going on there? When I get that peripheral sense that this person is of interest, then what I do is I launch right from this into this page that was created.

This is the insight engine. Now we’re live inside of the insight entire that we created, into the lead details page for the individual, and so as soon as someone converts on the site or clicks through from the email, we’re able to find a lot of things that are available for that individual, so image, name, title, those things, location, different social media links oftentimes you can see what’s publicly available for them. It might be Facebook or twitter or linked in or whatever. LinkedIn is the one I care about the most.

Where they work, they’re tenure there, and then, this si what I use all the tie. actual company information, so this is pulled in from various sources, so Dave, can you, without giving away all of our secrets, can you talk a little bit about how you created this and how you created all this different datasets?

Dave Mello: Sure. We tried to not reinvent the wheel as much as possible, and there are services out there that will give you a company insights, company intelligence, and also social intelligence, and we’ve tried to leverage those as much as possible, but in a way that actually presents a pretty good story here. LinkedIn, as you said, is the primary driver for a lot of the information, and I think rightly so, because as you said, that’s the information that you really care about.

Mark O’Brien: Absolutely. Down here, we can actually see everyone from this firm who has converted, so I can do ahead and click in to Stephanie or Martha or whoever else I wanted to, because we have tracked all them. Additionally, if I were to start searching for them up here in the search bar, then I might be able to see different things that people are associated with that organization as well, so if you didn’t come on because you were directed to by a web form or something, you could just go and say, well I wonder what this company’s done lately, and just start typing the URL, and see all the individuals prioritized by recency, which is really helpful, but in any case, when I get here after getting interested in looking at somebody, I can spend about 10 seconds here, we’re going through this in detail, but I spent maybe 10, maybe 15 seconds total looking at the screen.

Over here we’ve got a little story about how they’re engaged on our site, this is one of those things when I aid to Dave, hey, it’d be really great if right from here we could email somebody or view their LinkedIn page or go to their website right from here. Also, we could click right into Salesforce and view the opportunity and he had that update by the hour which is really exciting, and then there are conversions on our site, and down here, this is a tool that we actually created quite a while ago that continues to bring a mass amount of value to Newfangled is this overflow activities timeline, where we can actually click a play button here and basically watch a video of everything this individual has done on our site from the first time they came to visit to the most recent time, and so that, again, just gives me within a few moments a really clear idea of what this individual’s about, what the firm’s about, and then, does it make sense for me to get in touch with them?

That’s the handoff period, and after this, as soon as I looked at it, I basically make a judgment, and do I want to get in touch or not, and this is a question we get a lot from our prospects. Well, you know, you don’t want to get in touch with somebody and say you just watched a video of everything they did on your site, and that’s true, you don’t, you don’t.

Eventually we show this and it’s kind of interesting, but when you’re looking at all these conversions, you can choose something, like, oh, you know, you registered for the webinar or downloaded a past webinar last week on aligning marketing and sales for business development. It’d be very natural for me to get in touch to ask how he liked that webinar, and of course any prospect would know when to convert on an asset like that, that is reasonable provided the asset would be aware that that conversion happened, so it’s not spooky or weird or any of that stuff at all. It’s very, very natural.

What I’ve found has been interesting, it’s so far been a bit of an all or nothing sort of situation where either the prospect I’m emailing will get right back to me almost immediately and be really excited that I got in touch with them, and we’ll have a call and well go from there, or, I’ll hear nothing at all. That’s fine, and I don’t put them on an automation queue, or any of that, I don’t bombard them, I just send that one feeler, and if they’re interested, great, if they’re not, I’ll leave them alone.

Maybe a few months later if they’re still doing a lot on the site, I will get back in touch with them, but we’re not terribly aggressive about it because it’s just not our sales style. So this screen is great for me in terms of the machine telling you what to do next between the email I get every morning, all the conversion forms I see and then this screen right here, this is where I spend a lot of my time from a biz dev, form a sales perspective.

Chris Butler: Right, and if John was the contact with whom you were interacting on a Salesforce opportunity, that opportunity would also show on the screen. We are pulling that information in from Salesforce. It just so happens that John was not the contact with this opportunity. That’s a really helpful thing. An opportunity might be something that is conceived and closes in two weeks, that may not be that useful to you, but we’ve also seen sales cycles that can last up to a year, in which case you do want to login and see, okay, as this opportunity is being nurtured and endures over time, how does the content of the site play a role in that opportunity? What are they doing as they’re having these sales conversations?

That’s especially true for an opportunity you might be nurturing over six months, so we’ve really found that a tool like this needs to be aware of its role at various stages of that cycle, of that opportunity nurturing cycle, and something we should talk a little bit about is that so far, we haven’t looked at anything that’s machine learning oriented. What we’re doing here, is we’re doing synthesis. We’re going data synthesis from all the sources that are relevant, and the reason we’ve done that, the reason why that’s valuable is because someone like Mark or Lauren or even someone on the marketing side, we’ll talk more about that in a minute, the data they need in order to take the right actions is all over the place.

It’s trapped in all kinds of different silos. It might be in six different systems, and what we’ve done here is to build a system that intelligently curates that data, puts it in a different design so that it actually can support the needs of those two types of roles, and that sets up the machine learning stuff we’ll talk about in a minute, but so far, we’re just doing really good synthesis.

Dave Mello: These are all data points that then, like, you Mark, do what you do best, which is basically understand what to do.

Mark O’Brien: Right, and something we also talk about all the time is, when we’re speaking with different firms about their marketing benchmarks. You need to have 3 to 5,000 words of content each month. 3 to 5,000 contacts in your system. You’re emailing three to four times a month, 40% of your traffic come from search, you have a 3% conversion rate on 2,000 monthly visitors. All these numbers that are huge, huge, huge, huge numbers, and the biz dev person’s thinking like, I’m not going to follow up on any of that. I’m not going to look at those people, I’m not going to do any of that, and I don’t. That’s really great. I look at this, and that simplifies things greatly for me, and what it also does is, gives you piece of mind, because prior to this, I’ve always had the feeling like, all these people, these thousands of people, which ones that I’m sure there’s some that are ready for me to get in touch with that haven’t quite made that last step of getting in touch with us, and just, there was no way to do that before, and now there is.

Dave Mello: No, but we would use different kinds of tools. There are built in scoring mechanisms in other platforms, but none of them really got to the heard of the matter. None of them saw the patterns, and that’s where this shines.

Chris Butler: Yeah, we’ve been talking a lot behind the scenes about, well, struggling really, with the concept of lead-scoring over the last year or two, because lead-scoring is incredibly important aspect of marketing automation, but what it doesn’t really do is give someone on the biz dev give a sense of when they should get in touch with someone, so our perspective on marketing has been that marketing’s purpose is to deliver right fit prospects to the company so that they can be nurtured as an opportunity, and what it doesn’t do is tell you when.

It just brings them to you, and you don’t know if that moment they’re right, or if it’s later, and so what we’re really doing here is we’re doing something that’s not lead scoring, it’s more opportunity scoring. That’s different, and I hope to see that as a concept become a little bit more mature, because I think there’s a meaningful in lead scoring opportunity score.

Mark O’Brien: That’s a very good part. Let’s shift over to the marketing side of things here a little bit, Chris or Dave, could you walk us through what we’ve got here? What we’re looking at?

Chris Butler: I’ll talk about where this came from and I’ll talk about how Dave made it better. This first report here, what we were trying to achieve was have some kind of visualization of the various data points that are going to tell you if what you’re doing is working, fundamentally. What I’ve found over the years is that a marketer might start with something like Google Analytics, which honestly, I feel, is becoming less and less useful.

Google Analytics would be really powerful if we were doing eCommerce and wanted to do some really unique goal funnels. They might look at Analytics. They might look at their marketing automation suite and see how things are going in there. They might look at their email tool if that’s a separate tool. They might look at their CRM. The problem is, is that you really just need to figure out well, what is the expression of the story that I want to tell, and here we said, well, it’s really a combination of our publishing volume, the known visitors that are coming to our site, and those who are taking actions on the content we’re sending them.

What you see here is you see that synthesis. The known visitors are the light green, so that gives you the full volume per day of known visitors that are coming to your site, so known visitors being those people that we are tracking intentionally. The darker green with the happy face, those are the people who are taking action on that content, and so you can see how those patterns over time ebb and flow based upon never volume. Every firm publishes at a different pace, publishes a different amount, they might publish evenly week to week, they might publish evenly month to month, it depends on their culture and what their content strategist here has helped them to determine.

This is just giving you a sense of hey, big picture, where is the activity happening relative to the activity we’re putting out there, and it’s just a match of our activity vs. our prospect’s activity. Now, Dave was able to take this stream I had up as visualization and actually built it because this data is coming from a lot of different places.

Dave Mello: Correct. Yeah, so, we store a lot of this data locally, so that we can do stuff like this. We do pull from different sources, but really the goal of this is to have our own data silo so that we can create the reports so we can kind of learn from it in the ways that we want to, and we didn’t really know what all those were going to be when we started, so that was why, creating our own data silo made the most sense, because we didn’t know where it was going to take us, and that’s proved to be pretty successful so far, and that lets us do things like this.

Mark O’Brien: That’s something that’s so interesting to me, is all these different tools that you’ve created in the background. The different parties we’re using, and all the private data silos we’re creating and all of it, I envision you physically having this warehouse of stuff at your disposal, and to me, it’s all very mysterious, but it works.

In a big picture, we have our digital marketing model and that’s the six elements you hear us talk about a lot, which is the foundations positioning and then we’ve got the contact strategy, the content strategy and then the three tools of the website, the CRM, and the marketing automation tool, and then we have all these benchmarks, but what it all really boils down to is the content you’re creating, the emails you’re sending, and the activity that those two things are generating. That’s really all we, and they, care about, because those re the things that move the needle, and the needle is the personal activity, and so you’ll see as you go through this, that’s what we’re really trying to focus on is content, email, and visitor activity.

Chris Butler: Right, Google Analytics doesn’t know really what your intent for publishing is. It knows what pages exist and it knows when they show up, but it doesn’t know that your intent is to publish 3,000 words of indexable expert content even month. It has no idea. It doesn’t care. It’s totally agnostic to marketing intent, it’s just, all it knows is what’s where on the site, and that’s why it’s not really a good marketing analytics tool. It’s a good analytics tool, at a very foundational level, but that’s why pulling in here, the published word count gives you a sense of, are we actually making good on the promise that we made to publish that level of volume, and how does it affect our outcome?

Mark O’Brien: This screen here too is really helpful. We’ve got a few different reports: people active now on the site, the top page views, lead score changes, I’m just going through it quickly to protect the innocent. Salesforce opportunity visitors, conversions, referrers, all sorts of different things here, and actually let’s take a look at some of these with some borders in place so we can talk about in detail.

Active now. This is really exciting. When we actually publish content on a day we publish content, particularly within a few hours of publishing the content, they’ll be dozens of these little people with the green dots. The green dots mean that they were active within the past 15 minutes or so.

Dave Mello: I think it’s about an hour.

Mark O’Brien: The past hour.

Dave Mello: We don’t know exactly how long they’ve been sitting there reading the page. We can be a little more liberal with that.

Chris Butler: The standard timeout is 30 minutes on a page, so we have to do plus or minus.

Mark O’Brien: Then, when you mouse over these individuals, you see little breakout, and of course, when you click on them you go to that lead details page, which is looked at in detail, so that’s really helpful. Then we’ve got the top page views, and this is basically the physical report of that email I get but it’s live, so I get an email at the beginning of each morning that talks about yesterday, but I can log into the system anytime I want and I can just have it open in a tab all day long and I can go in here and see who then again it’s the page views for the day, but as you can see, you can toggle back and forth between days, you can show data by week or by day specifically, depending on how much information you want to see, and then again, it’s the page views, the visitor information, company information, and the location information as well, and somehow LinkedIn thinks you still work for Smash Magazine.

Chris Butler: Part of the reason for that is on my LinkedIn account, I’ve got a variety of different things there, and obviously this is pulling from a variety of sources, so that avatar is my daughter and in my podcast.The reason why is because the active cookie that is being looked at here is using my personal email address because I tend to not separate the two so I’m on our site using my personal email account all the time.

Dave Mello: We don’t show any of our internal people for our reports.

Mark O’Brien: Right, we filter all Newfangled activity, and I don’t have a personal email.

Chris Butler: You do. It’s Mark at newfangled.com.

Mark O’Brien: Then lead score changes. This can be really helpful just to see of all the people again, we’re trying to separate the wheat from the chaff and find that needle in the haystack and whatever other expression there, but we want to see who’s moving. Of all the people who are involved with us, and who are involved in our marketing tools system, which are the people that are taking big leaps and lead scores, so that means that they’ve all of a sudden binged on some content or hit high-quality pages, whatever it may have been, and this is really helpful again, just to keep track of the people we care the most about. Any other points you want to make about this?

Chris Butler: Well, again, you can see that these three that you happened to choose are actually in our opportunity world. They’re in Salesforce. You’re actually interfacing with them and nurturing those opportunities.

Dave Mello: Those people you’re probably pretty familiar with, but there can also be people that haven’t been on the site for three years but were an opportunity, and that’s also really interesting because they didn’t qualify then but maybe they do now.

Mark O’Brien: That’s exactly right. That’s happened quote a bit this year also. I’ll see former clients. For whatever reason they went to do something else, and we’ve changed a lot over the years as well, and I’ve used this for outreach with them and bring in a number of relationships coming from those other clients again now, because we saw that they were dormant for years but then something caught their eye and they came back and then I got in touch, and that’s actually, in terms of the speed to close, that’s among the fastest, the former client coming back, and because you know them, they know you, there was a good relationship in the past, when that person shows up and you get in touch with them and the need is right, that can be a very, very, nice close and a very quick close as well, which is exciting.

The actual conversions on the site, I think that’s pretty self-explanatory. That’s what we’ll show there, so now let’s get back here. One of the things we can look at live is companies, I don’t mind showing what companies visiting. We can go here and actually see which organizations are on the site, and the green dot, again, means they’ve been on in the past hour, which is many people here, I know many of these people are on the  webinar right now, which is great.

Dave Mello: This is actually live, it is actually cached by a few minutes, so don’t expect the site to see your company show immediately.

Mark O’Brien: Right. We’re not refreshing. That’s super helpful. Let’s get into emails and content, and then we’ll take some Q and A after that, I think.

Chris Butler: Sure, this piece that we’re about to see here is where the machine learning stuff comes into play. It’s also the stuff that’s most directly linked to that dog walking brainstorm I had, because what I immediately thought of when I was walking my dog is I had my smartphone in my pocket, and I was thinking about this book I was reading about machine learning, and I was thinking about this book I was reading about machine learning and I was thinking about a podcast I had heard that morning, and my phone buzzed. I got a text or something.

All of a sudden, the pieces connected where I thought, “You know, we have this opportunity to evaluate what we’re doing on the marketing side in a way that emulates the experience I just had, which is up to the minute communication,” and I put all those pieces together and thought, “You know, we do all of this outbound, and we then sort of ad hoc go back and do data analysis on it, and we help our clients do that once a month. Most of the time, we don’t do it because we do so much outbound, and we don’t really have a good way of objectively analyzing that pattern so that we can improve it over time.”

For instance, we have meetings internally where we’re still sort of debating baseline assumptions about outbound activity, like, what day or time is best to send on or the length of subject line or how long the email should be, or how many links it should have and we don’t really have any good way of confirming our suspicious of those things, because we’re human beings and we don’t know how to analyze that data, so we wanted to build something that would do that.

We’re going to look at a couple of things here. Right now, anyone who’s using the system, and that’s ourselves and a list of our clientele, every time they do an outbound campaign, they get an analysis on it from the system automatically, and the system is going to look through a variety of data points to tell them what worked about that campaign and what didn’t, and what they’re ideally using that to improve the next one, and this system will improve it’s ability to analyze it as it has more data to work with, so the more you do your outbound campaigns through the system, the more data it has to evaluate it’s assumptions about subject line length or character length of the actual message or the number of links.

We have a baseline recommendation that has some fuzzy logic number sin it, like a lot of plus or minus or ranges, but the more it looks at the response to your activity and the more it actually crunches your data, the better it’s going to adjust its recommendation and scores to your system. That’s what machine learning is. As you can see as Mark’s scrolling through, all the different points that we’re looking at. We’re looking at open rate. We’re looking at click through rate, we’re looking at spam rate, we’re looking at the link count, the image count, opt-out rate, the message length, and you can see some of the algorithmic detail on the right in terms of how we generate a score. Big picture, what’s awesome about this is you’re a marketer, you send out that campaign, and you start getting return on that campaign.

You start seeing people click through to things or you start seeing forms being filled out, but you really don’t know globally because you might have sent it to 15,000 people, and if you have the impression that 35 people might have come through since last night, you really don’t know if that’s good or not. That’s the whole point. I might have noticed a bunch of them because it was before you went to bed, and then the next morning, they’re 10 more but they’re all threaded in the same conversation. You really don’t know how well it did. This thing will tell you, that’s the whole point.

Mark O’Brien: This roundup here of all the emails sent is really helpful too along with the grades. That’s really helpful to see. Then we click into any of those if we’d like to. We can choose this one that we sent out recently that has an interesting style. I like this a lot, the email preview pane, so that’s helpful, and this an example of this plain-text email we sent last week, and it was one of the more effective emails we sent, it independently I think generated five leads right from the email. When I say lead, I don’t mean someone converted, I mean five people saying, “Hey, let’s talk about this,” which is really strong for us.

Chris Butler: Also counterintuitive for the typical content marketing approach that we take because this is a very specifically promotional email.

Mark O’Brien: Yeah.

Chris Butler: Yeah, this is a sales email. This isn’t content marketing like, “Let’s learn about this topic.”

Mark O’Brien: Yeah, Blair Enns has that ratio that you should send one promotional email to every three educational emails and we’ve done a really bad job adhering to that, meaning we never send anything promotional.

Chris Butler: Never.

Mark O’Brien: As we learned last week, we need to start doing that, so everyone on the webinar, just so you know, it’s not all going to be straight education. Getting up to 25% seems like a lot, honestly, from where we are now, but in increasing the promotion has been really, really helpful and exciting.

Chris Butler: It’s also relevant to a theme that I think we’ve talked about as an organization for a while, which is if you do content marketing right, you’re educating your prospects and that’s great, but they still need to know what you’re selling. They still need to know what you do for money, and you need to find intelligent and really straightforward ways of connecting the two, and we have ways of doing that in the structure of content marketing, and I know our strategist talk to our clients about that, but one of the easiest, dumbest, straightforward ways is actually sending emails that say, “We sell this, and here’s why.”

It’s also the way. There’s a bullet point list and there are pain points but I think it’s elegantly expressed, and I would imagine that a lot of people read that and said holy moly, these people know what I’m dealing with, I should give them a call.

Dave Mello: Just a second. We’re not showing this here, but we are seeing that there are different types of email that need to be scored differently. That’s something we’re currently working on is, how do we make that pretty much across the board for all of our clients in being able to categorize them, and as we see here, this is a very different type of campaign with different requirements and also different scores.

Mark O’Brien: Good point. Right. This would be very different from a blog digest or a webinar promotion or a white paper promotion, yeah.

Chris Butler: Yeah. We were talking about that beforehand. One of the things we wanted to share in this webinar is where is this all headed next, and machine learning is interesting because it assumes change. That’s what makes it what it is, and so this is an evolving system. We need to evolve it’s intent as well as how it does what it does.

To Dave’s point, we got docked on the email that Lauren sent the other day because our algorithm thought that it was too long of an email. It shouldn’t really think that because it’s a different type of email so it needs to be more sensitive to intent and it needs to be able to connect the dots on its own. It could do that if you sent more emails of different types and actually flagged them in certain ways, and we want to help it do that, so that’s where it’s going.

Mark O’Brien: Yeah, and this is interesting for us too in terms of our development, because this is a true software as a service platform, where, as we’re making updates everybody gets all the updates at once as we’re making them which is pretty cool as opposed to just building websitesYeah, it’s a whole different approach, which is really fun.

Let’s take a look briefly here at what’s going on with content, because again, we talked about, it’s really about emails and content and then the activity from visitors specifically, so let’s look at two of those three things, but let’s start with content here, and then we’ll spend a little time talking about the future, what our plans our next, and then take some Q and A.

Chris Butler: Okay.

Dave Mello: Sounds great.

Chris Butler: The content side is somewhat rudimentary right now, because we want it to be the foundation for some of the where is this going next stuff, and in particular, automated intelligence to coach you through this, so what it’s looking at, is, it’s looking at what you publish and when, and that’s helpful to know, because all of our clients, they have a content strategist they work with and they have their content matrices that tell them when they’re doing things.

This is just another view to see what actually happened. It’s more of a post-analysis. It gives you a list of all that content, shows you what type of content it was, the status of it, the word count and it’s publish date, which is really nice. If someone is at the marketing director level and they are not necessarily in the weeds, this is a really helpful way for them to stay in touch on what’s actually happened in their content department in the last whatever amount of time they want to look at.

Mark O’Brien: Excellent.

Dave Mello: In terms of where this is going, I don’t know.

Mark O’Brien: Yeah, let’s jump in.

Dave Mello: We’ve done a lot of scoring and low-level content analysis based upon emails. That same thing will be applied here, so the ability to know how successful an actual post was. How many people visited, did it really appeal to that actual target? All of that will fall right into this.

Chris Butler: When I had the initial idea, the idea of getting an email, three was three types of email I envisioned us being able to send to people using the system. The first one is that campaign specific analysis. You send the campaign, you get an echo immediately or, we decided to do the next morning, telling them how that campaign went. We want to give them time to actually have some return on the campaign. That can be tweaked, but we think 24 hours is the way to go.

The second type of email was a weekly evaluation of your site’s content, and not traffic. We’re not interesting in duplicating analytics. We’re interested in saying, “okay, you published these five pieces of content. How did they perform?” and that goes down to design analysis. We want to algorithm to include how effectively you laid that page out, how effectively the image support and text choices and the scanability of it contribute to engagement. How effectively the CTA choices, that kind of thing. That would happen on an automated basis. You’d get an email once a week.

The third one is probably the most exciting, and the most elusive, which is broad scale data analysis of light firms telling you how you should be evolving your marketing practices. In other words, it would be, what we’d love to hear is, hey, a firm like yours should be seeing this and should be doing this, and we have an enormous data repository now of a wide diversity of firms using this tool and all the data is being passed through it, and so we’re going to be writing some machine learning expressions to be able to say things like that, so a firm like yours that’s doing this kind of content marketing should be seeing this return in these ways quantified but also should be doing this kind of thing.

For instance, this firm might not be doing that much gated content, we want the machine to tell them yeah, your next best opportunity is that. Or your next best opportunity is eBooks, based upon your behavior or your next best opportunity is to do this kind of content in these types of topics, and we believe that we’re going to be able to do that in relatively short order based upon the data we’re already collecting, and that’s the future of this. Nobody has time to sit down and dream all this stuff up. In the marketing space you nee to go after this opportunity when its relevant and this machine is going to tell you how to do that.

Mark O’Brien: Yeah, that comparison thing is something I’m so excited about because it’s competitive. If they see that, okay, the average for the pool sending 3.2 emails per month and they’re sending .4 emails a month, it’s going to motivate them, and so being motivated by the crowd is something I think is going to help quite a bit.

Chris Butler: Especially if you know that that other firm who’s doing more than you is getting more than you. If that other firm is sending more emails that you are and actually seeing a better conversion rate or seeing a better opportunity score than you are, then, hell yeah, it’s not just about, I want to be as active as they are, I want to get what they’re getting.

Mark O’Brien: Right, and you won’t be able to see by firm name, but you’ll be able to see the trends.

Chris Butler: Correct. There is a privacy component here that’s sort of implicit. You know Mark was great with this stuff, he got permission from this client to show these names, but when we do that broad scale analysis, we’re going to be speaking in demographic generalities. We’re not going to be saying, “Here are the firms we’re comparing you to.” We’re going to be saying, “Firms of like nature,” based upon whatever characteristics are expressed in that particular communication.

Mark O’Brien: Okay, great. Let’s get into some Q and A here. If you do have any questions, just go into the questions poll and type them in and we’ll see them immediately. The first question is from Andy and Andy asks, are users only identified by name after they’ve filled out a conversion form?

Chris Butler: Yes.

Dave Mello: Sort of.

Chris Butler: I’ll let Dave answer.

Dave Mello: That is the primary mechanism that we will cookie a person, but we also can look at other sources to figure out who a person is. For instance, if they click through from an email campaign, we’re then able to poll the automation source to actually tell us who that person is.

Chris Butler: Right, they would not to have had to filled out a form in that session, per se. If we have cookied them in a session they started on their desktop and then they click on a form on their phone, we can dupe that experience instantly.

Dave Mello: Exactly, but if somebody does come to the website on a brand new computer, we won’t know who they are, but then to your point, if they do anything in the future that tells us who they are, we’ll be able to backfill all of that data, so to the best of our ability, we will know who is on the site and we’ll be able to actually be able to take that info and actually use it.

Chris Butler: Right. The root answer to Andy’s question is that, identification here comes down to the bedrock of voluntary identification. Someone needs to fill out a form and identify themselves. Before that, we are tracking them, but they have a number, and we don’t know anything about them until they tell us that stuff.

Mark O’Brien: Right. Nick has a question. Is this tool only compatible with Salesforce? It really would be helpful to talk about the necessary underlying technologies.

Dave Mello: Sure. We tried to build this to be technology agnostic as much as possible. Of course, that being said, we had to start somewhere, so we kind of built it around the tools that we use the most. That would be a set of actual plugins for WordPress as well as ActOn, and then the third piece is Salesforce. Of those three, the first two are pretty lightweight in terms of they can be replaced pretty easily. We can swap those out. Act-on and WordPress. Salesforce, I think, is here to say, and I don’t see us really working with any other CRMs in the short term.

Mark O’Brien: Right. I agree with that for sure. We’ve definitely tied our fate to Salesforce.

Dave Mello: It’s a great choice.

Mark O’Brien: The previous webinar was really all about that.

Chris Butler: What Dave’s really describing is this tool is essentially an API. It’s grabbing data from other places and messing with it. We’ve build a visual layer to show you what they looks like but it is connecting all those pieces. So in theory, we could have chosen a different CRM. There would be no good reason to do that at this point, so it’s about as agnostic as it ought to be, I think.

Mark O’Brien: Okay, great. Danixa has a question. She says, that she’s definitely a believer in analyzing, making sense of data, could we expand on the lead scoring system and do we have any examples of success stories from anyone that’s following the system?

We definitely have lots of examples of success stories from clients we’ve worked with. What we’re showing today is brand new. Newfangled is the only client to date, and I do have a lot of person success stories for business we’ve won, as part of our outreach based on what these reports told us, and so it’s been an amazing success. It’s surpassed my expectation for what I would get out of it actually.

That’s really exciting. What we’re also excited about is we’re going to be rolling this out to all of our clients who are on the WordPress platform starting in Q4 on a demo bases of data testing.

Chris Butler: I’m not sure if it’s understood for everyone on this webinar, but this is quite literally the introduction of this system to the world. Nobody’s seen it before, officially until today. We have some clients that have access to it and they don’t even know it, but Newfangled has been using it for several months now, and obviously we’re using it when it was something that didn’t even look like this before we finalized the design, but this is our way of introducing it to the world, and we’re excited about where it’s going.

Mark O’Brien: Mm-hmm (affirmative) wonderful. We’ll take one more question then we’ll break here. Which one do I choose? There’s a question, are there any social media tie=ins in the analytics portion of the insight dashboard?

Dave Mello: Yeah, quite a few. There are social media tie-ins in terms of pulling in social media stats about a particular person. We’re not analyzing social media activity. That’s a great idea, and I’ll probably have it build within the hour.

Chris Butler: That’s the thing. I think we’ve build a platform here for centralizing the kinds of things that we need to bring attention to and for some firms, social media is almost certainly one of those that they need to have a lens on and for us, it’s not been that big of a thing. We haven’t done a ton with social media in terms of real strategic outlay, but that’s a great idea and we should definitely bat it around.

Mark O’Brien: Thanks, Andy. Thanks for the question. Okay, we will get one more question in because it’s such a good one. Bob asks, on the difference between opportunity scoring and lead scoring, when Mark is checking the report every morning, is the system giving him one metric or several metrics as a signal to make an outreach decisions, or is he looking at specific metrics he’s chosen in making the decisions?

What is really is, the system is giving me a lot of data, so it’s giving that email every morning but it’s also anytime anyone converts on a form, I get a notification, obviously through email, and the main thing that develops inside of me is just a peripheral sense of this individual, because I’m being exposed to all the activity, the necessary activity on a regular basis, so the thing that has made the difference for us is that it’s showing us, take the form conversion, for example, that wasn’t nearly enough information for me to act on at all, but now that I’m seeing the most important people based on how we assess them coming up in my email every morning, and I’m seeing location and firm and page views and that kind of thing, the peripheral sense between that and the email activity, that encouraged me to click through and check out the information, and most is through my phone. We’re looking at desktop, but really, almost all of this I’m doing on my phone before I even get to work, and I’m checking out different sessions and things like that, and the phone view is actually even a little bit nicer because it orders things in a certain way.

Then, of course, once I get to that lead screen, then I know right away, and to be completely transparent, what I’ve looked at, I look at the rank of the individual who is on the site, what their title, I look at their location, and then I also look at their company information. The additional of the company info is a really, really, really big deal.

Chris Butler: I was going to mention that if there’s probably one single data point that represents the signal that Bob is asking about. You used to notice this all the time doing your own pattern recognition, like, oh, this form submission came through from this person at RP, I’ve seen a lot off RP. There’s other RP before, but it’s all intuitive. It’s all your own memory.

What this does is, as soon as he sees Bob’s little face in that pane, he scrolls down and he sees oh, RP has 15 different people on the site right now, that’s a huge deal. That is a major, meaningful difference between the firm that might just be a firm of one person or a firm of 300 people where only one person has ever visited the site.

Dave Mello: They’re actually talking about us internally and they’ve told other people.

Mark O’Brien: And is that one individual looking for a job or is this individual the owner? It could be the one person, but is it the right person? It all depends.  I hope that helps to explain that, Bob. This was fun, guys.

Dave Mello: Yeah, it was awesome.

Chris Butler: Thanks a lot for putting us all together.

Mark O’Brien: On a large scale and a medium scale.

Chris Butler: I think it’s been one of the more interesting projects we’ve worked on over the past few years internally anyway, and it’s been a real pleasure to be able to do something like this, especially given that the more we continue to nurture it, the more there doesn’t seem to be another tool out there that’s beating us to the punch. That’s sort of remarkable. It’s something I said to Dave that morning after we looked at my white board is, if this marketing automation suite that we know of isn’t building this right now, something’s wrong. Someone else should be jumping on this synthesis and proactive analysis tool and there are things like that out there, it’ snot like this is complete invention, but I haven’t seen anything of like kind. I haven’t seen anything that really competes with this just yet.

Mark O’Brien: Yep. The thought process then, as always, the future is the most exiting part of it. All right, hope you all have a great day, and thanks a lot for attending, and we’ll talk to you next webinar, which will be in December on some specific content marketing techniques. Until thin, I hope you all do well.

Chris Butler: Bye.

Dave Mello: Take care.

Chris Butler: Thank you.