Episode 209

podcast photo thumbnail
1x
-15
+60

00:00

00:00

[easy-social-share buttons=”facebook,twitter,google,linkedin,mail” counters=1 counter_pos=”topm” total_counter_pos=”leftbig” style=”icon_hover”]

Our agencies and the work we do will inevitably be impacted by artificial intelligence in the next 3-5 years. That’s just a fact. Odds are, it’s probably already changing the way you do business every day. There’s a lot of hype and buzz about jobs going away and most agencies are struggling to figure out how to keep up.

Questions of what data to use, how to analyze it, what tasks can be automated, and what AI can do for clients pop up all of the time. And it all shifts on what seems like a daily basis.

Fortunately, the agency space has a beacon to light the way into the uncharted territory of AI. His name is Chris Penn, and he is an authority on analytics, digital marketing, and marketing technology. A recognized thought leader, best-selling author, and keynote speaker, he has shaped four key fields in the marketing industry: Google Analytics adoption, data-driven marketing and PR, modern email marketing, and artificial intelligence/machine learning in marketing.

Chris is a generous guy who frequently shares his knowledge over his own podcast, Marketing Over Coffee, as well as through books like “AI for Marketers: A Primer and Introduction.” I was thrilled to welcome Chris as my guest for episode #208 of the Build a Better Agency podcast.

I barely scratched the surface of questions I had but we got a good start, talking about data analysis, keeping up with the flux of AI, and the tasks that agencies can automate to save time and money. The conversation was fascinating and I’m excited to share it with you.

A big thank you to our podcast’s presenting sponsor, White Label IQ. They’re an amazing resource for agencies who want to outsource their design, dev or PPC work at wholesale prices. Check out their special offer (10 free hours!) for podcast listeners here: https://www.whitelabeliq.com/ami/

What You Will Learn in this Episode:

  • What data actually looks like in the agency environment
  • The importance of hiring people with analytics and software skills to an agency
  • How to keep up with the ever-changing field of AI in an agency setting
  • Common mistakes agencies make around data and analytics
  • The most onerous tasks that can be automated using AI
  • The capabilities of machine learning and where humans come in during the process
  • What Chris does to keep current with technology in the agency space

The Golden Nuggets:

“The keyword here is repetitive. What are the things that are repetitive that you could make go faster or that you could deploy better technology for?” - @cspenn Share on X “The number one way I keep current with AI is by trying to solve problems for either the company or for clients.” - @cspenn Share on X “When agencies think about AI and data, they're always thinking about it through the lens of, “How can I help clients with this?” And they're probably not turning it around and saying, “Oh wait, I could use these tools for me.” - @cspenn Share on X “No amount of data analysis and machine learning can dig around in somebody's head and explain to you why they did something. You have to ask them.” - @cspenn Share on X “If you are truly net new, creating something from scratch from whole cloth, that is your value add as a human.” - @cspenn Share on X

Subscribe to Build A Better Agency!

Itunes LogoStitcher button

Ways to Contact Chris Penn:

Speaker 1:

If you’re going to take the risk of running an agency, shouldn’t you get the benefits too? Welcome to Agency Management Institutes, Build a Better Agency Podcast presented by White Label IQ. Tune in every week for insights on how small to midsize agencies are surviving and thriving in today’s market. We’ll show you how to make more money and keep more of what you make. We want to help you build an agency that is sustainable, scalable, and if you want down the road, sellable. With 25 plus years of experience as both an agency owner and agency consultant, please welcome your host, Drew McLellan.

Drew McLellan:

Hey, everybody Drew McLellan here with another episode of Build a Better Agency. We are going to dig into technology today. We’re going to talk about math, and all kinds of things that on the surface you think you don’t want to talk about, but I think actually you’re going to find it fascinating. Before I tell you a little bit about our guests and we jump into the conversation, just a few reminders. First and foremost, if you have not already grabbed a ticket for the Build a Better Agency Summit, May 19th and 20th of 2020. I know it seems very far away, but honestly it really is not. And the tickets are going fast, which is good news for me. I want you to be there. So remember we only have room for about 220 agency owners or leaders. And at that point we have maxed out the space.

So if you think you want to attend, please go over to the Agency Management Institute website< and click on the NAB where it says BABA summit for Build a Better Agency Summit. And you can read all about the speakers and get registered. Another thing I want to remind you about is remember that every week, we’re giving away something from either a podcast guests, or a lot of them write books or have courses or we’re giving away something from AMI. Like a seat in one of our workshops. And so all you have to do to be eligible to win one of those is go over to agencymanagemeninstitute.com/podcastgiveaway. All we need is your name and email address, and you are in the drawing forever and ever.

So even if you’ve already won, you are still in the drawing and you might win again. So take advantage of that because we really do give away some pretty cool stuff, and we are giving away some workshops in the coming weeks. So now would be a great time to head over there to do it. Also, I just want to put on your radar screen some of the conferences or workshops that we have coming up. Money Matters, where we spend two days talking about financial metrics, why you aren’t making money even though you’re busy, billing practices, how to know when you actually can afford to hire another person, tax strategies, all that kind of stuff for two days.

We’re going to do that on October 16th and 17th in Orlando, Florida. And then we’ve got two workshops in January that I want to make sure you know about. First one is building and nurturing your sales funnels. So that’s going to be all about how do you get someone’s attention and then keep their attention for as long as it takes for them to be ready to buy. As and you’ve heard me say it could be a day or it could be a decade. What are you doing to make sure that you stay relevant and interesting to them for that entire time so that on the day they’re ready to buy, you are top of mind. That’s going to be January 23rd and 24th.

That’s a Thursday, Friday, and then Monday, Tuesday of the week right after, Mercer Island Group is coming back and they’re doing a brand new workshop, brand new content. They have been studying the buyer’s journey that a prospect goes through, and the whole workshop is going to be, what are the milestones in that buyer’s journey and how can we win their attention and their affection and their interest at each of those stages? So it’s going to be fascinating and I’m super excited about it. That’s January 27th and 28th. So you could come to the workshop Thursday and Friday, play at Disney World Saturday and Sunday, and then do the workshop Monday and Tuesday. That doesn’t sound awesome, does it? I think it does.

So join us if you can for Money Matters, build a nurture your sales funnel or the prospect’s buying journey. We would love to have you there for any of those or all of those, and you can register for all of them on the Agency Management Institute website. All right, let me tell you a little bit about our guest. Although I suspect most of you know Chris Penn. So Chris Penn is probably one of the leading authorities on analytics, digital marketing, and marketing technology. He is a very respected thought leader. He’s a best-selling author, and keynote speaker. I just recently saw him speak, and he was just brilliant at the recent MAICON conference in Cleveland, but he is known as being a person who has literally shaped four key areas in the marketing industry. Google Analytics, adoption, and usage, data-driven marketing and PR, modern email marketing and AI and machine learning in marketing.

So he has this depth of knowledge around how we can take advantage of some of the technology that’s out there to actually improve the results we deliver for clients, and our own bottom line. I’m super excited to have him on the show. You also probably know him from he founded Pod Camp Conference. He is the co-host of a great podcast called Marketing Over Coffee. So I’m guessing that you have bumped into Chris at some point in time, read one of his books. I highly recommend AI For Marketers: a Primer and Introduction, and all of his other books like The Marketing Belt books are spectacular.

So I can’t say enough good things about him other than I know that I’m going to run out of time. There’s so many questions I want to ask him on my behalf and for you that I want to jump right into the conversation. So let’s do that. So without any further ado, Chris, welcome to the podcast. Thanks so much for-

Chris Penn:

Thank you for having me.

Drew McLellan:

So I’m guessing most of my listeners are very familiar with your work and are probably already fans, but just for the few that are not familiar, because you have an interesting background. And you are very much in alignment with many of the listeners in terms of having an agency life. Give everybody a little background on sort of how you came to be in the position you’re in today.

Chris Penn:

Sure. My background is originally in IT. I started as IT director at a bunch of companies. Early two thousands, went to a FinTech startup, and that’s when marketing became marketing technology. And as an IT person, I moved into marketing technology. Then end of that decade. I started really getting heavy into analytics. Analytics became data science became machine learning, became AI. Around 2012, I went to an agency, a PR firm for five years and change building a marketing technology practice in there. Then towards the end of that, I realized that that agency was going one way and I wanted to really focus in on AI and machine learning. And my account director split off and we became our own company. We founded Trust Insights, which is a data science consulting firm for marketers. And so that’s how we ended up here today is focusing on. We’re building an agency essentially for data science for marketers.

Drew McLellan:

I’m assuming that a lot of your clients, or at least some of your clients today are agencies, yes?

Chris Penn:

Yes, that’s right. We have brands and agencies. Pretty much anybody who has data and wants to make more money with it.

Drew McLellan:

Okay. So let’s just start there. We talk about data and big data. I don’t know if I’ve ever heard anyone talk about little data, but we use that word like everyone knows what it means. So for the non-techies listening, define that for us in the framework of our agencies and our clients?

Chris Penn:

Data is information. That’s the easiest way to frame it, right? And there’s all these different kinds. You mentioned big data and the running joke in the data science community is big data is that anything that doesn’t fit in an Excel spreadsheet. So if it’s in Excel it’s data, if it’s bigger than Excel it’s big data. And that’s actually not a bad way to look at it, because when you look at the way agencies use data right now, for the most part, the most sophisticated they get is typically an Excel spreadsheet. In that comes someone’s terrible pie chart in a PowerPoint.

So, companies have data and you’re surrounded by it. Every time you have a client call, you have data. If you are recording with permission, a client calls, you can take that audio, load it to a an automated transcription service, turn it into text, and then you can analyze, for example, your client calls. If you had say a year’s worth of client calls from all your different clients, and at the end of the year, you had a list of the clients that you’re retaining, and the clients that turned over. You could then run an analysis for example, and say, “Okay, what do all the calls, if you analyze the content of those calls, what are the calls have in common for clients we kept? What do they have in common for clients we didn’t keep?” And you have all this other data.

Who’s on the team? Is there an account manager? There’s certain account managers who keep clients like crazy or their account managers where there’s a lot of churn. You’re looking for patterns in your data to help you make better decisions. To say, “Yes, this team, and this topic is a winning combination for client retention, client upsells, things like that.” This combination over here, we’ve got to change some of them because clearly there’s a lot of churn, lower margins, unhappy team. That’s the way agency owners, and executives should be thinking about data, is not only the clients stand and what to do with it to help the clients be more successful. But what data are you analyzing internally?

One of the things I thought was most broken about the agency world when I was in a more traditional agency, was that so much is done on gut instinct, or feelings or the phrase I hate most of all, “This is the way we’ve always done it.”

Drew McLellan:

Right. Of course.

Chris Penn:

And not based on information you have that you could be using to make better decisions. I’m hoping even just the thought of, “Well, gosh, if we even just analyzed almost every team in the PR firm I worked in had like the account coordinator or the intern taking notes on every call.” Does anyone analyze those notes over time to see, “Hey, these are the things that keep coming up on client calls over and over again.” Maybe there’s even a new practice here that we could make some extra money on.

Drew McLellan:

Right. Yeah. I’m guessing you’re right. I’m guessing when agencies think about AI and data, they’re always thinking about it through the lens of how can I help clients with this. They’re probably not turning it around and saying, “Oh, wait, you can use these tools for me.”

Chris Penn:

What’s the joke that we always talk about with clients, with our agencies. It is the cobblers [crosstalk 00:11:35].

Drew McLellan:

Oh, I hate that phrase. Hate it, hate it, hate it.

Chris Penn:

But it’s a hundred percent accurate.

Drew McLellan:

Absolutely.

Chris Penn:

Exactly. So from a data perspective, what are you doing to grow your business?

Drew McLellan:

And for the agencies out there who are most of them, by the way, who are still scratching their heads saying, “I know I need to learn this, but I don’t even know where to start.” It does make sense that they would start experimenting on themselves, right?

Chris Penn:

Exactly. All of the innovations that we’ve come up with for our clients, originally the problems that we had. So I have this thing called the digital customer journey map. We had come up with some fancy name for it. But it’s just a chart. It’s a chart that shows you what digital channels contribute most to conversions, and in what order, because I needed to know that. I needed to know that for our company. Like how important is organic search? How important is social media? How important is speaking? And if I didn’t know that I couldn’t make decisions, but by charting it out with some custom built software, now I can go, “Oh, I should be doing more of this, and I shouldn’t bother with Instagram because it’s not working here.”

Drew McLellan:

How did you gather the… Like what were the data points that you use to build that out?

Chris Penn:

The first and most important one, and the one that every agency owner and operators should be doing, is what business problem do I have? In my case, the business question I had was what’s working. We’re a small startup, we’re three people. We have combined 120 hours a week, right? How do we allocate those? Figure if only 25% of that’s going to be allocated to growing the agency. The other 75% goes to clients. So, what do we do with that 25%? Well, that was our question. And so then the next step after you do the business requirements, the problem you’re trying to solve is what’s the analytic approach? How are we going to try and model this? And then you get to data requirements.

What do you have access to? For a lot of agencies they are fortunate in that they pay for a lot of tools on behalf of clients that they could be using for themselves. There’s a ton that’s free. So Google Analytics, totally free. If your agency is not using web analytics of some kind, Google Analytics is the gold standard. You’re doing it wrong because it’s huge amount of insight. You have email, you have social media data, you have news data. You have video data, you have all the stuff that your doing or should be doing. That’s your starting point. Then you explore it, you model it, and you make decisions on it.

Drew McLellan:

So the modeling part, you have a technology background. You are a data scientist. So for you, this is like me saying, “Hey, Chris, how do you make a bowl of cereal?” And you’re saying, “Well, Drew, you put the cereal in the bowl, you pour the milk over it and you’re done. Move on.” But for a lot of people who came up through the agency ranks in their account people or their creatives, they went into the agency business to avoid math and data, right? Now the realization on all of us is, “Oh, crap, I have to learn this stuff.” So when you talk about the modeling, are there resources or places that we could go to look at examples and go, “Oh, I can see how I could plug my data into this example and wallah, I’m going to get some answers.”

Chris Penn:

It depends on how you define the word model. In the context I’m using it, I’m using it in terms of machine learning. And there is a way to take a pre-trained model and tune it, but that’s pretty far down a technical rabbit hole. The number one thing that as an agency owner, especially if you don’t have a quantitative background should be doing is all your new hires should have a balance of creative and the quantitative. Every new hire or at least a substantial number should be able to do basic statistics. If someone doesn’t know how to use R or Python or SPSS or SAS or Tableau might not be as good a hire. Once you have that talent on staff, then you can start asking for their help to say like, “Hey, here’s a question I want to answer too, what impact does Instagram have on my agency? What impact does LinkedIn have on my agency?” That person should be able to help you with, or if you can’t have the head count, then you find a partner, a vendor to help you do it with.

There aren’t a ton of agency specific pre-made models, because everyone’s is going to be different. Even your own models change over time as you do different things and you just try different things. But there are clearly defined best practices and software that can make the task easier. I would compare the software, the solutions out there to kitchen appliances. A Kitchen aid stand mixer is going to make whipping eggs, and kneading dough a heck of a lot easier than doing it by hand, but you still need to know how to cook. If you pour dough in there, your not going to have a good time.

Drew McLellan:

And this is part of what you do with your clients, right? So there was a case study on your website about a tracking recruiting firm. Because I think that was a great example of like, “Here’s part of the problem. I think when people hear machine learning and AI, agencies are wrestling with, how do I apply that to actual things we do for clients?” I thought that, that case study that very brief case study you got on your website, was a great example of like, “Oh, here’s how we thought through this problems.” Can you just walk us through that?

Chris Penn:

So this is one, I assume the one we’re talking is the recruiting one. This agency, it’s a truck driving agency actually. It was recruiting drivers, and one of the things that they said is we don’t know why are our applications are down? We’re not getting as much conversion as we thought. What we did was we took all the job postings that they were recruiting for, and analyze the words and phrases and language that they used in those job postings. Then we looked at the calls from their call center. No, 17,000 audio recordings. We had machine learning transcribe those calls to written texts that could be analyzed the same techniques that we did for the job listings.

Then we put the two side by side and said, “Look at the words and phrases that are in your job listings.” CDL, class this, class license, this decoration, this whatever. And then look at the conversations that the drivers are having with the recruiters. Starting pay, pay per mile, weekends, vacation, what type of loads are you carrying? What drivers were saying, and what was in the job ads, weren’t even dining at the same restaurant. So of course, if you’re a driver, and because they’re in short supply, they control the market, because there’s a talent shortage.

They can look at your job ends go, there’s nothing in it for me here. I see nothing in this that interests me. This is all about you, and only you, not me. So, bye guys, I’m going to go talk to a company that is talking about the things that I as a driver want to talk about, what is my starting pay? Will I be home on holidays and things like that. And if you’re an agency owner, look. You don’t even need texts and analysis for the most part unless you’re like a mega agency. Look at what people say in your Glassdoor comments. Look what they say about working at your agency, especially if they left or if they left and they weren’t terminated for cause. What do people say? And does it reflect the way that you talk to your current employees about what it’s like to work there?

One of the things that at the company I used to work for, Glassdoor employee reviews were always saying things like crazy workload, no room for advancement, political infighting. Then when management would talk to employees, everybody loves each other. There’s this massive disconnect that people are like, management’s out of touch. They don’t know what’s going on. I’m going to go work someplace else. So if you’re an agency owner, you should be doing this. This is not AI. This is using you in intelligence.

Drew McLellan:

Right. So if I’m an agency owner and we haven’t dabbled in any of those. Maybe we have Google Analytics plugged into our website or all of that, where do I start to educate myself and what are some sources of resources that will help me sort of figure out the baby steps that I need to take, because a few weeks ago, maybe it was a month ago that the newsletter that you send out on Sundays, which by the way, everybody should subscribe to. It’s awesome. But you were talking about your greatest fear. I think you’d set it up as when you have a nightmare, it’s about this, right? But your greatest fear is not being able to keep up. When I read it, I thought, “Holy crap, whose Penn can’t keep up? What hope do I have to keep up?” Because I don’t have a data background.

I’m like all of these listeners that I’m a copywriter by trade. So math is not where I spent my time in college, or my early part of my career. So I know that every agency owner out there is feeling this. The rate of change is faster and faster and faster, and how in the world do I keep up? I know that you have said in the past that for agency leadership, the greatest risk in terms of being successful down the road is not keeping up. So how do we keep up? How do you keep up?

Chris Penn:

Well, let’s table me because I’m an anomaly.

Drew McLellan:

Fair enough.

Chris Penn:

For agency owners, and for any business, there’s sort of a seven step process in your journey towards AI. It’s not something you could do immediately. Just like you wouldn’t immediately become executive chef at a Gordon Ramsey restaurant, right? There’s a long process to get there to build up those skills. And like becoming an executive chef, your first step is take a cooking class and see if you can win the basics. Can you chop a carrot without injuring yourself, or anybody else? That process in this realm is first becoming data competent. Do you have data? Do you know where it is? Do you know what data you have access to? Then becomes the next step of learning how to do analysis. Learning how to derive usable, like what happened from your data?

That can be very much the realm of spreadsheets and things. What I see happen over and over at agencies is what my colleague Avinash Kaushik calls data puking. Where as an agency, like you dump a 48 slide PowerPoint on a client saying here’s all your data. The client’s like, “What is this? Is this chart?”

Drew McLellan:

What does it mean? What does it tell me? What is it exactly to do? What does it tell me to stop doing?

Chris Penn:

Exactly. And so what happened? You’re becoming data-driven and developing analytics capabilities. This is essential. And there are tremendous resources for learning how to do that very simply. For Google Analytics, there’s the analytics academy. Go to analyticsacademy.withgoogle.com. It’s totally free. You can learn some of those practices. If you want to learn data analysis in general, there are totally free courses from IBM. Go to cognitiveclass.ai and take a bunch of courses there. They’re all free. Like, $0. And every agency owner, even if you don’t want to do this, you should absolutely be your staff through some of these trainings, because again, $0 and they’re valuable skills.

Drew McLellan:

Agency owners should do it too. If we don’t know what we’re asking our people to understand, and we can’t have an intelligent conversation with them around it. Then we become the ones that are irrelevant.

Chris Penn:

Yes, there is that. And then after analysis of what happened, comes qualitative research. Being able to answer why. Like, when you talk to a client, what is your offboarding procedure look like? This is something that you should be asking as an agency owner. What is your exit interview for a client look like? What questions you ask them? Why are you leaving? Why did you not renew your attainment? Digging in deep down as to like, “Damn, these are the reasons that we didn’t do more business with you.” Because this is something that a lot of people particularly in AI, don’t get. No amount of data analysis and machine learning can dig around into somebody’s head and explain to you why they did something. You have to ask them. So that’s step three.

Once you know what happened, once you know why it happened, then you could start to do things like prediction and forecasting. That’s where you start getting into stuff like data science, where you are like any form of science, hypothesis testing. Like, “Oh, well, we think clients leave because of these reasons.” Great. Now you’ve got a hypothesis. Now, you test it and you run tests and optimizations and things like any scientist would to prove or disprove your hypothesis. Once you’ve gained competence in data science, then you move on to things like machine learning and AI. But it really all begins at that very basic level. What do I have? What questions do I have? What happened? Why? What should I do about it? Those are skills that don’t require technical background. They require you to ask great questions.

Drew McLellan:

Yeah. Which in theory is what agency folks are supposed to be great at? Yeah.

Chris Penn:

In theory.

Drew McLellan:

Yeah. Right. So how does all of that… actually, I’m going to put AI aside for a second, and go back to what you said before. Which is A, we should all be using analytics and we should all be using it properly. But I know in your work that you have discovered that a lot of folks aren’t using it very well or at all. So what mistakes we making around analytics? Specifically around the data that we all think we have besides the data puking, which I think many agencies are guilty of. What other mistakes are we making around the simple data that we do have access to?

Chris Penn:

It comes down to something from 1964. HJ love it. It was a business professor, came up with this framework, which has been used and used or defined in the decades. Since then people process platform, people process technology. What happens in the agency world especially is everyone’s so focused on the technology or the platform. Google Analytics, machine learning, data science, R, Python, you name it, but we ignore the people and the process. And that’s where agencies have gone the most wrong. Yes, you have Google Analytics. Good, you’ve got the platform down, but what’s your process for using it? To document it? Do you have governance around it? Do you understand what’s in it? Do you even know if you’re violating regulations? I was in a client’s analytics last week. They had personal identifiable information in Google Analytics, which is forbidden.

Google will kick you off the platform if they discover that. They had one form on their website that was passing email addresses straight into Google Analytics, which is a no-no. In addition, anyone who has access to their analytics could now take their entire customer database and run away with it, which is a security risk in many other ways. So there’s a process breakdown there. And then the people side is where agencies have gone the most wrong. They’re not hiring the right people. They’re not investing in the people they already have. And there’s sort of this legacy of just get it done as fast as possible, maximize the margins, maximize productivity and not focus on what I call system thinking. When I looked back at the way, in this case, it was a PR agency.

PR teams solve problems. They tried to solve the problem as it was instead of taking a step back and saying, “Can we build a system, a set of processes, a methodology for solving this problem in perpetuity.” I’ll give you a very concrete example. Social media content curation. We generally say in social media marketing 60, 70, 80% of what you share should be industry news, things that were relevant to your followers, and 20% should be promotional. You don’t want to be like all me all the time. That takes time. At this old company, I used to watch the account coordinators spend four to six hours a day curating content for clients. What a waste of time, and what a waste of a human being.

What we did at Trust Insights because we don’t have that as a startup. As we automated, we took a step back and said, “Is this a problem, a problem that technology can solve? Is it repetitive enough? Is it structured enough that we could automate at least part of it?” And the answer is yes. So we built software that ingests the feeds from like 1500 blogs, analyzes them by topic, analyzes them by how many times has been shared and clicked on so far, and then selects the best of the best. Spits out a file and says, “Okay, go load this into buffer or the software of your choice.”

So what used to be a four to six hour task, it is now a four to six minute task, and frees us up to do other things with that time. As an agency owner, you should be thinking to yourself, “Okay, what tasks do I have within my agency that are repetitive, that are huge time sinks that I should tackle first and try and solve from that [inaudible 00:29:56], solve it forever with a system that’s a blend of people in technology.” Agencies don’t think that way, and that is what’s holding you back the most.

Drew McLellan:

So I’m sure that when you said the words, and so we built software to analyze blah, blah, blah. Most people went, “I’m out. We don’t know how to do that.” So if someone doesn’t have folks like you on staff, how do they either vet tools out there to do these things or find a partner? So for example, if I wanted to replicate that very thing, because every agency is like, “Oh my god, we do that for five clients and we’re doing it for ourselves. And we’re doing a lousy job for ourselves.” Because as you said earlier, cobbler’s children have no shoes. We don’t have time to create our own content blah, blah, blah. I want the thing that Chris just talked about.

How do we who can’t build it ourselves, or am I thinking that you’re talking about building a skyscraper, and really you’re talking about putting some Legos together? Can we do it ourselves? And if not, how do we enjoy the benefits of a tool like that when we can’t build it?

Chris Penn:

Fundamentally, if you’re talking buy or build, right? You buy when you have money, and you don’t have time. You build when you have time, and you don’t have money assuming that the skill exists. The first thing you should do is figure out, do you have the skill anywhere on staff? And if you don’t know the answer to that, one of the first things you should absolutely do is these skills audit of your staff. You may find that the intern that on team seven actually has two years of programming experience that they got into college, and they just decided they didn’t want to do that full-time, but they may have the skill. Unless you have a deep skills audit of your people, you won’t know that that’s potentially an untapped resource within your own walls.

If you just don’t have it, period. If you know hey, we’ve only gotten students from art school, and not one of the technology arts schools. Then you start looking at a couple of different approaches. If you can clearly define and outline the business requirements, and put together like a mock flowchart of here’s exactly what I want to have happen. No technology, just whiteboard it out, you can get a contractor to build software for you. There are contractors planet wide, some of whom are incredibly affordable. Looking at like Croatia, Ukraine, Singapore, all these different places that as long as you have a very clear definition of what you want, you can hire someone to build a for you. Or you can partner with companies like mine to just have it as service software. As a service which on every week, a file drops in your inbox. And you pay money for somebody else’s software.

But there are many different pathways to that solution. So if it sounds like that’s something we need to do, then pick a solution based on the time, and the money you have available. But I would start first at doing that skills audit, because if you don’t know what you have in-house, it’s a wasted opportunity. Especially since that person who has maybe a year or two of those skills would probably be delighted to dust off say, “Wow, I never thought I could do this at an ad agency, or a PR agency or whatever.”

Drew McLellan:

Well, and it seems to me that if you don’t have anyone on your team that has these skills, or can at least talk about sort of all of this from sort of a programmers or technology point of view, this is not going away. This is not-

Chris Penn:

It’s not a fad.

Drew McLellan:

Right. It’s not the Foursquare of agency life. It’s going to be around. So if you don’t have some, but any one human being on your team that can talk this talk and understand it when other people are talking to them, you really put yourself at risk, and at a disadvantage of making bad purchasing choices, and sort of being duped as you start to outsource this stuff, right?

Chris Penn:

Very much so. There’s a term in the industry called AI washing where every company is saying, “Oh, we use artificial intelligence.” And the Financial Times did a blistering expos a recently that said 35% of companies that say they use AI in fact have zero in their products and services. So yeah, becoming educated, developing a detector of what our Spanish and Mexican friends would call [foreign language 00:34:19]. It’s absolutely essential, and being able to ask good questions and understand the answers. And again, you don’t have to have like fingers on keyboard writing this code to do that.

You can take courses in the management of AI to understand like, what should this thing be able to do for us? What are its limitations? The same is true of data science and analytics. Again, I would point people towards the free classes at cognitiveclass.ai. There’s a couple online big data basics and data analysis basics. No code, but learn the lingo, learn the kinds of questions you should be asking.

Drew McLellan:

Yeah. I want to ask you more about sort of our education process, but first let’s take a quick break and then we’ll come right back. I wanted to take just a quick second and remind you about one of the core offerings of Agency Management Institute. And that is our peer networks. So we offer them both for agency owners, and also what we call key executives. So if you’re a traction follower, these would be for your integrators. These are sort of your right hand people who help you run the business day in and day out. So from the owner’s perspective, imagine a Vistage Group or an EO Group, only everyone around the table owns an agency.

These folks become like your board of advisors. They become trusted friends that you learn a lot about their business, and they learn a lot about yours. So not only do you learn from us the facilitators, but you’re constantly learning from your peer group as well. And the same thing happens in the key executive groups. We bring them together and we help them learn how to help you bring your vision to life as an agency owner. If you want to check out either of these peer groups, you can go over to the AMI website and look under the networks tab. And there, you will find information on both our live, and our virtual agency owner peer groups, and also our key executive group. Check it out, and if you’re interested, let us know. We’re happy to have a conversation. Okay, let’s get back to the episode.

All right. We are back with Chris Penn, and we are talking about AI and technology and data and all of those things in terms of how do we as agency folks wrap our arms around this. And how do we start using it to benefit ourselves and our clients? So we were both just recently at the MAICON Conference, which I thought was spectacular. But one of the things that you talked about in your presentation was what you started to talk about before the break. Which were sort of these sort of seven layers of how to sort of wrap your head and business practice around AI. I was telling you before we hit the record button, that I was a little surprised that it’s not until the middle stages when we’re actually using technology to automate activity.

And so I think that for a lot of agencies, that’s immediately where their brain goes is I should be using AI to automate things. So to your point earlier, what are some things like sourcing curated content that are taking an inordinate amount of human time? And I think analytics analysis is one of those as well. Agencies are beginning, and when I say beginning, I don’t know about you, but I I’m talking just scratching the surface of starting to play with this sort of stuff. So in your work, what are some of the problems that you’re seeing agencies, or your brand clients solve that you feel are sort of at the rudimentary? Like these are some good starting places to think about using AI to be more efficient and effective. Because it’s not just a faster, it’s also better, right?

Chris Penn:

Better, faster, cheaper. That’s what every agency wants, or every client wants. Better, faster, cheaper. So the question that as an agency owner, you have to ask yourself is what is currently terrible, slow and expensive, and repetitive. The key word here is repetitive. What are the things that are repetitive that you could make go faster or that you could deploy better technology for? Here’s a real simple one. Every agency I’ve ever been at has at least one human at every meeting taking notes. Why? That is a repetitive process that adds very little value compared to the power of having a smartphone laying on the table with a transcription app running in real time that transcribes what people are saying and makes that available with timestamps to the people after the meeting. And you get the audio, you get the words, you don’t need the intern to transcribe stuff.

If there are takeaways, that’s probably something that the manager, whoever should be using. Anyway, if you use Google Hangouts, there’s automated closed captions that appear across the bottom of the screen. So you can actually see it being written out in real time. The app that I use that I recommend that anyone use and you have to let your people on the phone know this is done. It has to be secured permission, but it’s called otter.ai. Otter, like the animal. O-T-T-E-R.ai. It’s like six bucks a month for a hundred hours of transcription time. That’s every conference call you’re going to have with every client that month. Transcribed, delivered to you. You can tag the different speakers who speaking for what, and it’s such a time saver. Because now, imagine you’re an agency with eight teams, and those eight teams each have an intern, and the intern spending an hour a day in client calls transcribing stuff, right?

You’re now talking freeing up almost a full work week of someone’s time. Not doing that job anymore. Find a different job to do. It feel like, “Well, I don’t know what else to have the intern do.” Then don’t have the intern. Save that time, save that money. That’s how agency owners should be thinking about this. What are the repetitive tasks that are low, or no value that we can just find a technology solution for that will free up time for us to then pursue the higher value stuff like machine learning and data science and deploying these things. Another real simple example is web analytics reporting. There a number of agencies that crank out hideous PowerPoints every week for every client, right? There is this free service called Google Data Studio. You build the dashboard once, it pulls the Google Analytics in real time, you give that URL to your client and say, “Here’s your report that is up to the minute every time you open the webpage.” You no longer need the 48 slide PowerPoint.

What’s really fun is you actually put Google Analytics on the Data Studio dashboard, and you see how many times the client checks it, and you realize, “Hey, they checked it like once a quarter.” But it then shifts the burden to the client to say like, “Yeah, you’re getting up to the minute data.” And because it’s Data Studio, you manually select the pieces that you want. It’s only the data that you care about that delivers the deal that reports on the impact that you’re asking for. Now, the catch is, and the reason why agencies don’t do this is because A, it requires some good infrastructure, and good forethought and planning to collect that data in the first place. Second, sometimes it says that you as an agency, ain’t doing such a hot job.

Drew McLellan:

Right. But in theory, if we’re manually preparing that report, we still have to own up to that sooner or later, right? In theory?

Chris Penn:

Indeed. Ethical, well worth agency, yes.

Drew McLellan:

Yes. Right. That’s all of my listeners are.

Chris Penn:

All your listeners are, but boy, I can share some war stories from other places where that is not the case.

Drew McLellan:

Yeah. But having something pull the data is different than having the data analyzed, and offering up suggestions for here’s where I think the puck is going go there. So in most cases, is AI crunching data, and then are they then handing it off to the humans because I’m reading more and more about like Chase coming out and saying, “You know what? The machines are a better copy than our people did.” So, where does that handoff happen?

Chris Penn:

It depends on the skill and sophistication of the system. For most agencies today, the machines will crunch the data, and then our value add is the humans is to do that interpretation to say like, “Oh, okay, we have a gradient descent model here with these 17 predictors, and these top three predictors of the strongest. So Mr. Client, you need to double the posting that you do on Instagram, and you need four more pictures of cats a week.” It is our experience as humans to interpret the analysis and transform that into insights based on our wisdom and experience. For the example that you’re talking about, the example you’re talking about is what’s called natural language generation. Where we train the machine to create stuff based on an enormous training library. That’s fairly advanced applications of machine learning.

It is within reach, but you can actually technically do it for free as long as you know what you’re doing with the code, and you have the training library to crank that out. I did an example recently taking a certain prominent politicians tweets and feeding it to this writing software. And it mirrored the tone exactly. The half broken grammar, the weirdness spellings of capitalizations. But I told it, I said, “Software, I want you to declare war on North Korea.” It came up with all these things like, “These are all very bad people. Those Asians.” I should disclose to those listeners listening, I am Asian. I am Korean by descent. So it’s fine. But generating net new content is something that AI is very good at, especially because, and this is no surprise to everyone in the agency world, a lot of what happens in agency life that we call creative isn’t all that creative, because the clients don’t want it to be.

The client’s like make me an ad that is a square box with these words on it and make 40 different variations of it. It’s not really creative. It’s a template. The rule is this, if you do it with a template today, a machine does it without you tomorrow. You do it with a template today, a machine does it without you tomorrow because a template is a trainable thing that a machine can learn to do. If you are truly net new creating something from scratch, from whole cloth, that is your value add as human. If you’re like, I want to dream up this thing that’s never been seen before. Never been done before. There’s no pre-existing examples. That is true creative, and that’s very hard for machine to do. But make me have another PowerPoint slide, that’s pretty easy.

Drew McLellan:

Right. So if you were back in the agency space, give me 10 things your agency would be doing using AI, or some sort of technology that you think most agencies are doing by hand today. So if I say Chris, I own an agency, a traditional agency. I know that your agency is much more finite in its focus. What would you say, with your knowledge, because you have a knowledge that most of us don’t have. You go, “Oh, I know 10 things where I can make more money, be faster, more efficient, and cheaper bang, bang, bang, bang. ”

Chris Penn:

Yeah. Number one, call transcription. All calls to this agency will be recorded and transcribed by AI. No one will be taking notes. Boom, you’ve saved probably 20 hours of labor a week. That’s super low hanging fruit, and that should just go away. Ideally, the meetings should go with them too. That’s number one. Number two, content curation, super low hanging fruit. Super easy to do. No AI is even involved. Like you could actually do that with traditional programming. That is something that a lot of digital agencies do that they shouldn’t be doing.

Number three, lead scoring. At an agency right now so many companies use good software. HubSpot, Marketo, Pardot, you name it. But they haven’t ever touched the lead scoring feature. They’ve never actually done an analysis to figure out what really results in a qualified lead for our agency. You run that analysis. You build a system that updates the scoring quarterly, pushes it into the software, instantly better results. Instantly better. You can say, okay, our lead quality is going to go up because we’ve put together a model.

number four, content analysis for clients, to be able to say like, “We’re going to tell you what content is going to perform better, period.” And there’s any number of data inputs like SEO tools, social media tools, news gathering tools, video tools, but you can put together, and you can know like this is what is engaging content. And they can repurpose that for ads, for PR, for content, for SEO, you name it. That’s a no brainer. Number five, predictive analytics across the board. Anything that you have times you use data for, you should absolutely be forecasting. We do this with SEO. Here’s the list of your keywords. Here’s the last five years of data that we’re going to train the algorithm on, and predict week by week. Here’s what’s going to trend each week. So when to publish the right content for your agency, for your agency’s blog, for your agency’s Instagram account. This is what you should be doing, and when.

Number six, driver analysis. What things yield conversions. You have the data in Google Analytics. You know what channels? You should a hundred percent be running that analysis. We do it for ourselves every week. We run the analysis say, “Okay, what channels are contributing to the converged we care about to generate leads for our agency.” Number seven, most valuable content on your website. Guess what? Google Analytics collects this data already about what pages nudge somebody to conversion. The challenge is it’s not in the web interface. You have to extract it from the API.

If I had a traditional agency with a good agency blog, I would be running this software weekly to say, these are the pages I need to go back and optimize, because they nudge people towards conversion. Very, very straightforward stuff. Number eight, I should be deploying and only using the machine learning features in advertising systems for my agency. Google AdWords, for example, has tons of incredible AI built into it. But a lot of people whose knowledge is stuck five years back are using the old settings because what they know instead of letting AdWords do the optimization for them, and just say, “Well, you know what Google, you have the big iron. You do it. Just give me leads.” And letting it do that.

Number nine, competitive analysis. Using data science and machine learning. Again, people don’t do this nearly enough of the APC space and they really should be. What are your competitors posting? What are your aspirational competitors posting? If you’re an independent PR firm with 50 to a hundred people, you should be analyzing Edelman digital figuring out what is it that they’re publishing that’s working great. And finally, number 10, analysis of your customers themselves. What are your customers doing? What is their footprint? What data they leaving behind that you can analyze to either help them be better at what they do, or in some cases, if you’ve got a terrible client, you should be identifying that client’s competitors and saying, “This is our pitch list, and here’s exactly what we’re going to say to win that business, and Genesis of this loser.

Drew McLellan:

Okay. I just want to make it very clear to the listeners that I did not preface in our conversation. Chris, I’m going to ask you for 10 that has not been in our email exchange, or this happened in our verbal exchanges. So the fact that you just rattled that off like that was so much more impressive than anyone listening understands. Because they’re all thinking now, well, Drew and Chris set that up in advance. So I just want to give you full props for that, that you did that off the cuffs.

Chris Penn:

Thank you. That’s what we do for Trust Insights, because we have no choice. We’re a startup. We have $0 in investment. We couldn’t get anyone to invest in the company when we started it. So like, “Okay, we’ve got to build all this stuff ourselves, a dewar for ourselves so that we can survive.”

Drew McLellan:

So, how do you keep current?

Chris Penn:

The number one way I keep current is by trying to solve problems for either the company or for clients. Because when they bring us interesting and challenging problems, I can then go, “Okay, well, how are other people solving this problem or problem like this that is structurally the same.” And then as you Google, as you hit up stack overflow in all the different things you realize, “Oh, these are new advancements.” So that’s one aspect. The second aspect is in our content curation software, one of my benchmarks, one of my bench tests of both good social media curation, and good email newsletters is that if when I’m putting together, if I don’t learn something. As I’m putting together a newsletter. It’s not a good newsletter.

So more often than not every day, every week when I’m putting together the company newsletter, I’m reading what the software spit out I’m going, “Oh, crap. I missed that.” Go read the article like, “Huh, I didn’t know that was available now.” And so just by the nature of bringing in 1500 blogs and analyzing them and pulling the top articles from each, the software also provides a little snippet. Like, the most important sentence in that article. I can look at a hundred articles go, “Oh, I need to go read that one. I need to go get that one.” Then that’s not news. So, it was really the machine’s hand delivering information combined with solving tough customer problems really keeps me up-to-date.

Drew McLellan:

So I know my listeners are wondering, you’ve mentioned this software curation or that content curation software a couple of times. I know it’s something you built for yourself, but is that something that your clients have access to, or an agency if they wanted to reach out to you could tap into?

Chris Penn:

We do. We actually do it for two clients. One of whom is an agency and they get a weekly file of, “Here’s the stuff.” And it’s after the initial input call, and asking like, “Okay, what’s good, what’s bad, what is relevant, what’s not?” At that point it’s almost fully automated.

Drew McLellan:

Cool. All right. If folks want to sign up for your weekly newsletter, if they want to follow you, if they want to track you down, what are the best ways for them to find you?

Chris Penn:

Two websites. trustinsights.ai is my company. I co-founded with my CEO, Katie Robbert. And then my personal blog, christopherspenn.com. Those two places you can get to everything else from there, but those are the two places to go.

Drew McLellan:

I could talk to you for about another four hours. Two things. One, thank you very much for being so generous with what you know, and being so quick to say yes when I asked you to be on the show. I knew this would be an amazing episode, and you have proven me correct. So, first of all, thank you.

Chris Penn:

Thank you.

Drew McLellan:

And secondly I am going to ask right in front of everybody, I would really like to have you back at some point in time in the not near future, because I have more questions than I had a chance to get to. So I would love to have you back, and pick your brain a little more if you’re game for that.

Chris Penn:

Sounds good.

Drew McLellan:

Good. Thank you so much. Appreciate it.

Chris Penn:

Thank you for having me.

Drew McLellan:

All right guys. Well actually, if you did not grab the show notes, which has all of this, or you didn’t transcribe it as you were listening, or you didn’t take notes the old fashioned way, you missed out. This episode is packed with stuff to do. And even if you just pick one, just pick one, and start playing. The bottom line is here’s what my takeaway from Mclellan was, most agencies have barely scratched the surface on this. I don’t care where you’re at in the process, unless you’re super far ahead, but you’re not far behind. So you have plenty of time to learn this, but not learning it isn’t really an option.

This is like, for those of you that have been around for a while, remember when people were talking about websites and how they going to be a fad, right? And can you imagine not understanding how the internet works today in our business? That’s what this is for us. This is the next wave of that. And so we’ve got to get smart and we’ve got to learn how to leverage these tools, and we have to use them on both ourselves and our clients.

So please take some of the takeaways that Chris gave you and put them to you. So, one thing I wanted to note. So as you know, we have a sponsor of the… presenting sponsor of the podcast is White Label IQ. They do all kinds of white label, wholesale price, design dev, and PPC. So as you’re listening to Chris, and you’re thinking I don’t have developers on my team. How do I start experimenting with this. White Label might be a great place to go, and A, thank them for being the presenting sponsor of the podcast, but B they might be a great partner for you. So give them a call and explore that.

In the meantime, I will be back next week with another guest who’s going to really get you thinking differently like Chris did today. And in the meantime, you can track me down at drewatagencymanagementinstitute.com. I will talk to you next week. Thanks for listening. That’s all for this episode of AMI’s Build a Better Agency podcast. Be sure to visit agencymanagementinstitute.com to learn more about our workshops, online courses, and other ways we serve small to mid-sized agencies. Don’t forget to today so you don’t miss an episode.